Don't Retrain, Just Rewrite: Countering Adversarial Perturbations by Rewriting Text

no code implementations • 25 May 2023 • Ashim Gupta, Carter Wood Blum, Temma Choji, Yingjie Fei, Shalin Shah, Alakananda Vempala, Vivek Srikumar

For example, on sentiment classification using the SST-2 dataset, our method improves the adversarial accuracy over the best existing defense approach by more than 4% with a smaller decrease in task accuracy (0. 5% vs 2. 5%).

Adversarial Robustness Classification +4

Cascaded Gaps: Towards Gap-Dependent Regret for Risk-Sensitive Reinforcement Learning

no code implementations • 7 Mar 2022 • Yingjie Fei, Ruitu Xu

In this paper, we study gap-dependent regret guarantees for risk-sensitive reinforcement learning based on the entropic risk measure.

reinforcement-learning Reinforcement Learning (RL)

Exponential Bellman Equation and Improved Regret Bounds for Risk-Sensitive Reinforcement Learning

no code implementations • NeurIPS 2021 • Yingjie Fei, Zhuoran Yang, Yudong Chen, Zhaoran Wang

The exponential Bellman equation inspires us to develop a novel analysis of Bellman backup procedures in risk-sensitive RL algorithms, and further motivates the design of a novel exploration mechanism.

reinforcement-learning Reinforcement Learning (RL)

Entropic Risk-Sensitive Reinforcement Learning: A Meta Regret Framework with Function Approximation

no code implementations • 1 Jan 2021 • Yingjie Fei, Zhuoran Yang, Zhaoran Wang

We study risk-sensitive reinforcement learning with the entropic risk measure and function approximation.

Efficient Exploration reinforcement-learning +1

Dynamic Regret of Policy Optimization in Non-stationary Environments

no code implementations • NeurIPS 2020 • Yingjie Fei, Zhuoran Yang, Zhaoran Wang, Qiaomin Xie

We consider reinforcement learning (RL) in episodic MDPs with adversarial full-information reward feedback and unknown fixed transition kernels.

Reinforcement Learning (RL)

Risk-Sensitive Reinforcement Learning: Near-Optimal Risk-Sample Tradeoff in Regret

no code implementations • NeurIPS 2020 • Yingjie Fei, Zhuoran Yang, Yudong Chen, Zhaoran Wang, Qiaomin Xie

We study risk-sensitive reinforcement learning in episodic Markov decision processes with unknown transition kernels, where the goal is to optimize the total reward under the risk measure of exponential utility.

Q-Learning reinforcement-learning +1

Achieving the Bayes Error Rate in Synchronization and Block Models by SDP, Robustly

no code implementations • 21 Apr 2019 • Yingjie Fei, Yudong Chen

We study the statistical performance of semidefinite programming (SDP) relaxations for clustering under random graph models.

Clustering Stochastic Block Model +1

Hidden Integrality and Semi-random Robustness of SDP Relaxation for Sub-Gaussian Mixture Model

no code implementations • 17 Mar 2018 • Yingjie Fei, Yudong Chen

The error of the integer program, and hence that of the SDP, are further shown to decay exponentially in the signal-to-noise ratio.

Clustering

Exponential error rates of SDP for block models: Beyond Grothendieck's inequality

no code implementations • 23 May 2017 • Yingjie Fei, Yudong Chen

In this paper we consider the cluster estimation problem under the Stochastic Block Model.

Stochastic Block Model