Tutorial on
Evolutionary Computation, Tunneling and Connections to Quantum Computing
Instructor
|
Darrell Whitley
Colorado State University
United States
|
|
|
|
Brief Bio
Prof. Darrell Whitley has been active in Evolutionary Computation since 1986, and has published more than 250 papers. These papers have garnered more than 33,000 citations. Dr. Whitley’s H-index is 73. He introduced the first “steady state genetic algorithm” with rank based selection, and has worked on dozens of real world applications of evolutionary algorithms. He has served as Editor-in-Chief of the journal Evolutionary Computation, and served as Chair of the Governing Board of ACM Sigevo from 2007 to 2011. He is a Fellow of the ACM recognized for his contributions to Evolutionary Computation, and he was awarded the 2022 IEEE PIONEER Award in Evolutionary Computation.
|
Abstract
This tutorial is mainly about Evolutionary Algorithms, but it will highlight ways in which Evolutionary Algorithms and Quantum Computing share the ability to efficiently tunnel between local optima and to exploit problem representations with low nonlinearity. New theoretical results offer new insights to explain why “Partition Crossover” is so successful at tunneling between local optima on NP Hard problems such as MAX-kSAT and the Traveling Salesman Problem. Partition Crossover can also be applied to Quadratic Unconstrained Boolean Optimization (QUBO) problems, which exploit a keystone representation in Quantum Optimization. The tutorial will briefly review key basic concepts from Quantum Computing. Problem transforms that are used in Quantum Computing to reduce problem nonlinearity also offer critical advantages for Evolutionary Algorithms due to the fact that powerful evolutionary operators can also exploit problem representations with low nonlinearity.
Keywords
Evolutionary Algorithms, tunneling, quantum computing.
Aims and Learning Objectives
This tutorial will look at some of the best Evolutionary Algorithms for MAX-kSAT and the TSP, and explore connections between Evolutionary Algorithms and Quantum Computing.
Target Audience
Anyone interesting in EAs for MAXSAT, TSP, QUBO or NK Landscapes, or who is curious about Quantum Optimization.
Prerequisite Knowledge of Audience
A basic understanding of Evolutionary Algorithms and Genetic Algorithms.
Detailed Outline
1) Quantum Tunneling
2) A basic review of Quantum Gate
3) A fast look at Quantum Fourier methods
4) Quadratic Unconstrained Boolean Optimization
5) Transform and QUBO
6) Deterministic Partition Crossover
7) Deterministic constant time improving moves
8) Fourier and Walsh methods for Quantum and EAs
9) Theoretical results for tunneling in EAs
Tutorial on
LLM-Driven AI Techniques Design and Discovery
Instructor
|
Adam Viktorin
Tomas Bata University in Zlin
Czech Republic
|
|
|
|
Brief Bio
Not Available
|
Abstract
Abstract
Large language models (LLMs) are reshaping the way we invent and automate AI techniques - moving beyond hyperparameter tuning and automated selection into automated design and discovery of algorithms, architectures, and end-to-end pipelines, that close the loop between ideation and evaluation. This tutorial maps the rapidly evolving landscape—covering an overview of frameworks like EASE, LLaMEA, AlphaEvolve, FunSearch, and emerging Gen-AI–driven AI assistants—then contrasts the broader remit of EASE (Effortless Algorithmic Solution Evolution) with the metaheuristic-first orientation of LLaMEA (Large Language Model Evolutionary Algorithm). We place special emphasis on EASE as a practical, fully modular framework for iterative closed-loop generation and evaluation; beyond algorithm code, EASE can also iteratively generate text and graphics. The format of the tutorial is demo-driven (no hands-on): short videos and narrated walkthroughs of the EASE frontend/backend illustrate task setup, fitness/evaluation loops, and result inspection; we also provide overview links to LLaMEA variants, documentation, and benchmarking environment. We highlight guardrails (testing, analysis, time/resource caps), and evaluation/benchmarking practice. Attendees will gain practical criteria for choosing LLaMEA vs EASE, methods for responsible evaluation, and steps for adopting LLM-driven discovery in AI research.
Keywords
LLM-driven design; AutoML; Automated algorithms discovery; evolutionary computation; metaheuristic synthesis; EASE; LLaMEA; program & artifact synthesis
Aims and Learning Objectives
The goal is to equip researchers and practitioners with a conceptual toolbox and concrete workflows to responsibly adopt LLM-driven AI techniques, design, and algorithms in their own work. Participants will:
- Differentiate frameworks: understand LLaMEA (metaheuristic-first discovery) vs EASE (broader AI technique & artifact generation), including typical problem scopes and trade-offs.
- Understand EASE and LLaMEA workflows.
- Specify discovery tasks with measurable objectives and guardrails.
- Benchmark responsibly.
Target Audience
AI/EC/ML researchers and practitioners aiming to automate algorithm or pipeline discovery and compare frameworks
AI and data scientists exploring LLM-in-the-loop workflows for optimization and model/algorithm design.
Prerequisite Knowledge of Audience
Conceptual: basic ML and CI/metaheuristics, plus fundamentals of (usage of) LLMs.
Detailed Outline
- Introduction & landscape – why automated algorithm discovery matters
- Core paradigms: FunSearch, LLaMEA, ReEvo, AlphaEvolve, EASE
- Deep dive into EASE architecture
- EASE: frontend/backend architecture + live examples
- LLaMEA architecture and variants: LLaMEA-HPO & LLaMEA-BO, BLADE
- Evaluation & safety, benchmarks
- Frontiers & open challenges (bias, compute cost, integration)
- Open Q&A
Tutorial on
Trustworthy Autonomous Systems: Challenges and Enablers
Instructor
|
Joanna Isabelle Olszewska
University of the West of Scotland
United Kingdom
|
|
|
|
Brief Bio
Dr Joanna Olszewska BSc(Hons) MSc(EPFL) PhD(UCL) CEng CSci FBCS FHEA is a British Computer Scientist. She is an Asst. Professor with UWS, UK, and leads research in Algorithms and Softwares for Trustworthy Intelligent Vision Systems. Senior Member of IEEE, she stands on the IEEE AI Standard Committee and she is part of the IEEE Global Initiative for Ethical Considerations in AI and Autonomous Systems. She is Co-Chair of the IEEE RAS Technical Committee on the Verification of Autonomous Systems. ACM Distinguished Speaker, she has participated in panels about the Future of AI and standardization efforts in robotics and autonomous systems (e.g. at ICRA). She has delivered invited talks (e.g. at the Canadian Mathematical Society, ENS Paris, DDD Scotland Industry forum), keynote talks (e.g. at SBESC), tutorials at conferences (e.g. at ICAART), webinars (e.g. at IEEE) as well as podcasts (e.g. at BCS) and interviews (e.g. at BBC). She has been TPC member of over 100 international conferences (e.g. IJCAI) and has chaired over 70 webinars/seminars/workshops (e.g. IROS). She has served as Technical Program Chair of the IEEE International Conference on Engineering Reliable Autonomous Systems (ERAS). She has been appointed Guest Editor for the Knowledge Engineering Review Journal, Cambridge University Press, and Associate Editor for the Frontiers in Artificial Intelligence journal, the Engineering Applications of Artificial Intelligence journal, and the Machine Learning with Applications journal, Elsevier. Fellow of the British Computer Society, Chartered Engineer, Chartered Scientist, and Fellow of StandICT.eu, she has contributed to 20+ ISO/IEC/IEEE standards in various roles, e.g. Vice-Chair of ISO/IEC/IEEE 41062. She has published 150+ papers and one book – ‘Artificial Intelligence and Software Testing: Building systems you can trust’ – winner of the Independent Press Award.
|
Abstract
Abstract
With Artificial Intelligence (AI) technologies becoming ubiquitous in people’s daily life, trustworthiness in these intelligent and autonomous systems is crucial. For this purpose, this tutorial will outline the major challenges of trustworthy autonomous systems, while it will highlight the enablers to develop and deploy such systems. In particular, the tutorial will cover methods and guidelines for designing, developing, testing and verifying autonomous systems. Furthermore, this tutorial will present IEEE and ISO/IEC/IEEE standards for the design and development of autonomous systems so they meet not only technical but also societal expectations – leading thus to the trustworthiness in autonomous systems.
Keywords
Trustworthy Autonomous Systems (TAS), Explainable Artificial Intelligence (XAI), Software and System Reliability, Artificial Intelligence Testing, Verification of Autonomous Systems (VAS), Artificial Intelligence (AI) and Robotics Standards, Ontologies Supporting Ethical AI, Autonomous Systems, Intelligent Systems, Knowledge-Based Systems, Decision-Support Systems, Human-AI Collaboration, Human-Robot Interaction, Robotic Systems, Intelligent Vision Systems.
Aims and Learning Objectives
1. The tutorial aims to present technical approaches to develop trustworthy autonomous systems.
2. The tutorial aims to increase awareness about the social, ethical and legal aspects as well as the standardization efforts in the context of the development and deployment of trustworthy autonomous systems.
3. The tutorial aims to provide real-world case studies and examples of developed and deployed trustworthy autonomous systems.
Target Audience
Stakeholders from Academia and Industry as well as regulators from Governmental and Professional Bodies
Prerequisite Knowledge of Audience
Suitable for students and scholars of different disciplines who are interested in the study, analysis, design, modelling and implementation of interpretable and explainable AI systems
Detailed Outline
- Introduction of Trustworthy Autonomous Systems:
-- Definition of Trustworthy Autonomous Systems
-- Technical and Societal Challenges of Trustworthy Autonomous Systems
- Key Enablers of Trustworthy Autonomous Systems:
-- Processes and Techniques to Design, Test, Verify and Assess Trustworthy Autonomous Systems
-- Guidelines and Standards for AI, Robotics, Autonomous and Intelligent Systems (AIS)
- Applications of Trustworthy Autonomous Systems in context of smart manufacturing and smart cities