2019 Evolutionary Algorithms Review
This 2019 review introduces a new taxonomy for evolutionary algorithms based on User Control Attributes (limiters, explainability, causality, fairness, correction) and surveys traditional and specialized EAs, their applications, challenges, and future directions.
Episodes
Chapters
01Title
A review of evolutionary algorithms proposes a new taxonomy focusing on control, explainability, causality, bias, and corrective measures.
1:34Explained02Preface
An Evolutionary Algorithm replaces manual chemist experimentation to explore chemical problem spaces.
1:47Explained03Introduction
AI science sits at the boundary of philosophy and science, combining theoretical ideas with practical engineering.
1:23Explained04User Control Attributes
Rule-based ML is transitioning to outcome-oriented systems, with UCA (limiters, explainability, causality, fairness, and correction) and trust considerations.
1:50Explained05Control Attributes in ML
Modern ML evaluates models by control attributes including limits, explainability, causality, fairness, and the ability to correct.
1:43Explained06End of Moore's Law
Advances in silicon are hitting physical and economic limits, signaling the end of Moore’s Law.
1:33Explained07SpiNNaker
SpiNNaker uses a million interconnected ARM cores to run Spiking Neural Networks, roughly equating to about 1% of a human brain.
1:36Explained08Hybrid Evolutionary Algorithms
Hybridizing evolutionary algorithms with neural networks is a growing approach to automate complex problems.
1:28Explained09Introduction
EAs balance exploitative local search with explorative stochastic search.
1:19Explained10Figure 2.1
The figure illustrates how hardware capability and algorithmic efficiency co-evolve, expanding the unknown search space.
1:26Explained11End of Dennard Scaling
End of Dennard scaling shifts power-density concerns and enables new hardware concepts and deep learning resurgence.
1:42Explained12No Free Lunch Theorem
No single algorithm dominates across all problems; domain knowledge is needed to achieve efficiency.
1:35Explained13Evolutionary Algorithms Overview
Evolutionary algorithms are population-based metaheuristics driven by selection, variation, and reproduction.
1:21Explained14Fitness and Objectives in EAs
Fitness functions quantify success and guide selection; multi-objective optimization seeks Pareto-optimal solutions.
1:28Explained15Introduction
EAs tackle problems where traditional methods struggle due to resources, dimensionality, or complexity.
1:27Explained16Figure 2.1
A captioned figure linking hardware progress to algorithmic capability and unknown problem spaces.
1:51Explained17End of Dennard Scaling
Shifts in power dynamics and new hardware concepts accompany the deep-learning resurgence.
1:29Explained18No Free Lunch Theorem
No single algorithm outperforms all problems; domain-specific knowledge enhances efficiency.
1:27Explained19Overview of Evolutionary Algorithms
EAs are population-based metaheuristics that evolve solutions through selection, reproduction, and variation.
1:15Explained20Fitness and Objectives in EAs
Fitness measures guide selection; many-objective optimization seeks trade-offs along Pareto fronts.
1:20Explained21Introduction
EAs apply when traditional exploitative or stochastic methods fail due to complexity or resources.
1:23Explained22Figure 2.1
Shows the relationship between hardware improvement and algorithmic progress over time.
1:28Explained23End of Dennard Scaling
Describes how diminishing power density leads to reliance on new computing models and AI advances.
1:31Explained24No Free Lunch Theorem
No universal solver exists; domain knowledge increases efficiency for specific problems.
1:25Explained25Overview of Evolutionary Algorithms
EAs are population-based metaheuristics that use evolution-inspired operators to explore problem spaces.
1:38Explained26Fitness and Objectives in EAs
Fitness functions evaluate performance and multi-objective optimization seeks balanced trade-offs.
1:51Explained27Introduction
EAs address problems where classic methods struggle due to complexity or resource limits.
1:19Explained28Figure 2.1
Depicts how hardware scaling and algorithmic progress enable exploration of unknown problem regions.
1:16Explained29End of Dennard Scaling
Points to new hardware paradigms and AI-driven software growth beyond Dennard limits.
1:17Explained30No Free Lunch Theorem
Domain knowledge is essential for efficient problem solving when using EAs.
1:35Explained31Evolutionary Algorithms Overview
EAs are population-based methods using selection, variation and reproduction to search spaces.
1:20Explained32Fitness and Objectives in EAs
Fitness guides progression; multi-objective EAs aim for Pareto-optimal compromises.
1:22Explained33Introduction
Introductory discussion on when EAs are advantageous relative to traditional methods.
1:43Explained34Figure 2.1
Illustrates hardware vs software growth and the unknowns at the frontier.
1:44Explained35End of Dennard Scaling
Hardware scaling constraints drive exploration of novel computing paradigms.
1:40Explained36No Free Lunch Theorem
Algorithm performance is problem-dependent; no universal best method.
1:36Explained37Overview of Evolutionary Algorithms
EAs are Darwinian metaheuristics consisting of population, variation, selection.
1:27Explained38Fitness and Objectives in EAs
The fitness function determines success and guides evolution; multi-objective issues exist.
1:39Explained39Introduction
EAs tackle hard problems where standard methods fail or are inefficient.
1:50Explained40Figure 2.1
Depicts the interplay of hardware and algorithmic progress over time.
1:41Explained41End of Dennard Scaling
Denotes the shift to new computing approaches alongside AI trends.
1:48Explained42No Free Lunch Theorem
No single algorithm excels for all problems; domain knowledge improves efficacy.
1:10Explained43Evolutionary Algorithms Overview
EA families like GAs, GP, GE, CGP, PushGP explore problem spaces via evolution.
1:01Explained44Fitness and Objectives in EAs
Fitness quantifies progress; multi-objective optimization balances conflicting goals.
1:48Explained45Introduction
EAs provide a framework for solving complex, high-dimensional problems.
1:54Explained46Figure 2.1
Illustrates how hardware capabilities enable broader exploration of problem spaces.
1:49Explained47End of Dennard Scaling
Outlines transitions to alternative computing models to sustain AI progress.
1:17Explained48No Free Lunch Theorem
No universal algorithm exists; domain knowledge improves search efficiency.
1:33Explained49Overview of Evolutionary Algorithms
EAs are population-based metaheuristics evolving solutions via selection and variation.
1:22Explained50Fitness and Objectives in EAs
Fitness evaluation drives selection; multi-objective optimization seeks Pareto-optimal fronts.
1:50Explained51Introduction
Introduces when EAs outperform traditional approaches.
1:42Explained52Figure 2.1
Links hardware progress with AI algorithmic capabilities and unknown regions.
1:35Explained53End of Dennard Scaling
Signals a move to new hardware paradigms and deep learning resurgence.
1:02Explained54No Free Lunch Theorem
No single algorithm is universally best; domain knowledge improves efficiency.
1:07Explained55Evolutionary Algorithms Overview
Overview of evolutionary algorithms and their metaheuristic nature.
1:26Explained56Fitness and Objectives in EAs
Fitness functions guide evolution; multi-objective optimization navigates trade-offs.
1:24Explained57Introduction
Discussion of when EAs are advantageous over traditional methods.
1:52Explained58Figure 2.1
Illustrates the relationship between hardware improvements and algorithmic progress.
1:15Explained59End of Dennard Scaling
End of scaling drives exploration of new computing models.
1:26Explained60No Free Lunch Theorem
No universal algorithm exists; domain knowledge improves efficiency.
1:27Explained61Overview of Evolutionary Algorithms
EA families evolve solutions via population-based processes.
1:19Explained62Fitness and Objectives in EAs
Fitness measures success; multi-objective optimization seeks Pareto frontiers.
1:14Explained63Introduction
EAs solve complex, high-dimensional problems where other methods fail.
1:35Explained64Figure 2.1
Shows hardware and software progress and the unknown frontiers.
1:42Explained65End of Dennard Scaling
Describes a shift to advanced hardware and AI-enabled systems.
1:10Explained66No Free Lunch Theorem
Domain-specific knowledge is essential for efficient search.
1:16Explained67Evolutionary Algorithms Overview
Overview of population-based metaheuristics.
1:21Explained68Fitness and Objectives in EAs
Fitness guides evolution; multi-objective optimization balances several goals.
1:52Explained69Introduction
Explains why EAs are used for challenging problems.
1:08Explained70Figure 2.1
Depicts hardware progress enabling new problem-solving capabilities.
1:02Explained71End of Dennard Scaling
Describes the move to novel hardware models to support AI.
1:26Explained72No Free Lunch Theorem
No universal algorithm exists; domain knowledge drives efficiency.
1:17Explained73Overview of Evolutionary Algorithms
EA families such as GA, GP, GE, CGP, and PushGP evolve programs and designs.
1:26Explained74Fitness and Objectives in EAs
Fitness calculations drive selection; multi-objective may yield Pareto-optimal sets.
1:10Explained75Introduction
Introduces the role of EAs in solving complex problems.
1:31Explained76Figure 2.1
Relates hardware capability to algorithmic performance and unknown regions.
1:22Explained77End of Dennard Scaling
Discusses shifts to new computing paradigms and AI advances.
1:52Explained78No Free Lunch Theorem
No single algorithm is best for all problems; domain knowledge helps.
1:35Explained79Evolutionary Algorithms Overview
Overview of evolutionary algorithms and their metaheuristic nature.
1:19Explained80Fitness and Objectives in EAs
Fitness measures success; multiple objectives require Pareto optimization.
1:13Explained81Introduction
Explains when EAs are advantageous relative to traditional methods.
1:22Explained82Figure 2.1
Illustrates hardware progression and algorithmic capacity growth.
1:12Explained83End of Dennard Scaling
Hardware scaling limits spur new computing approaches for AI.
1:44Explained84No Free Lunch Theorem
There is no universal best algorithm; domain knowledge improves search.
1:02Explained85Overview of Evolutionary Algorithms
Introduction to EAs as population-based metaheuristics.
1:47Explained86Fitness and Objectives in EAs
Fitness guides evolution; multi-objective optimization seeks Pareto-optimal solutions.
1:33Explained87Introduction
Discusses when EAs are favored over traditional optimization methods.
1:58Explained88Figure 2.1
Illustrates the interplay of hardware progress and algorithmic capabilities.
1:41Explained89End of Dennard Scaling
Describes the transition to new hardware paradigms to sustained AI progress.
1:33Explained90No Free Lunch Theorem
No universal best algorithm; domain knowledge enhances efficiency.
1:23Explained91Traditional Techniques Overview
An overview of established EAs used in industry and research.
1:47Explained92Evolutionary Strategy (ES)
ES evolves continuous parameters using mutation and selection, with CMA-ES and CMSA-ES variants.
1:30Explained93Genetic Algorithms (GA)
GA optimizes fixed-length strings representing variables or parameters using crossover and mutation.
1:30Explained94Genetic Programming (GP)
GP evolves executable programs or equations using tree-like representations.
1:34Explained95Genetic Improvement (GI)
GI optimizes existing working code to improve performance or correctness.
1:23Explained96Grammatical Evolution (GE)
GE evolves programs by evolving grammars (BNF) to generate code.
1:29Explained97Linear Genetic Programming (LGP)
LGP uses linear programs for sequential problems and low-level optimizations.
1:45Explained98Cartesian Genetic Programming (CGP)
CGP uses Cartesian graphs with small populations to solve structured problems.
1:34Explained99Differential Evolution (DE)
DE optimizes by weighted differences and self-organizes populations.
1:40Explained100Gene Expression Programming (GEP)
GEP uses fixed-length strings encoding expression trees to generate valid programs.
1:19Explained101Specialized Techniques
Covers exotic and hybrid EAs beyond traditional methods.
1:41Explained102Auto-constructive Evolution
Entities evolve themselves without a central controller to form offspring.
0:57Explained103Neuroevolution
Uses genetic algorithms to optimize neural networks and architectures.
0:32Explained