Preparing for Black Swans: The Antifragility Imperative for Machine Learning
Not on the map. Why?
Abstract (arXiv)
Operating safely and reliably despite continual distribution shifts is vital for high-stakes machine learning applications. This paper builds upon the transformative concept of ``antifragility'' introduced by (Taleb, 2014) as a constructive design paradigm to not just withstand but benefit from volatility. We formally define antifragility in the context of online decision making as dynamic regret's strictly concave response to environmental variability, revealing limitations of current approaches focused on resisting rather than benefiting from nonstationarity. Our contribution lies in proposing potential computational pathways for engineering antifragility, grounding the concept in online learning theory and drawing connections to recent advancements in areas such as meta-learning, safe exploration, continual learning, multi-objective/quality-diversity optimization, and foundation models. By identifying promising mechanisms and future research directions, we aim to put antifragility on a rigorous theoretical foundation in machine learning. We further emphasize the need for clear guidelines, risk assessment frameworks, and interdisciplinary collaboration to ensure responsible application.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 83 of its 138 references that are in the map. First posted on arXiv 18 May 2024.
Builds on
83 of 138 references.
References · 83
- 2020Language Models are Few-Shot Learnerscited by 32,199
- 2021On the Opportunities and Risks of Foundation Modelscited by 3,961
- 2019Benchmarking Neural Network Robustness to Common Corruptions and Perturbationscited by 2,857
- 2023Retrieval-Augmented Generation for Large Language Models: A Surveycited by 2,185
- 2020The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalizationcited by 1,842
- 2020WILDS: A Benchmark of in-the-Wild Distribution Shiftscited by 1,366
- 2020Gradient Surgery for Multi-Task Learningcited by 1,210
- 2020Meta-Learning in Neural Networks: A Surveycited by 944
- 2023A Comprehensive Survey of Continual Learning: Theory, Method and Applicationcited by 860
- 2022A Generalist Agentcited by 802
- 2019Introduction to Online Convex Optimizationcited by 793
- 2014Neural Turing Machinescited by 630
38 more
- 2016RL2: Fast Reinforcement Learning via Slow Reinforcement Learningcited by 490
- 2021Conflict-Averse Gradient Descent for Multi-task Learningcited by 441
- 2019Online Learning: A Modern Introduction Using Convex Optimizationcited by 407
- 2016Learning to reinforcement learncited by 369
- 2017Safe Model-based Reinforcement Learning with Stability Guaranteescited by 347
- 2021Safe Learning in Robotics: From Learning-Based Control to Safe Reinforcement Learningcited by 337
- 2019Control Barrier Functions: Theory and Applicationscited by 332
- 2021Unsolved Problems in ML Safetycited by 291
- 2020Towards Continual Reinforcement Learning: A Review and Perspectivescited by 268
- 2013(More) Efficient Reinforcement Learning via Posterior Samplingcited by 233
- 2015Contextual Markov Decision Processescited by 213
- 2014Robots that can adapt like animalscited by 203
- 2018Online Learning: A Comprehensive Surveycited by 203
- 2017Hamilton-Jacobi Reachability: A Brief Overview and Recent Advancescited by 186
- 2022In-context Reinforcement Learning with Algorithm Distillationcited by 186
- 2019Adaptive Gradient-Based Meta-Learning Methodscited by 183
- 2013Non-stationary Stochastic Optimizationcited by 173
- 2019Provably efficient RL with Rich Observations via Latent State Decodingcited by 169
- 2022A Review of Safe Reinforcement Learning: Methods, Theory and Applicationscited by 169
- 2019A Generalized Algorithm for Multi-Objective Reinforcement Learning and Policy Adaptationcited by 166
- 2022Safe Control with Learned Certificates: A Survey of Neural Lyapunov, Barrier, and Contraction methodscited by 148
- 2023Supervised Pretraining Can Learn In-Context Reinforcement Learningcited by 137
- 2023Human-Timescale Adaptation in an Open-Ended Task Spacecited by 136
- 2018Adaptive Online Learning in Dynamic Environmentscited by 130
- 2016Matrix Completion has No Spurious Local Minimumcited by 124
- 2020Learning the Pareto Front with Hypernetworkscited by 122
- 2020Learning to be Safe: Deep RL with a Safety Criticcited by 109
- 2020Conservative Safety Critics for Explorationcited by 107
- 2022Discovered Policy Optimisationcited by 103
- 2023A Tutorial on Meta-Reinforcement Learningcited by 102
- 2015Online Optimization : Competing with Dynamic Comparatorscited by 99
- 2017Leave no Trace: Learning to Reset for Safe and Autonomous Reinforcement Learningcited by 86
- 2018Evolved Policy Gradientscited by 86
- 2023Retrieving Multimodal Information for Augmented Generation: A Surveycited by 84
- 2022A Theoretical Study on Solving Continual Learningcited by 78
- 2021RL for Latent MDPs: Regret Guarantees and a Lower Boundcited by 69
- 2019Meta-learning of Sequential Strategiescited by 68
- 2022Safe Reinforcement Learning by Imagining the Near Futurecited by 67
50 most cited of 83 All 83 (JSON)
Built on by
Citing papers per year: 2024: 1.
1 paper.