Call for papers

Self-Improving Systems: The Convergence of Evolutionary Strategies and Machine Intelligence using XAI and LLMs

IEEE International Conference on Evolving and Adaptive Intelligent Systems 2026 

IEEE EAIS 2026

September 21 - 23, University of Pisa, Pisa, Italy 

Aim and scope 

The primary aim of this special session is to catalyze a fundamental paradigm shift in artificial intelligence by pioneering the deep integration of Evolutionary Algorithms (EAs), Machine Learning (ML), Explainable AI (XAI), and Large Language Models (LLMs). We seek to move beyond static, single-purpose AI models toward the creation of truly self-improving systems; adaptive, resilient, and transparent computational entities capable of continuous learning and autonomous optimization in complex, dynamic environments. The ultimate objective is to create systems that not only solve complex problems but also understand, explain, and autonomously refine their own problem-solving processes, thereby evolving into genuine collaborative partners in scientific discovery and engineering innovation.

The scope of this session is resolutely interdisciplinary, welcoming contributions across theoretical, algorithmic, and applied domains. We specifically encourage work that demonstrates how this convergence creates emergent capabilities that surpass individual component technologies. Relevant topics encompass, but are not limited to: Architectures for Self-Improvement, where EAs and neural networks form closed-loop mutual enhancement cycles; LLMs as Evolutionary Components, employing large models for intelligent operations in code synthesis and strategy generation; XAI for Guiding Evolution, using interpretability to open black-box processes and steer search toward transparent solutions; Automated Discovery applications in science, AutoML, and robotics; Causal Evolution & Symbolic Regression for uncovering underlying data structures; and critical studies on Trust, Safety, and Governance alongside Theoretical & Empirical Analyses of hybrid system properties. This forum aims to unite researchers from evolutionary computation, machine learning, NLP, and XAI to chart the course toward autonomous, trustworthy self-improving systems collectively.

 Main topics of interest

Authors are invited to submit their original and unpublished work to this special session; the topics of interest (but not limited) are the following:

Architectural & Algorithmic Innovations

• Novel hybrid architectures combining ES with Deep Reinforcement Learning (DRL).

• Frameworks for continuous and lifelong learning inspired by evolutionary processes.

• Co-evolutionary systems where multiple interact and improve each other.

• EAs for hyperparameter tuning and neural architecture search (NAS) of LLMs and other large-scale models.

• Memory-augmented evolutionary algorithms for learning from historical search trajectories.

• Quality-Diversity (QD) algorithms and their integration with generative models.

The Role of LLMs & Generative Models

• Using LLMs as intelligent operators for mutation, crossover, and initial population generation.

• Prompt engineering and optimization through evolutionary algorithms.

• LLMs for fitness function design, approximation, or natural language description.

• Evolutionary fine-tuning of LLMs for specific domains or capabilities.

• Knowledge extraction from LLMs to seed or guide evolutionary search.

• Generating synthetic data or environments for evolutionary training using LLMs.

• Code generation and program synthesis via LLM-driven evolution.

Explainability, Transparency & Trust (XAI)

• XAI methods for interpreting the decision-making of evolved models and policies.

• Visualizing and understanding the fitness landscape and population dynamics.

• Using interpretability metrics as objectives in multi-objective evolutionary optimization.

• Explainable and verifiable reward function shaping in evolved systems.

• Auditing, monitoring, and debugging autonomous learning cycles with XAI.

Applications & Use Cases

• Robotics: Evolving adaptive control policies, morphology, or behavior trees.

• Scientific Discovery: Automated hypothesis generation, experimental design, and analysis in fields like biology, chemistry, and physics.

• Automated Machine Learning (AutoML): End-to-end system design and configuration.

• Game AI & Interactive Entertainment: Procedural content generation, NPC behavior design, and balancing.

• Engineering & Design: Optimizing complex systems, circuits, antennas, and materials.

• Finance & Economics: Evolving trading strategies and forecasting models.

• Supply Chain & Logistics: Dynamic optimization of routing and resource allocation.

Safety, Ethics & Governance

• Algorithmic alignment: Ensuring long-term goals remain stable during self-improvement.

• Constraint handling and guaranteed safe exploration.

• Detecting and mitigating bias and unfairness in autonomously evolved systems.

• Governance models and control mechanisms for potentially high-impact AI systems.

Submission guidelines

Please follow the submission guidelines from the IEEE EAIS 2026 Submission Website. Special session papers are treated the same as regular conference papers. Please specify that your paper is for the Special Session called: Self-Improving Systems: The Convergence of Evolutionary Strategies and Machine Intelligence using XAI and LLMs. All papers accepted and presented at EAIS 2026 will be published on the IEEE Xplore Digital Library.

Important dates

• Paper Submission: March 15, 2026

• Notification of acceptance: May 15, 2026

• Camera Ready Submissions: June 15, 2026

• Author Registration: June 30, 2026

• Conference Dates: 21-23 September 2026

To get more information, access to the next url: https://ai.dii.unipi.it/eais2026/

Organizers

• Dr. Diego Oliva

Universidad de Guadalajara, México

 diego.oliva@cucei.udg.mx

• Dr. Jorge Ramos-Frutos

Tecnológico Nacional de México Jiquilpan, México

jorgearmando.rf@jiquilpan.tecnm.mx