🏢 EPFL, Switzerland
Randomized algorithms and PAC bounds for inverse reinforcement learning in continuous spaces
·1647 words·8 mins·
loading
·
loading
Machine Learning
Reinforcement Learning
🏢 EPFL, Switzerland
This paper presents randomized algorithms with PAC bounds for solving inverse reinforcement learning problems in continuous state and action spaces, offering robust theoretical guarantees and practica…
Fast Proxy Experiment Design for Causal Effect Identification
·2057 words·10 mins·
loading
·
loading
AI Theory
Causality
🏢 EPFL, Switzerland
This paper presents efficient algorithms for designing cost-optimal proxy experiments to identify causal effects, significantly improving upon prior methods.
Analyzing & Reducing the Need for Learning Rate Warmup in GPT Training
·2347 words·12 mins·
loading
·
loading
Natural Language Processing
Large Language Models
🏢 EPFL, Switzerland
This study reveals that modifying optimizers to normalize updates based on angular changes and gradient signal-to-noise ratio significantly reduces the need for learning rate warmup in GPT training.