English
Related papers

Related papers: Evolutionary Stability Against Multiple Mutations

200 papers

Robustness of deep learning models is a property that has recently gained increasing attention. We explore a notion of robustness for generative adversarial models that is pertinent to their internal interactive structure, and show that,…

Machine Learning · Computer Science 2019-10-11 Zhi Xu , Chengtao Li , Stefanie Jegelka

The study of adversarial robustness has so far largely focused on perturbations bound in p-norms. However, state-of-the-art models turn out to be also vulnerable to other, more natural classes of perturbations such as translations and…

Machine Learning · Computer Science 2019-09-17 Logan Engstrom , Brandon Tran , Dimitris Tsipras , Ludwig Schmidt , Aleksander Madry

We investigate the stability against small deformations of strings dangling into AdS_5-Schwarzschild from a moving heavy quark-anti-quark pair. We speculate that emission of massive string states may be an important part of the evolution of…

High Energy Physics - Theory · Physics 2010-10-27 Joshua J. Friess , Steven S. Gubser , Georgios Michalogiorgakis , Silviu S. Pufu

We show that there may exist an inherent tension between the goal of adversarial robustness and that of standard generalization. Specifically, training robust models may not only be more resource-consuming, but also lead to a reduction of…

Machine Learning · Statistics 2019-09-10 Dimitris Tsipras , Shibani Santurkar , Logan Engstrom , Alexander Turner , Aleksander Madry

Complex networks often have a modular structure, where a number of tightly- connected groups of nodes (modules) have relatively few interconnections. Modularity had been shown to have an important effect on the evolution and stability of…

Physics and Society · Physics 2014-04-21 Saray Shai , Dror Y. Kenett , Yoed N. Kenett , Miriam Faust , Simon Dobson , Shlomo Havlin

Conditions for Bayesian posterior robustness have been examined in recent literature. However, many of the proofs seem to be long and complicated. In this paper, we first summarize some basic lemmas that have been applied implicitly or…

Statistics Theory · Mathematics 2023-11-06 Yasuyuki Hamura

This work tackles an intriguing and fundamental open challenge in representation learning: Given a well-trained deep learning model, can it be reprogrammed to enhance its robustness against adversarial or noisy input perturbations without…

Machine Learning · Computer Science 2024-10-08 Zhichao Hou , MohamadAli Torkamani , Hamid Krim , Xiaorui Liu

The present article introduces a reference framework for discussing resilience of computational systems. Rather than a property that may or may not be exhibited by a system, resilience is interpreted here as the emerging result of a dynamic…

Systems and Control · Computer Science 2015-04-13 Vincenzo De Florio

In environments that vary frequently and unpredictably, bet-hedgers can overtake the population. Diversifying bet-hedgers have a diverse set of offspring so that, no matter the conditions they find themselves in, at least some offspring…

Populations and Evolution · Quantitative Biology 2023-07-06 Csenge Petak , Lapo Frati , Melissa H. Pespeni , Nick Cheney

Counterfactual explanations (CEs) are advocated as being ideally suited to providing algorithmic recourse for subjects affected by the predictions of machine learning models. While CEs can be beneficial to affected individuals, recent work…

Machine Learning · Computer Science 2024-02-06 Junqi Jiang , Francesco Leofante , Antonio Rago , Francesca Toni

Genetic robustness, the preservation of evolved phenotypes against genotypic mutations, is one of the central concepts in evolution. In recent years a large body of work has focused on the origins, mechanisms, and consequences of robustness…

Populations and Evolution · Quantitative Biology 2018-01-17 Qijun He , Fenix W. Huang , Christopher Barrett , Christian M. Reidys

A method for constructing evolution equations admitting a master symmetry is proposed. Several examples illustrating the method are presented. It is also noted that for certain evolution equations master symmetries can be useful for…

Exactly Solvable and Integrable Systems · Physics 2009-11-07 F. Finkel , A. S. Fokas

Evolution Strategies (ES) emerged as a scalable alternative to popular Reinforcement Learning (RL) techniques, providing an almost perfect speedup when distributed across hundreds of CPU cores thanks to a reduced communication overhead.…

Machine Learning · Statistics 2018-11-13 Víctor Campos , Xavier Giro-i-Nieto , Jordi Torres

For a non-linear MIMO feedback system, the robustness against uncertain time-variations in the feedback loop is investigated in an input-output framework. A general sufficient condition in terms of a bound on average rates of time-variation…

Systems and Control · Electrical Eng. & Systems 2020-03-03 Yu-chen Sung , Sagar V. Patil , Michael G. Safonov

In this paper, we study analytically the statistics of the number of equilibria in pairwise social dilemma evolutionary games with mutation where a game's payoff entries are random variables. Using the replicator-mutator equations, we…

Populations and Evolution · Quantitative Biology 2021-09-15 Manh Hong Duong , The Anh Han

We tackle here a specific, still not widely addressed aspect, of AI robustness, which consists of seeking invariance / insensitivity of model performance to hidden factors of variations in the data. Towards this end, we employ a two step…

Machine Learning · Computer Science 2022-03-04 William Paul , Philippe Burlina

Super-stability and strong stability are properties of a matching in the stable matching problem with ties. In this paper, we introduce a common generalization of super-stability and strong stability, which we call non-uniform stability.…

Computer Science and Game Theory · Computer Science 2024-08-30 Naoyuki Kamiyama

We present a new, scalable alternative to the structured singular value, which we call $\nu$, provide a convex upper bound, study their properties and compare them to $\ell_1$ robust control. The analysis relies on a novel result on the…

Optimization and Control · Mathematics 2022-04-13 Olle Kjellqvist , John C. Doyle

Improving model robustness against potential modality noise, as an essential step for adapting multimodal models to real-world applications, has received increasing attention among researchers. For Multimodal Sentiment Analysis (MSA), there…

Multimedia · Computer Science 2022-11-28 Huisheng Mao , Baozheng Zhang , Hua Xu , Ziqi Yuan , Yihe Liu

The mutation rate of a well adapted population is prone to reduction so as to have a lower mutational load. We aim to understand the role of epistatic interactions between the fitness affecting mutations in this process. Using a multitype…

Populations and Evolution · Quantitative Biology 2016-09-20 Ananthu James