English
Related papers

Related papers: Momentum Accelerates Evolutionary Dynamics

200 papers

Biodiversity and extinction are central issues in evolution. Dynamical balance among different species in ecosystems is often described by deterministic replicator equations with moderate success. However, fluctuations are inevitable,…

Populations and Evolution · Quantitative Biology 2014-11-25 Tsung-Cheng Lu , Yi-Ko Chen , Hsiu-Hau Lin , Chun-Chung-Chen

Evolution is a fundamental process that shapes the biological world we inhabit, and reinforcement learning is a powerful tool used in artificial intelligence to develop intelligent agents that learn from their environment. In recent years,…

Neural and Evolutionary Computing · Computer Science 2023-06-19 Taboubi Ahmed

Motivated by present activities in (statistical) physics directed towards biological evolution, we review the interplay of three evolutionary forces: mutation, selection, and genetic drift. The review addresses itself to physicists and…

Statistical Mechanics · Physics 2016-11-23 Ellen Baake , Wilfried Gabriel

We use analytical techniques based on an expansion in the inverse system size to study the stochastic evolutionary dynamics of finite populations of players interacting in a repeated prisoner's dilemma game. We show that a mechanism of…

Populations and Evolution · Quantitative Biology 2012-04-20 Alex J. Bladon , Tobias Galla , Alan J. McKane

The main topic of this thesis is the analysis of evolution equations reflecting issues in ecology and population dynamics. In mathematical modelling, the impact of environmental elements and the interaction between species is read into the…

Analysis of PDEs · Mathematics 2021-03-08 Elisa Affili

We investigate a model of evolutionary dynamics on a smooth landscape which features a ``mutator'' allele whose effect is to increase the mutation rate. We show that the expected proportion of mutators far from equilibrium, when the fitness…

patt-sol · Physics 2009-10-30 David A. Kessler , Herbert Levine

Momentum method has been used extensively in optimizers for deep learning. Recent studies show that distributed training through K-step averaging has many nice properties. We propose a momentum method for such model averaging approaches. At…

Machine Learning · Computer Science 2021-10-05 Guojing Cong , Tianyi Liu

Evolutionary biology shares many concepts with statistical physics: both deal with populations, whether of molecules or organisms, and both seek to simplify evolution in very many dimensions. Often, methodologies have undergone parallel and…

Populations and Evolution · Quantitative Biology 2015-03-19 Harold P. de Vladar , Nick H. Barton

Momentum has become a crucial component in deep learning optimizers, necessitating a comprehensive understanding of when and why it accelerates stochastic gradient descent (SGD). To address the question of ''when'', we establish a…

Machine Learning · Computer Science 2023-06-16 Jingwen Fu , Bohan Wang , Huishuai Zhang , Zhizheng Zhang , Wei Chen , Nanning Zheng

We show how to calculate the likelihood of dynamical large deviations using evolutionary reinforcement learning. An agent, a stochastic model, propagates a continuous-time Monte Carlo trajectory and receives a reward conditioned upon the…

Statistical Mechanics · Physics 2020-08-26 Stephen Whitelam , Daniel Jacobson , Isaac Tamblyn

To approximate convolutions which occur in evolution equations with memory terms, a variable-stepsize algorithm is presented for which advancing N steps requires only O(N log(N)) operations and O(log(N)) active memory, in place of O(N^2)…

Numerical Analysis · Mathematics 2007-05-23 María López-Fernández , Christian Lubich , Achim Schädle

Under constant selection, each trait has a fixed fitness, and small mutation rates allow populations to efficiently exploit the optimal trait. Therefore it is reasonable to expect mutation rates will evolve downwards. However, we find this…

Populations and Evolution · Quantitative Biology 2022-08-23 Brian Mintz , Feng Fu

Is evolution always gradual or can it make leaps? We examine a mathematical model of an evolutionary process on a fitness landscape and obtain analytic solutions for the probability of multi-mutation leaps, that is, several mutations…

Populations and Evolution · Quantitative Biology 2022-10-12 Mikhail I. Katsnelson , Yuri I. Wolf , Eugene V. Koonin

Migration is a key mechanism for expansion of communities. In spatially heterogeneous environments, rapidly gaining knowledge about the local environment is key to the evolutionary success of a migrating population. For historical human…

Populations and Evolution · Quantitative Biology 2016-12-07 Dong Wang , Michael W. Deem

Population dynamics on a rugged landscape is studied analytically and numerically within a simple discrete model for evolution of N individuals in one-dimensional fitness space. We reduce the set of master equations to a single Fokker-Plank…

adap-org · Physics 2015-06-30 Igor Aranson , Lev Tsimring , Valerii Vinokur

Understanding the evolutionary dynamics of reinforcement learning under multi-agent settings has long remained an open problem. While previous works primarily focus on 2-player games, we consider population games, which model the strategic…

Multiagent Systems · Computer Science 2020-06-30 Shuyue Hu , Chin-Wing Leung , Ho-fung Leung , Harold Soh

A new approach to understanding evolution [Val09], namely viewing it through the lens of computation, has already started yielding new insights, e.g., natural selection under sexual reproduction can be interpreted as the Multiplicative…

Populations and Evolution · Quantitative Biology 2016-04-26 Ruta Mehta , Ioannis Panageas , Georgios Piliouras , Prasad Tetali , Vijay V. Vazirani

The training of deep residual neural networks (ResNets) with backpropagation has a memory cost that increases linearly with respect to the depth of the network. A way to circumvent this issue is to use reversible architectures. In this…

Machine Learning · Computer Science 2021-07-23 Michael E. Sander , Pierre Ablin , Mathieu Blondel , Gabriel Peyré

We study effects of strategy-dependent time delays on equilibria of evolving populations. It is well known that time delays may cause oscillations in dynamical systems. Here we report a novel behavior. We show that microscopic models of…

Populations and Evolution · Quantitative Biology 2021-04-13 Jacek Miȩkisz , Marek Bodnar

A simple model of macroevolution is proposed exhibiting both the property of punctuated equilibrium and the dynamics of potentialities for different species to evolve towards increasingly higher complexity. It is based on the phenomenon of…

Biological Physics · Physics 2007-05-23 Siegfried Fussy , Gerhard Groessing , Herbert Schwabl