CITS4404 Artificial Intelligence & Adaptive Systems

Slides:



Advertisements
Similar presentations
G5BAIM Artificial Intelligence Methods
Advertisements

Constraint Optimization We are interested in the general non-linear programming problem like the following Find x which optimizes f(x) subject to gi(x)
Parameter control Chapter 8. A.E. Eiben and J.E. Smith, Introduction to Evolutionary Computing Parameter Control in EAs 2 Motivation 1 An EA has many.
1 Evolutionary Computational Inteliigence Lecture 6b: Towards Parameter Control Ferrante Neri University of Jyväskylä.
Parameter Control A.E. Eiben and J.E. Smith, Introduction to Evolutionary Computing Chapter 8.
Evolutionary Computational Intelligence Lecture 10a: Surrogate Assisted Ferrante Neri University of Jyväskylä.
Evolutionary Computational Inteliigence Lecture 6a: Multimodality.
Multimodal Problems and Spatial Distribution Chapter 9.
Evolutionary Computational Intelligence Lecture 8: Memetic Algorithms Ferrante Neri University of Jyväskylä.
Intro to AI Genetic Algorithm Ruth Bergman Fall 2004.
Metaheuristics The idea: search the solution space directly. No math models, only a set of algorithmic steps, iterative method. Find a feasible solution.
Image Registration of Very Large Images via Genetic Programming Sarit Chicotay Omid E. David Nathan S. Netanyahu CVPR ‘14 Workshop on Registration of Very.
Genetic Algorithms: A Tutorial
IE 607 Constrained Design: Using Constraints to Advantage in Adaptive Optimization in Manufacturing.
An efficient distributed protocol for collective decision- making in combinatorial domains CMSS Feb , 2012 Minyi Li Intelligent Agent Technology.
Global Optimisation Global optimisation: a search through a set of solutions S for the solution x* that is best according to some measure of fitness.
PRACTICALITIES CITS4404 Artificial Intelligence & Adaptive Systems.
Multimodal Optimization (Niching) A/Prof. Xiaodong Li School of Computer Science and IT, RMIT University Melbourne, Australia
Artificial Intelligence Methods Neural Networks Lecture 4 Rakesh K. Bissoondeeal Rakesh K. Bissoondeeal.
Evolving Virtual Creatures & Evolving 3D Morphology and Behavior by Competition Papers by Karl Sims Presented by Sarah Waziruddin.
1 “Genetic Algorithms are good at taking large, potentially huge search spaces and navigating them, looking for optimal combinations of things, solutions.
GENETIC ALGORITHM A biologically inspired model of intelligence and the principles of biological evolution are applied to find solutions to difficult problems.
2005MEE Software Engineering Lecture 11 – Optimisation Techniques.
Niching Genetic Algorithms Motivation The Idea Ecological Meaning Niching Techniques.
/ 26 Evolutionary Computing Chapter 8. / 26 Chapter 8: Parameter Control Motivation Parameter setting –Tuning –Control Examples Where to apply parameter.
Evolutionary Design (2) Boris Burdiliak. Topics Representation Representation Multiple objectives Multiple objectives.
CITS7212: Computational Intelligence An Overview of Core CI Technologies Lyndon While.
Evolutionary Computing Chapter 11. / 7 Chapter 11: Non-stationary and Noisy Function Optimisation What is a non-stationary problem? Effect of uncertainty.
Evolutionary Computing Chapter 12. / 26 Chapter 12: Multiobjective Evolutionary Algorithms Multiobjective optimisation problems (MOP) -Pareto optimality.
Multimodal Problems and Spatial Distribution A.E. Eiben and J.E. Smith, Introduction to Evolutionary Computing Chapter 9.
Genetic Algorithms. Solution Search in Problem Space.
Genetic Algorithms And other approaches for similar applications Optimization Techniques.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Intelligent Exploration for Genetic Algorithms Using Self-Organizing.
Paper Review for ENGG6140 Memetic Algorithms
Optimization Problems
Genetic Algorithms.
School of Computer Science & Engineering
OPERATING SYSTEMS CS 3502 Fall 2017
CSCI 4310 Lecture 10: Local Search Algorithms
Particle Swarm Optimization (2)
Digital Optimization Martynas Vaidelys.
School of Computer Science & Engineering
C.-S. Shieh, EC, KUAS, Taiwan
Hypothesis Testing and Confidence Intervals (Part 1): Using the Standard Normal Lecture 8 Justin Kern October 10 and 12, 2017.
Announcements HW4 due today (11:59pm) HW5 out today (due 11/17 11:59pm)
Multimodal Problems and Spatial Distribution
Chapter 4 Beyond Classical Search
Subject Name: Operation Research Subject Code: 10CS661 Prepared By:Mrs
Comparing Genetic Algorithm and Guided Local Search Methods
Advanced Artificial Intelligence Evolutionary Search Algorithm
Dr. Arslan Ornek IMPROVING SEARCH
Multimodal Problems and Spatial Distribution
Working with Evolutionary Algorithms
Genetic Algorithms: A Tutorial
Optimization Problems
Heuristic Optimization Methods Pareto Multiobjective Optimization
Where did we stop? The Bayes decision rule guarantees an optimal classification… … But it requires the knowledge of P(ci|x) (or p(x|ci) and P(ci)) We.
Multi-Objective Optimization
“Hard” Optimization Problems
Parameter control Chapter 8.
Multimodal Problems and Spatial Distribution
Boltzmann Machine (BM) (§6.4)
CS 188: Artificial Intelligence Fall 2008
CHAPTER I. of EVOLUTIONARY ROBOTICS Stefano Nolfi and Dario Floreano
Computer Vision Lecture 19: Object Recognition III
Traveling Salesman Problem by Genetic Algorithm
Evolutionary Ensembles with Negative Correlation Learning
Parameter control Chapter 8.
Genetic Algorithms: A Tutorial
Parameter control Chapter 8.
Presentation transcript:

CITS4404 Artificial Intelligence & Adaptive Systems practicalities CITS4404 Artificial Intelligence & Adaptive Systems

Issues in applying CI techniques Global optimisation, some definitions Fitness progression Generality Role of domain knowledge Niching and speciation Memetic optimisation Multi-objective problems Co-evolution Constraints Noisy and dynamic problems Offline vs online learning Supervised, reinforcement, unsupervised learning Experimental methodology, performance measures Parameter tuning and/or control

Three basins of attraction Global optimisation Given a function f and a set of solutions S, search for x*  S such that x  S: f(x*) beats f(x) The graph below depicts a simple 1D fitness landscape Three basins of attraction x* Global optimum Local optima http://www.cs.vu.nl/~gusz/ecbook/ecbook-course.html

Definitions The function f giving the fitness for each member of S is called the fitness landscape The best solution wrt f is called the global optimum Note that there may be multiple (equal) global optima Non-global optima that are better than other “similar solutions” are called local optima The part of the landscape dominated by an optimum is called its basin of attraction Diversity refers to the distribution of a set of solutions over the fitness landscape Diversity preservation techniques try to ensure a good distribution

Heuristic optimisation CI techniques are heuristic optimisers, also known as generate-and-test optimisers Searching procedures that use rules (inspired by nature) to decide which solution(s) to try next The simplest heuristic optimisers are hill-climbers Given one solution, generate similar solutions, and keep the best of them A hill-climber is guaranteed to find a local optimum They can exploit their basin of attraction, but they lack the ability to explore the entire landscape properly CI techniques use populations and other tricks to promote both exploration and exploitation More about this later under memetic algorithms

Fitness progression CI techniques are anytime algorithms The fitness of the best known solution will improve over time, but usually under the law of diminishing returns The convergence rate (the rate of improvement) tends to fall over time Several shorter runs may be better than one long run Best fitness in population Time (number of generations) Progress in 2nd half Progress in 1st half

Smart initialisation Initial solutions can be generated randomly, or using domain knowledge, often called smart initialisation This can improve both results and time performance, but it may also introduce bias into the search Best fitness in population Time (number of generations) T: time needed to reach equivalent fitness of “smart” initialisation T

CI vs problem-specific methods CI techniques are general-purpose Robust techniques giving good performance over a range of problems Special problem-tailored method Computational intelligence approach Performance Random search P Scale of “all” problems

Scale of “all” problems Domain knowledge Performance can sometimes be improved by incorporating “expertise” into the process EA 4 EA 3 EA 2 Performance EA 1 P Scale of “all” problems

Contd. Too little domain knowledge makes the search space bigger and can make the search inefficient cf. EA1 Too much domain knowledge can exclude novel solutions cf. EA4 But care must be taken! “If you tell the system what the solution looks like, that’s what it’ll give you!” [R.L. While] But this can all be highly non-obvious…

Niching and speciation Most interesting problems are multi-modal Sometimes we want to discover more than just the global optimum i.e. we want to discover y* and z*, as well as x* This might be important to offer extra robustness Often it is hard for a fitness function to capture everything x* y* z*

Contd. Each basin of attraction is called a niche, or sometimes a species Niching can be achieved in two broad ways Implicit niching is achieved by modifying the solution representation Explicit niching is achieved by promoting dissimilar solutions, or penalising similar solutions Both techniques rely on having some distance metric between solutions

Memetic algorithms CI techniques are good at exploration – finding high peaks in the fitness landscape – but are less good at exploitation of those peaks Memetic algorithms combine CI with some local-search technique that is good at exploitation AKA Baldwinian, Lamarckian, or cultural algorithms Hill-climbing is the classic example CI finds the best basin of attraction, then hill-climbing climbs the peak The two techniques can be applied in series, or in parallel

Multi-objective optimisation In many problems, solutions are assessed wrt several criteria e.g. speed vs safety vs price for car designs Fitness is now a vector, not a scalar, which complicates selection Vectors have only a partial ordering, rather than a total ordering The “solution” to a multi-objective problem is a set of solutions offering different trade-offs between the objectives It is important to make no a priori assumptions about trade-off weights or the shape of the solution set

Contd. Two objectives f1 and f2, both being maximised f2 Rank = 0 A and B are non-dominated Each solution is plotted by its values in the objectives A Rank = 0 B Rank = 0 X dominates Y if it is better in all objectives Rank = 1 The rank of X is the number of solutions that dominate X Rank = 1 Rank = 0 Rank = 4 P Q P dominates Q Selection is based on ranks Rank = 2 f1

Co-evolution Fitness is sometimes assessed by the interactions between solutions, rather than in isolation e.g. build a team which is the best in the AFL The above is an example of competitive co-evolution Aim for an “arms race” between solutions to drive improvement through parallel adaptation The alternative is co-operative co-evolution Decompose a problem into simpler sub-tasks (layered learning), and combine the sub-solutions to solve the original problem

Constraints A constraint is a requirement placed on a solution, as opposed to a measure of quality e.g. the length of this component must be less than X, or the power of that component must be more than Y A feasible solution is one that satisfies all constraints An infeasible solution fails at least one constraint All solutions Feasible Feasible Feasible

Contd. There are many different constraint-handling techniques Separatist: consider objectives and constraints separately Purist: discard all infeasible solutions when they arise Repair: repair infeasible solutions when they arise Penalty: modify the fitness function to penalise infeasible solutions MOOP: add an extra objective function that measures “degree of infeasibility” Hybrid: some combination of the above

Noisy fitness landscape Noisy problems A noisy fitness function arises when the fitness calculations aren’t perfect Fitness landscape Noise landscape Noisy fitness landscape

Contd. The algorithm can now only estimate performance Again, this complicates selection Bad solutions might get lucky and survive Good solutions might get unlucky and die These behaviours cause undesirable long-term effects The learning rate is reduced Learning may not be retained The usual approach to this is resampling Evaluate the fitness multiple times and average the results But how many times is sufficient? A second common approach is to try to bound the error Basically to assume the error won’t exceed a certain magnitude

Dynamic problems With some problems, the fitness landscape changes over time, maybe due to Temporal effects External factors System adaptation The system needs to adapt to this change Requires online learning http://www.natural-selection.com

Offline vs online learning Offline learning is where a system learns before use The strategy is fixed once training is completed Requires comprehensive training data Only feasible in well-understood environments Online learning is where a system learns while in use The strategy is adapted from each instance encountered Initial decisions are made from incomplete training data Usually much greater time-pressure to improve

Learning paradigms Supervised learning is (offline) training by comparing a system’s responses to expected responses in training data Reinforcement learning is (online) training using feedback from the environment to assess the quality of responses Unsupervised learning is (offline) training with no training data No real question presented The system looks for patterns in the data

Unsupervised learning Question: tell me about this data Question: tell me about this data

Supervised learning Question: what is an apple?

Experimental methodology CI techniques are stochastic Their results are non-deterministic Thus we should never draw conclusions from a single run Always perform a “large” number of runs Assess results using statistical measures Assess significance using statistical tests When comparing algorithms, it is crucial to make all comparisons fair Give each the same amount of resource Use the same performance measures Try different competition limits

Performance measures Offline performance measures Effectiveness (algorithm quality) Success rate (percentage of “good’ runs) Mean best fitness at termination Efficiency (algorithm speed) CPU to “completion” Number of solutions evaluated Online performance measures Population distribution Fitness distribution Improvement rate

Parameter tuning How do we decide on the various constants in a run of the system? e.g. bigger population, or more generations? How big should mutations be? This can be difficult! Sub-optimal values can seriously degrade performance Choosing good values can take significant time Exhaustive search is usually impractical Good values may become bad during a run

Parameter control Can we get the system to choose parameters automatically? i.e. allow settings to vary during the run Three main alternatives are used Deterministic: change parameters according to some pre-determined schedule, e.g. based on the passage of time Adaptive: change parameters according to some measure of the search progress Self-adaptive: encode the “scope of change” into the solution representation in some way One important goal is to reduce the prevalence of “magic numbers” in the system Still, finding good settings is not easy