1. CS 540 - Fall 2016 (Shavlik©), Lecture 12, Week 6
2/23/2019
Today’s Topics
Exam Review Thursday 10/20/16 (see note in Moodle)
Genetic Algorithms
Mutation
Crossover
Fitness-proportional Reproduction
Premature Convergence
Building-block Hypothesis
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

2. Genetic Algorithms (GAs)
Use ideas of
Survival of fittest (death)
Combination of ‘genetic material’ (sex)
(‘Taxes’ play a role in some algo’s)
Mutation (randomness)
Mixing of genes from parents more important than mutation (contrary to popular press)
About 25,000 human genes
For simplicity, assume two variants of each
So 225,000 possible combo’s to explore!
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

3. Basic FRAMEWORK for GAs (many possible ALGORITHMS)
Create initial population of entities
Evaluate each entity using a fitness function
Discard worst N% of entities
K times, stochastically grab ‘best’ parents (fitness proportional reproduction)
Combine them (crossover) to create new entities
Make some random changes (mutation)
Goto 2
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

4. Representing Entities as Bit Strings
Assume we represent our problem as a bit string (but any data structure ok for GAs)
Cross Over (example on next slide)
Pick two entities, A and B
Choose a cross-over location
Copy first part of A and last part of B
Copy first part of B and last part of A
Mutation
Randomly flip 1 or more bits
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

5. CS 540 - Fall 2016 (Shavlik©), Lecture 12, Week 6
Crossover Example
Entity A 1
Randomly chosen ‘cross over’ point
Entity B 1
Child C 1
Child D 1
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

6. Aside: My Family Phones
My cell phone (#’s changed for anonymity) My wife’s cell phone Our daughter’s cell phone
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

7. CS 540 - Fall 2016 (Shavlik©), Lecture 12, Week 6
Typical Design
Discard Worst HALF of Population
Generate Children to Refill Population
Keep Parents and Generated Children
‘Flip’ a Small Faction of Bits (eg, 0.1%)
Flip bits in all member of population
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

8. Fitness-Proportional Reproduction
Let Fi be the fitness of entity i
Assume Fi are non-negative (if not, use eFi as the fitness for the GA)
Let Ftotal = ∑ Fi // Sum the fitness of all the entities
Prob (entity i chosen) = Fi / Ftotal
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

9. CS 540 - Fall 2016 (Shavlik©), Lecture 12, Week 6
Roulette-Wheel View - spin arrow and see where it stops (pie-wedge size proportional to fitness)
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

10. A GA Approach to Supervised ML
Assume we want to learn a model of the form below (and all of our N features are numeric)
if [ ∑ weighti ₓ featurei ] > threshold
then return POS else return NEG
Representation of Entities?
See next slide
Fitness?
Accuracy on TRAIN set plus maybe some points for being different from rest of population
Role of Tuning Set?
Could chose best member of population when done
If we use ALL of population (an ‘ensemble’), could weight each’s predictions based on tuneset accuracy
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

11. Possible Representation of Entities
. . . … … … … …
16 bits for weight1
16 bits for weightN
16 bits for threshold
Notes
1) we might only use 16 bits so weights are small (Occam’s Razor)
2) first bit could be SIGN (or use “2’s complement”)
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

12. CS 540 - Fall 2016 (Shavlik©), Lecture 12, Week 6
Design Tip
Design your space of entities so that most are viable (ie, get a non-zero fitness)
Otherwise will waste a lot of cpu cycles generating useless entities
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

13. Premature Convergence (‘Inbreeding’)
If not careful, entire population can become minor variations of a small number of ‘bit vectors’
Eg, consider crossing over A and child_of_A
Result will be  ¾ a copy of A
Solutions
Don’t crossover with ‘recent’ descendent
Mutate more (but might destroy good traits)
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

14. GAs as Searching a Space
Consider the space defined by single-bit mutations
What is a CROSSOVER?
Grab any two nodes (might not be adjacent)
‘hyper jump’
to a possibly
distant 3rd node
101…01
001…01
101…00
001…00
101…10
011…10
001…00
etc
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

15. Building-Block Hypothesis
GAs work well when overall task has subtasks
And fitness function gives credit for being able to solve subtasks
Crossover ‘mixes and matches’ solutions to subtasks
Eg, consider building cars
Need to engine, wheels, windows, brakes, etc
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

16. Which Fitness Function Better for GAs?
State Space
Fitness
State Space
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

17. CS 540 - Fall 2016 (Shavlik©), Lecture 12, Week 6
Genetic Programming
Entities need not be bit strings
Often ‘genetic programming’ used for richer rep’s of entities
Decision trees
Neural networks
Code snippets
Etc
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

18. CS 540 - Fall 2016 (Shavlik©), Lecture 12, Week 6
In-Class HW
Design Genetic Programming Approach for Creating Good Decision Trees
Think for 1-2 Mins before Raising Hand
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6

19. CS 540 - Fall 2016 (Shavlik©), Lecture 12, Week 6
GA Wrapup
Can come up with quite creative solutions since many possibilities considered
Might be too undirected?
Designing good fitness functions can be a challenge
Make more sense as computing power 
10/13/16
CS Fall 2016 (Shavlik©), Lecture 12, Week 6