human computation

22
Human Computation Steven Emory CS 575

Upload: albany

Post on 21-Jan-2016

71 views

Category:

Documents


2 download

DESCRIPTION

Steven Emory CS 575. Human Computation. Overview. What is Human Computation? History of Human Computation Examples of Human Computation Bad Example Good Example Challenges in Human Computation. Definition of Computation. Normally we rely on computers to do all the work On input: - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Human Computation

Human Computation

Steven EmoryCS 575

Page 2: Human Computation

Overview

What is Human Computation?

History of Human Computation

Examples of Human Computation

Bad Example

Good Example

Challenges in Human Computation

Page 3: Human Computation

Definition of Computation

Normally we rely on computers to do all the work

On input:

Step #1

Step #2

...

Step #n

Ouput

Page 4: Human Computation

Definition of Human Computation

Some or all steps are solved by human(s)

On input:

Step #1

Step #2 (ask a human)

Step #3 (ask a human)

...

Step #n

Ouput

Page 5: Human Computation

History of Human Computation

1980's: Interactive Genetic Algorithms

2000's: Human-Based Genetic Algorithms

2000's: Outsourced Human Spam

2000's: Interactive Guessing Games

Page 6: Human Computation

Quicksort Example

Bad Example: Quicksort

Input: Unsorted array

Select a pivot (human selection)

Swap last element with pivot element

Partition array using pivot element

Insert pivot element into correct position

Repeat above steps for left/right side partitions

Output: Sorted array

Page 7: Human Computation

Quicksort Problems

Solved faster by computer alone

Not rewarding

Boring

Painful

Page 8: Human Computation

Photomosaic Example

Photomosaic Algorithm

Can be solved by computer alone.

On input “image gallery,” “source image”1.) Tile source image.2.) From left-to-right, top-to-bottom, compare each image in the image gallery to each tile in the source image, inserting the gallery image with the lowest error.3.) Output photomosaic.

Page 9: Human Computation

Photomosaic Problems

Noticable visual artifacts

Could use a randomized algorithm instead

Can we do better using Human Computation?

Of course!

Page 10: Human Computation

Photomosaic Solutions

Optimize important features first

Objects (eyes, nose, mouth)

Edges (chin, facial edges, hair, body)

Revised algorithm:

On input “image gallery,” “source image”1.) Tile source image.2.) Select important features (ask human).3.) Optimize (randomly) important features.4.) Optimize (randomly) unimportant features.5.) Output photomosaic.

Page 11: Human Computation

When to Use

When the problem:

is hard for a computer, but easy for a human

can be done better by a human

needs a human-like (artistic/creative) solution

is not boring to a human (music, art, games)

is rewarding (financially or emotionally)

is one and done

Page 12: Human Computation

Human Computation Challenges

User interface design

Coordinating many human participants

Analagous to distributed computing

Honesty

Prolonged computation

Page 13: Human Computation

reCAPTCHA Example

Used to digitize old books (make e-books)

OCR normally works 99% of the time

OCR accuracy drops for older books

Old paper

Old printing techniques

Solution: Ask humans to determine words OCR fails to classify

When enough humans agree, consider it solved

Page 14: Human Computation

reCAPTCHA Example

Page 15: Human Computation

Metadata Example

Algorithm: Assign metadata to images

Useful for content/multimedia management systems (i.e. Istockphoto)

No algorithm exists for image labeling

Luis Von Ahn's solution: The ESP Gamehttp://www.espgame.org

ESP Game Demo

Page 16: Human Computation

Metadata Example

Problem:

Getting humans to agree correctly

Page 17: Human Computation

Electric Sheep Example

Brief fractal explanation:

Iterative process based on chaos, dynamical systems

Newton's Method FractalSolve az3 + bz2 + cz + d = 0 for complex numbersa, b, c, d are fractal parameters

Cubic equation has 3 rootsRed = converges to root #1Green = converges to root #2Blue = converges to root #3Black = fails to converge

Page 18: Human Computation

Electric Sheep Example

Page 19: Human Computation

Electric Sheep Example

Interactive genetic algorithm

Humans evaulate fitness

Animated fractal parameters are mutated

Algorithm has been running for years

Implemented as a screensaver

http://www.electricsheep.org/

Page 20: Human Computation

Electric Sheep Example

Page 21: Human Computation

Conclusions

There many hard/impossible to solve problems

Nothing shameful about using Human Computation

Applications in art, music, computer vision, security, content management

Page 22: Human Computation

References

Luis Von Ahn

Google Talk lecture on Human Computation

http://www.cs.cmu.edu/~biglou/

http://www.espgame.org

The Art of Artificial Evolution (2008)

http://www.springerlink.com/content/r68831/?p=d04c460aa17749eb8153fec3d0507f68&pi=9