← Back to all news
arXiv Machine Learning

On Sex, Evolution, and the Multiplicative Weights Update Algorithm

The paper revisits a theory linking sexual reproduction to the multiplicative weights update algorithm (MWUA), originally proposed by Chastain et al. It removes the earlier assumptions of weak selection and product distribution, showing that marginal allele distributions from population dynamics match those produced by MWUA in a general setting. The authors also revise conclusions about convergence and fitness guarantees in coordination games, arguing that sexual dynamics do not impose additional properties on the population distribution beyond convergence.

Related stories

arXiv Machine Learning
Oct 5

Operator Calculus for Population-Based Optimization: Modular Convergence and Finite-Population Guarantees

The paper introduces an operator calculus for population-based optimization, treating each update rule—such as mutation, selection, or recombination—as an operator. By establishing explicit regularity and small‑step conditions, the authors show that the leading effects of these operators add, allowing convergence guarantees to be reused as modular building blocks. The framework differentiates between finding and retaining good solutions, reducing the population’s mean objective, and concentrating candidates near an optimizer, and it identifies additional approximation conditions required for finite‑budget guarantees. Applications to distribution adaptation, recombinative evolution, and consensus dynamics are verified, including nonconvex cases, and controlled experiments illustrate how component effects vary with population geometry and performance measures.

By Pekka Malo, Lauri Viitasaari, Patrik Nummi, Antti Suominen, Ankur Sinha, Olli Tahvonen
More like this →
arXiv AI
Sep 28

Genetic Algorithms with Optimization Guided Operators

The paper introduces a new framework for genetic algorithms where mutation and recombination operators are guided by machine‑learning optimization rather than random changes. It shows that such operators can improve objective values but at higher computational cost, and demonstrates three key phenomena: the necessity of solution‑pool diversity for parity learning, the simultaneous need for generation, mutation, and recombination to achieve near‑optimal solutions, and a phase transition in Gaussian settings where positive drift yields exponential speedup.

By Anna Brandenberger, Ilan Doron-Arad, Elchanan Mossel
reinforcement-learning
More like this →
arXiv Machine Learning
Sep 17

The evolution of sex for artificial intelligence: a population-genetic framework for multigenerational model populations

The paper draws a parallel between AI model development and population genetics, treating successive model generations as analogous to sexual and asexual reproduction. It demonstrates that training models on peers’ outputs reproduces classic genetic processes such as the Wright–Fisher model, while combining parents’ weights can either cancel or preserve inherited advantages depending on the method. Experiments across recurrent, feedforward, and language models confirm these analogies and reveal architecture‑specific biases, including the Fisher‑Muller effect and reproductive isolation when lineages learn conflicting conventions.

By Giorgio F. Gilestro
llms
More like this →
arXiv Machine Learning
Jun 15

Operator Calculus for Population-Based Optimization: A Mean-Field Convergence Theory

arXiv:2606. 14289v1 Announce Type: cross Abstract: Population-based and distributional optimization methods, from evolution strategies and consensus-based optimization to covariance-matrix adaptation and stochastic gradient methods viewed as distributional dynamics, are widely used for nonconvex or black-box problems, yet their convergence analyses remain fragmented across algorithm-specific techniques.

By Pekka Malo, Lauri Viitasaari, Patrik Nummi, Antti Suominen, Ankur Sinha, Olli Tahvonen
More like this →
arXiv AI
Aug 6

Beyond the Dirac Delta: Mitigating Diversity Collapse in Reinforcement Fine-Tuning for Versatile Image Generation

arXiv:2601. 12401v2 Announce Type: replace-cross Abstract: Reinforcement learning (RL) has emerged as a powerful paradigm for fine-tuning large-scale generative models, such as diffusion and flow models, to align with complex human preferences and user-specified tasks.

By Jinmei Liu, Haoru Li, Zhenhong Sun, Chaofeng Chen, Yatao Bian, Bo Wang, Daoyi Dong, Chunlin Chen, Zhi Wang
llmsdiffusionreinforcement-learningfine-tuningsafety
More like this →
arXiv AI
Sep 24

Evolutionary Stability Does Not Guarantee Learning Accessibility: A Multi-Agent Reinforcement Learning Perspective on Cooperation Emergence

The paper investigates whether evolutionary stability guarantees that learning agents can achieve cooperative outcomes in a multi‑agent setting. Using a three‑agent governance game, the authors compare the evolutionary basin of attraction with learning basins derived from independent Q‑learning, scaled Boltzmann exploration, and SA–EA BQL. They find that while the evolutionary basin covers the entire sampled grid, only ε‑greedy Q‑learning attains a substantial learning basin, whereas the other methods fail to sustain cooperation, highlighting a disconnect between population‑level stability and finite‑sample learning accessibility.

By Yijie Wang
agentsreinforcement-learning
More like this →
arXiv AI
Oct 16

T-GADE: Thermodynamical Generative-AI-Driven Evolution of LLM Artifacts

T‑GADE is an evolutionary computation framework that integrates large language models (LLMs) to generate and evolve structured artifacts. It retains classic population and genetic operators while using LLMs for crossover, mutation, and a self‑repair operator called HoxLM. The framework introduces temperature‑controlled free‑energy selection and occupancy rules, and demonstrates improved performance on an online bin‑packing task compared to a baseline Evolution of Heuristics approach.

By Kyoko Ogawa, Naoki Mori
llms
More like this →