Researchers from Carnegie Mellon University and Princeton University have proposed a new method called Distance Aware Bottleneck (DAB) for quantifying uncertainty in deep learning. Other recent papers focus on topics such as multi-objective Bayesian optimization, online algorithms for optimizing performance metrics, and control-theoretic approaches for reinforcement learning.
Distributionally Robust Constrained Reinforcement Learning under Strong Duality. https://t.co/U2xy2zHL1C
Open Problem: Order Optimal Regret Bounds for Kernel-Based Reinforcement Learning. https://t.co/ImDwQt9Oyr
A Provably Efficient Option-Based Algorithm for both High-Level and Low-Level Learning. https://t.co/k7c7qVFihb
GOAL: A Generalist Combinatorial Optimization Agent Learner. https://t.co/IdZQqePyRR
Learning to Cover: Online Learning and Optimization with Irreversible Decisions. https://t.co/SyiPDSO6gZ
A General Control-Theoretic Approach for Reinforcement Learning: Theory and Algorithms. https://t.co/RXxTsNBdwf
Stochastic Re-weighted Gradient Descent via Distributionally Robust Optimization https://t.co/zVDUX40Pnb #imagenet #sgd #benchmarks
A General Online Algorithm for Optimizing Complex Performance Metrics. https://t.co/5clrUfgmfe
Preferential Multi-Objective Bayesian Optimization. https://t.co/ZfCd9NxyMq
Preferential Multi-Objective Bayesian Optimization https://t.co/8v6o9UVFYn
[LG] A Rate-Distortion View of Uncertainty Quantification I Apostolopoulou, B Eysenbach, F Nielsen, A Dubrawski [CMU & Princeton University] (2024) https://t.co/FuKwV5o3ME - The paper proposes Distance Aware Bottleneck (DAB), a new method for quantifying uncertainty in deep… https://t.co/aLTF4BTVEr