some connections between open problems · 10/25/2018 · write : u!1cat, i.e. (x) is the...
TRANSCRIPT
![Page 1: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/1.jpg)
Some connections betweenopen problems
Nicolai Kraus
University of Nottingham
HoTTEST, 25 Oct 2018
![Page 2: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/2.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 3: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/3.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 4: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/4.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 5: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/5.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).
UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 6: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/6.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 7: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/7.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 8: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/8.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)
NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 9: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/9.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).
APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 10: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/10.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 11: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/11.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 12: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/12.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 13: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/13.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 14: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/14.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 15: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/15.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 16: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/16.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 17: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/17.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 18: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/18.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 19: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/19.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 20: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/20.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 21: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/21.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 22: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/22.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 23: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/23.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 24: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/24.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 25: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/25.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 26: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/26.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 27: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/27.jpg)
Open Problems
Define semi-simplicial types
A theory of∞-categories
appli-cations
HoTT“eating” itself
“Minimal”truncation UP
Open HoTT problems(suspension of set, . . . )
completesemi-Segal,cf. Capriottiet al.’16-18
trivial
viahigherCwFs?
reflection, cf.Shulman’14
(-1): K.’15
teaser:this talk
SEMISIMPLICIAL TYPES
(UF 2012/13, Lumsdaine et al.)
A0 :UA1 :A0 × A0 → UA2 : Π(x, y, z : A0), A1(x, y)
× A1(y, z)× A1(x, z)→ U
δ
x
y z
w
f g
h
jk
Example:A0 ≡ x, y, z, wA1(x, y) ≡ f, gA1(x,w) ≡ h,. . .A2(.., g, j, h) ≡ δ
PROBLEM: Find F : N→ U1such that F(n) 'type of tuples (A0, . . . ,An).UNSOLVED in “book-HoTT”,solved in Voevodsky’s HTS,our 2LTT (arXiv:1705.03307)
NOTE: semi is what allows theabove encoding.
HIGHER CATEGORIES
TASK: develop (∞, 1)-cat’s inHoTT (internally)WHY: occur everywhere inHoTT, e.g. “HIT H is initialin the category of H-algebras”(not captured by AKS)NOTE: “Semisimplicial types” ≈functors ∆op
+ → U(“simpl. types” ≈ ∆op → U ,others see K.-Sattler’17).APPROACH (if given SST):mimic Rezk’s Segal spaces (re-place space by type);issue: semi
COMPLETE SEMI-SEGAL TYPE
Def:• semisimpl. type (A0, A1, . . .)• Segal cond. (aka horn filling)• completeness (Harpaz’15)(note: Segal, compl. are prop)
(comp.)
(assoc.)
Completeness:
x xee
x xid
ee
Special case:
• A1 family of sets(⇒ only need A0, . . . , A3)' AKS categories
(also works one level higher).Capriotti-K.’18
TYPE THEORY IN TYPE THEORY
QUESTION (SHULMAN’14): DoesHoTT with (n+1) universesmodel HoTT with n universes?
Can we implement HoTT inHoTT?
Some partial results, e.g.Altenkirch-Kaposi’s type the-ory in type theory, Escardó,Xu, Buchholtz, Lumsdaine,Weaver, Tsementzis, . . .
PROBLEM:
• syntax should be a set• set-truncate ⇒ HoTT
not a model• idea: use ∞-cat’s to add
all coherences (cf. Al-tenkirch)
TT-IN-TT ⇒ SST: generate SSTterms, reflect them into thehost type theory.
ELIMINATING TRUNCATIONS
What is ‖A‖n → B(for untruncated B)?
E.g. if B is set:(‖A‖−1 → B)
' Σ(f : A→ B),Π(x, y : A), fx = fy
We have constructions of trun-cations (v Doorn, Rijke, K.)which give elimination princi-ples into arbitrary types –but difficult to apply!
USING HIGHER CAT’S
• Write η : U → ∞CAT,i.e. η(X) is the ∞-groupoid of X (RF re-placement of X).• Write coskeln :∞CAT→∞CAT for the [n]-coskeleton (removes cellsabove dimension n)
CONJECTURE: (‖A‖n → B)' (coskeln+1(ηA)→ ηB)
RESULT:
Works for n ≡ −1 (K.’15)
ELEMENTARY OPEN PROBLEMS
• Free ∞-group over a set:is it a set?• Suspension of a set:
is it a 1-type?• Adding a path to a 1-
type: what can we say?
What do these have to do withthe other questions?
![Page 28: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/28.jpg)
“Elementary” ProblemsWedge of A-many circles:
A 1 WA
Or, as HIT:
data WA
base : WA
loops : A→ base = base
Free group FG(A) is Ω(WA).If A is a set⇒ is FG(A) a set?
Slightly more general:
A 1
1 ΣA
If A is a set, is ΣA a 1-type?
![Page 29: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/29.jpg)
“Elementary” ProblemsWedge of A-many circles:
A 1 WA
Or, as HIT:
data WA
base : WA
loops : A→ base = base
Free group FG(A) is Ω(WA).If A is a set⇒ is FG(A) a set?
Slightly more general:
A 1
1 ΣA
If A is a set, is ΣA a 1-type?
![Page 30: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/30.jpg)
“Elementary” Problems (cont.)Adding a path:
2 1
B B
If B is a 1-type, is the pushoutB still a 1-type?
Generalization of the above:
A C
B D
If A is a set and B,C are 1-types, is D a 1-type?
Notes: (1) Yes, if LEM; (2) Probably no, if we try togeneralize further (thanks to P. Capriotti for the left example):
2 1
∥∥S2∥∥2
+∥∥S2
∥∥2
∥∥S2∥∥2∨∥∥S2
∥∥2
S1 1
1 S2
![Page 31: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/31.jpg)
“Elementary” Problems (cont.)Adding a path:
2 1
B B
If B is a 1-type, is the pushoutB still a 1-type?
Generalization of the above:
A C
B D
If A is a set and B,C are 1-types, is D a 1-type?
Notes: (1) Yes, if LEM; (2) Probably no, if we try togeneralize further (thanks to P. Capriotti for the left example):
2 1
∥∥S2∥∥2
+∥∥S2
∥∥2
∥∥S2∥∥2∨∥∥S2
∥∥2
S1 1
1 S2
![Page 32: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/32.jpg)
“Elementary” Problems (cont.)Adding a path:
2 1
B B
If B is a 1-type, is the pushoutB still a 1-type?
Generalization of the above:
A C
B D
If A is a set and B,C are 1-types, is D a 1-type?
Notes: (1) Yes, if LEM; (2) Probably no, if we try togeneralize further (thanks to P. Capriotti for the left example):
2 1
∥∥S2∥∥2
+∥∥S2
∥∥2
∥∥S2∥∥2∨∥∥S2
∥∥2
S1 1
1 S2
![Page 33: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/33.jpg)
The Difficulty
How to characterize path spaces inA C
B D
?
If A is a set and B,C are 1-types, we want path spaces in Dto be sets.
Favonia-Shulman’s HoTT version of the Seifert-van Kampentheorem characterizes path spaces as lists.
![Page 34: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/34.jpg)
The Difficulty
How to characterize path spaces inA C
B D
?
If A is a set and B,C are 1-types, we want path spaces in Dto be sets.
Favonia-Shulman’s HoTT version of the Seifert-van Kampentheorem characterizes path spaces as lists.
![Page 35: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/35.jpg)
The Difficulty
How to characterize path spaces inA C
B D
?
If A is a set and B,C are 1-types, we want path spaces in Dto be sets.
Favonia-Shulman’s HoTT version of the Seifert-van Kampentheorem characterizes the set-truncation of path spaces asset-quotients of lists.
(. . . so this does not answer the question.)
![Page 36: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/36.jpg)
The Difficulty
How to characterize path spaces inA C
B D
?
If A is a set and B,C are 1-types, we want path spaces in Dto be sets.
Favonia-Shulman’s HoTT version of the Seifert-van Kampentheorem characterizes the set-truncation of path spaces asset-quotients of lists.
(. . . so this does not answer the question.)
The problem:
Forming “coherent quotients” without truncating is hard.
![Page 37: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/37.jpg)
Quotienting by directed relationsWhat I need from a directed relation:
• a set W : U , together with a function deg : W → N;• a family ;: W ×W → U ;• such that (w ; v)→ deg(w) > deg(v).• for (w, v, u : W ) such that w ; v and w ; u, we gett : W together with v ;rt t and u ;rt t.
Here, ;rt is the refl-trans closure of ;, as in:
data ;rt: W → W → Unil : w : W → (w ;rt w)
cons : w, v, u : W → (w ;rt v)
→ (v ; u)→ (w ;rt u)
Idea: We want (W/ ;).
![Page 38: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/38.jpg)
Quotienting by directed relationsWhat I need from a directed relation:
• a set W : U , together with a function deg : W → N;• a family ;: W ×W → U ;• such that (w ; v)→ deg(w) > deg(v).• for (w, v, u : W ) such that w ; v and w ; u, we gett : W together with v ;rt t and u ;rt t.
Here, ;rt is the refl-trans closure of ;, as in:
data ;rt: W → W → Unil : w : W → (w ;rt w)
cons : w, v, u : W → (w ;rt v)
→ (v ; u)→ (w ;rt u)
Idea: We want (W/ ;).
![Page 39: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/39.jpg)
General case – undirected relationConsider any set X with ∼: X → X → U .Write (X/ ∼) for the set-quotient, with [−] : X → (X/ ∼).
Fact: For any x, y : X, the canonical map
Φ : (x ∼rst y)→ ([x] = [y]) (1)
is surjective.
Consequence: If Y is a 1-type, then the type
(X/ ∼)→ Y (2)
is equivalent to the type of triples (f, p, q) where
f : X → Y
p : Π(a, b : X), (a ∼ b)→ f(a) = f(b)
q : Π(a : X), (l : a ∼rst a), p∗(a, l) = refl
(3)
![Page 40: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/40.jpg)
General case – undirected relationConsider any set X with ∼: X → X → U .Write (X/ ∼) for the set-quotient, with [−] : X → (X/ ∼).
Fact: For any x, y : X, the canonical map
Φ : (x ∼rst y)→ ([x] = [y]) (1)
is surjective.
Consequence: If Y is a 1-type, then the type
(X/ ∼)→ Y (2)
is equivalent to the type of triples (f, p, q) where
f : X → Y
p : Π(a, b : X), (a ∼ b)→ f(a) = f(b)
q : Π(a : X), (l : a ∼rst a), p∗(a, l) = refl
(3)
![Page 41: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/41.jpg)
General case – undirected relationConsider any set X with ∼: X → X → U .Write (X/ ∼) for the set-quotient, with [−] : X → (X/ ∼).
Fact: For any x, y : X, the canonical map
Φ : (x ∼rst y)→ ([x] = [y]) (1)
is surjective.
Consequence: If Y is a 1-type, then the type
(X/ ∼)→ Y (2)
is equivalent to the type of triples (f, p, q) where
f : X → Y
p : Π(a, b : X), (a ∼ b)→ f(a) = f(b)
q : Π(a : X), (l : a ∼rst a), p∗(a, l) = refl
(3)
![Page 42: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/42.jpg)
General case – undirected relation
f : X → Y
p : Π(a, b : X), (a ∼ b)→ f(a) = f(b)
q : Π(a : X), (l : a ∼rst a), p∗(a, l) = refl
“Polygon” in X:
••
•
•
••
•
•
∼∼
∼
∼
∼∼
∼
∼
f : maps points to points in Yp: maps lines to equalities in Yq: polygon in Y is trivial
![Page 43: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/43.jpg)
General case – undirected relation
f : X → Y
p : Π(a, b : X), (a ∼ b)→ f(a) = f(b)
q : Π(a : X), (l : a ∼rst a), p∗(a, l) = refl
“Polygon” in X:
••
•
•
••
•
•
∼∼
∼
∼
∼∼
∼
∼
f : maps points to points in Yp: maps lines to equalities in Yq: polygon in Y is trivial
![Page 44: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/44.jpg)
What do we gain from a directed relation?
“Polygon” in W , ; directed,vertices are the degrees:
4
3
2
7
6
4
9
1
;;
;
;
;;
;
;
![Page 45: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/45.jpg)
What do we gain from a directed relation?
“Polygon” in W , ; directed,vertices are the degrees:
4
3
2
7
6
4
9
1
;;
;
;
;;
;
;
1
5
![Page 46: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/46.jpg)
What do we gain from a directed relation?
“Polygon” in W , ; directed,vertices are the degrees:
4
3
2
7
6
4
9
1
;;
;
;
;;
;
;
1
5 2
![Page 47: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/47.jpg)
What do we gain from a directed relation (cont)?
Lexicographical order on ordered lists on N is well-founded
⇒ this process necessarily terminates with the trivial polygon
⇒ every “polygon” can be disassembled into “confluencepolygons”
⇒ inf : X → Y
p : Π(a, b : X), (a ; b)→ f(a) = f(b)
q : Π(a : X), (l : a ;rst a), p∗(a, l) = refl
it is enough if q quantifies over those shapes given from theconfluence condition!
q : confluence shapesp7→ commuting polygons
![Page 48: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/48.jpg)
What do we gain from a directed relation (cont)?
This gives a fairly easy way to construct functions
(W/ ;)→ Y
if Y is a 1-type!
Back to the SvK theorem:
A (set) C (1-type)
B (1-type) D (???)
SvK theorem (Fav.-Shulm.): (lists/ ;) ' ‖eq-types in D‖0Now we can get: (lists/ ;)→ ‖eq-types in D‖1(which is the hard part of “2nd hom-groups of D are trivial”).
![Page 49: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/49.jpg)
What do we gain from a directed relation (cont)?
This gives a fairly easy way to construct functions
(W/ ;)→ Y
if Y is a 1-type!
Back to the SvK theorem:
A (set) C (1-type)
B (1-type) D (???)
SvK theorem (Fav.-Shulm.): (lists/ ;) ' ‖eq-types in D‖0Now we can get: (lists/ ;)→ ‖eq-types in D‖1(which is the hard part of “2nd hom-groups of D are trivial”).
![Page 50: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/50.jpg)
What do we gain from a directed relation (cont)?
This gives a fairly easy way to construct functions
(W/ ;)→ Y
if Y is a 1-type!
Back to the SvK theorem:
A (set) C (1-type)
B (1-type) D (???)
SvK theorem (Fav.-Shulm.): (lists/ ;) ' ‖eq-types in D‖0Now we can get: (lists/ ;)→ ‖eq-types in D‖1(which is the hard part of “2nd hom-groups of D are trivial”).
![Page 51: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/51.jpg)
How to generalize?
f : X → Y
p : Π(a, b : X), (a ∼ b)→ f(a) = f(b)
q : confluence shapesp7→ commuting polygons
Next step (Y is a 2-type): q maps the boundary of anypolyhedron whose sides are confluence shapes to a trivialpolyhedron in Y ?Seems difficult. . .
Idea: In many examples (including the SvK-example), wehave strong confluence (reflexive instead of refl-trans closure).
⇒ Only need to consider cubes!Seems much more doable (even in arbitrary dimensions?)
Thank you for your attention!
![Page 52: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/52.jpg)
How to generalize?
f : X → Y
p : Π(a, b : X), (a ∼ b)→ f(a) = f(b)
q : confluence shapesp7→ commuting polygons
Next step (Y is a 2-type): q maps the boundary of anypolyhedron whose sides are confluence shapes to a trivialpolyhedron in Y ?Seems difficult. . .
Idea: In many examples (including the SvK-example), wehave strong confluence (reflexive instead of refl-trans closure).
⇒ Only need to consider cubes!Seems much more doable (even in arbitrary dimensions?)
Thank you for your attention!
![Page 53: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/53.jpg)
How to generalize?
f : X → Y
p : Π(a, b : X), (a ∼ b)→ f(a) = f(b)
q : confluence shapesp7→ commuting polygons
Next step (Y is a 2-type): q maps the boundary of anypolyhedron whose sides are confluence shapes to a trivialpolyhedron in Y ?Seems difficult. . .
Idea: In many examples (including the SvK-example), wehave strong confluence (reflexive instead of refl-trans closure).
⇒ Only need to consider cubes!Seems much more doable (even in arbitrary dimensions?)
Thank you for your attention!
![Page 54: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/54.jpg)
How to generalize?
f : X → Y
p : Π(a, b : X), (a ∼ b)→ f(a) = f(b)
q : confluence shapesp7→ commuting polygons
Next step (Y is a 2-type): q maps the boundary of anypolyhedron whose sides are confluence shapes to a trivialpolyhedron in Y ?Seems difficult. . .
Idea: In many examples (including the SvK-example), wehave strong confluence (reflexive instead of refl-trans closure).
⇒ Only need to consider cubes!Seems much more doable (even in arbitrary dimensions?)
Thank you for your attention!
![Page 55: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/55.jpg)
ReferencesMentioned or related papers and talks, roughly in order ofoccurrence. (Many papers have been published “formally”, clickablearXiv links are for convenience.)• Vladimir Voevodsky. A type system with two kinds of identity
types. 2013.• Danil Annenkov, Paolo Capriotti, and Nicolai Kraus. Two-level
type theory and applications. 2017. arXiv:1705.03307• Paolo Capriotti. Models of Type Theory with Strict Equality.
2016. arXiv:1702.04912• Emily Riehl and Michael Shulman. A type theory for synthetic∞-categories. 2017. arXiv:1705.07442
• Benedikt Ahrens, Chris Kapulkin, and Michael Shulman.Univalent categories and the Rezk completion.2015.arXiv:1303.0584
• Steve Awodey, Nicola Gambino, and Kristina Sojakova.Homotopy-initial algebras in type theory. 2017.arXiv:1504.05531
![Page 56: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/56.jpg)
References (cont.)• Nicolai Kraus and Christian Sattler. Space-valued
diagrams, type-theoretically (extended abstract). 2017.arXiv:1704.04543• Yonatan Harpaz. Quasi-unital ∞-Categories. 2015.
arXiv:1210.0212• Charles Rezk. A model for the homotopy theory of
homotopy theory. 2001. arXiv:math/9811037• Paolo Capriotti and Nicolai Kraus. Univalent Higher
Categories via Complete Semi-Segal Types. 2018.arXiv:1707.03693• Michael Shulman. Homotopy Type Theory should eat
itself (but so far, it’s too big to swallow). 2014. Blogpost.• Thorsten Altenkirch. Towards higher models and syntax
of type theory. HoTTEST talk, 10 May 2018.
![Page 57: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/57.jpg)
References (cont.)• Ambrus Kaposi. Type theory in a type theory with
quotient inductive types. PhD thesis, 2016.• Thorsten Altenkirch and Ambrus Kaposi. Type Theory in
Type Theory Using Quotient Inductive Types. 2016.• Martín Hötzel Escardó and Chuangjie Xu. Autophagia.
2014. Agda file online.• Nicolai Kraus. The General Universal Property of the
Propositional. 2015. arXiv:1411.2682• Paolo Capriotti, Nicolai Kraus, and Andrea Vezzosi.
Functions out of Higher Truncations. 2015.arXiv:1507.01150• Nicolai Kraus. Truncation Levels in Homotopy Type
Theory. PhD thesis, 2015.• Floris van Doorn. Constructing the Propositional
Truncation using Non-recursive HITs. 2016.arXiv:1512.02274
![Page 58: Some connections between open problems · 10/25/2018 · Write : U!1CAT, i.e. (X) is the 1-groupoid of X (RF re-placementofX). Writecoskel n: 1CAT ! 1CAT for the [n]-coskeleton (removes](https://reader036.vdocuments.site/reader036/viewer/2022071018/5fd1ecd07cce4c413636fd2c/html5/thumbnails/58.jpg)
References (cont.)
• Nicolai Kraus. Constructions with non-recursivetruncations. 2016.• Egbert Rijke. The join construction. 2017.
arXiv:1701.07538• Ulrik Buchholtz, Floris van Doorn, and Egbert Rijke.
Higher Groups in Homotopy Type Theory. 2018.arXiv:1802.04315• Nicolai Kraus and Thorsten Altenkirch. Free Higher
Groups in Homotopy Type Theory. 2018.arXiv:1805.02069• Kuen-Bang Hou (Favonia) and Michael Shulman. The
Seifert-van Kampen Theorem in Homotopy Type Theory.2016.