Publications (93)
Realizable Continuous-Space Shields for Safe Reinforcement Learning
Kyungmin Kim, Davide Corsi, Andoni Rodriguez +5
While Deep Reinforcement Learning (DRL) has achieved remarkable success across various domains, it remains vulnerable to occasional catastrophic failures without additional safegua…
Neural Erosion: Emulating Controlled Neurodegeneration and Aging in AI Systems
Antonios Alexos, Yu-Dai Tsai, Ian Domingo +2
Creating controlled methods to simulate neurodegeneration in artificial intelligence (AI) is crucial for applications that emulate brain function decline and cognitive disorders. W…
A theory of capacity and sparse neural encoding
Pierre Baldi, Roman Vershynin
Motivated by biological considerations, we study sparse neural maps from an input layer to a target layer with sparse activity, and specifically the problem of storing input-ta…
Memorization: A Close Look at Books
Iris Ma, Ian Domingo, Alberto Krone-Martins +2
To what extent can entire books be extracted from LLMs? Using the Llama 3 70B family of models, and the "prefix-prompting" extraction technique, we were able to auto-regressively r…
Learning in the Machine: To Share or Not to Share?
Jordan Ott, Erik Linstead, Nicholas LaHaye +1
Weight-sharing is one of the pillars behind Convolutional Neural Networks and their successes. However, in physical neural systems such as the brain, weight-sharing is implausible.…
Reconstruction of Unstable Heavy Particles Using Deep Symmetry-Preserving Attention Networks
Michael James Fenton, Alexander Shmakov, Hideki Okawa +5
Reconstructing unstable heavy particles requires sophisticated techniques to sift through the large number of possible permutations for assignment of detector objects to the underl…