Publications

Neural Active Learning Beyond Bandits

Yikun Ban, Ishika Agarwal, Ziwei Wu, Yada Zhu, Kommy Weldemariam, Hanghang Tong, Jingrui He

Published in ICLR, 2024

We study both stream-based and pool-based active learning with neural network approximations. A recent line of works proposed bandit-based approaches that transformed active learning into a bandit problem, achieving both theoretical and empirical success. However, the performance and computational costs of these methods may be susceptible to the number of classes, denoted as K, due to this transformation. Therefore, this paper seeks to answer the question: “How can we mitigate the adverse impacts of K while retaining the advantages of principled exploration and provable performance guarantees in active learning?” To tackle this challenge, we propose two algorithms based on the newly designed exploitation and exploration neural networks for stream-based and pool-based active learning. Subsequently, we provide theoretical performance guarantees for both algorithms in a non-parametric setting, demonstrating a slower error-growth rate concerning K for the proposed approaches. We use extensive experiments to evaluate the proposed algorithms, which consistently outperform state-of-the-art baselines.

Download here

QuickAns: A Virtual TA

Ishika Agarwal, Shradha Sehgal, Varun Goyal, Prathamesh Sonawane

Published in AIML Systems, 2023

QuickAns is a virtual teaching assistant designed to help course staff who use Campuswire as their Q&A platform. It reads Campuswire posts from digest emails, and sends a potential answer to the course staff. At this stage, the course staff can review the answer for any logistical issues, and answer a student’s question in a matter of minutes.

Download here

HiSaRL: A Hierarchical Framework for Safe Reinforcement Learning

Zikang Xiong, Ishika Agarwal, Suresh Jagannathan

Published in SafeAI @ AAAI, 2021

We propose a two-level hierarchical framework for safe reinforcement learning in a complex environment. The high-level part is an adaptive planner, which aims at learning and generating safe and efficient paths for tasks with imperfect map information. The lower-level part contains a learning-based controller and its corresponding neural Lyapunov function, which characterizes the controller’s stability property. This learned neural Lyapunov function serves two purposes. First, it will be part of the high-level heuristic for our planning algorithm. Second, it acts as a part of a runtime shield to guard the safety of the whole system. We use a robot navigation example to demonstrate that our framework can operate efficiently and safely in complex environments, even under adversarial attacks.

Download here