OSI Lab Seminars
[previously] The ``RLHF Study Group’’ has been dissolved on 2025.01.05.
[previously] OSI Lab has divided the global seminar into several divisions based on the topics. I was in charge of the Theory division til 2024.06.11, when it dissolved and converted to ``RLHF Study Group.''
[previously] OSI Lab (led by Prof. Se-Young Yun) holds a weekly seminar where each of 2 members, whose orders are assigned based on a fixed circular order, presents a paper of his/her own interest. The seminar is called off only when it overlaps with some major conference or exam period.
- Bandits 101: A Maximally Non-technical Tutorial
- Flooding with Absorption: An Efficient Protocol for Heterogeneous Bandits over Complex Networks
- Conference Day - Theory Division
- Improved Regret Bounds of (Multinomial) Logistic Bandits via Regret-to-Confidence-Set Conversion
- Introduction to Reinforcement Learning with Human Feedback (RLHF): A Theoretically Biased Overview
- Community Detection in Block Models: From SBMs to Block Markov Chains
- A Primer on (Combinatorial Semi-) Bandits
- Fair Streaming Principal Component Analysis: Statistical and Algorithmic Viewpoint
- Improved Sample Complexity for Reward-free Reinforcement Learning under Low-rank MDPs
- Exact Dynamics of Stochastic Gradient Descent in High Dimensions and Volterra (Integral) Equations
- Entropic variants of SGD
- Conference Week (AISATS & ICML 2022)
- From Generalized Linear Bandit to Logistic Bandit: An Overview
- Gradient Descent on Infinitely Wide Neural Networks: Global Convergence and Generalization
- Clustering in Block Markov Chains
- Fast and Efficient MMD-based Fair PCA via Optimization over Stiefel Manifold
- Poisson Learning: Graph Based Semi-Supervised Learning At Very Low Label Rates
- Landscape and training regimes in deep learning
- How Powerful are Graph Neural Networks?
- Conference Week (NeurIPS 2020)
- Heavy-tail behaviour of SGD - Part 2
- Heavy-tail behaviour of SGD - Part 1