15 papers
Agents' Last Doubt: Knowing When to Act and When to Step Back in Tool-Using Agents
Qisai Liu, Qingyue Jiao, Bangwei Guo +7
Tool-using LLM agents repeatedly face decision points under uncertainty, choosing between autonomous action, additional information gathering, and human escalation. These decisions…
Continuous-Time Probabilistic Correctors for Uncertainty-Aware Physics-Based Spacecraft Trajectory Forecasting
Muhammad Bilal Shahid, Zhanhong Jiang, Soumik Sarkar +1
Long-horizon spacecraft trajectory forecasting suffers from error accumulation due to the absence of corrective observations in the forecast regime, making reliable uncertainty est…
Lighting-aware Unified Model for Instance Segmentation
Qisai Liu, Alloy Das, Zhanhong Jiang +4
Foundation models like the Segment Anything Model (SAM) demonstrate impressive zero-shot generalization but frequently degrade under diverse real-world illumination, particularly f…
Distributed Direct Preference Optimization
Zhanhong Jiang
Preference-based reinforcement learning (RL) is a key paradigm for aligning policies with human judgments, yet its theoretical behavior in distributed settings where preference dat…
TabQL: In-Context Q-Learning with Tabular Foundation Models
Qisai Liu, Zhanhong Jiang, Timilehin Ayanlade +4
We propose Tabular Q-Learning (TabQL), a reinforcement learning framework that replaces the conventional parametric Q-network in Deep Q-Learning (DQN) with a tabular foundation mod…
COOPO: Cyclic Offline-Online Policy Optimization Algorithm
Qisai Liu, Zhanhong Jiang, Joshua Russell Waite +3
Offline reinforcement learning struggles with distributional shift and constrained performance due to static dataset limitations, while online RL demands prohibitive environment in…
ADKO: Agentic Decentralized Knowledge Optimization
Lucas Nerone Rillo, Zhanhong Jiang, Nastaran Saadati +4
We present Agentic Decentralized Knowledge Optimization (ADKO), a framework for collaborative black-box optimization across autonomous agents that achieves sample efficiency, priva…
LexiSafe: Offline Safe Reinforcement Learning with Lexicographic Safety-Reward Hierarchy
Hsin-Jung Yang, Zhanhong Jiang, Prajwal Koirala +3
Offline safe reinforcement learning (RL) is increasingly important for cyber-physical systems (CPS), where safety violations during training are unacceptable and only pre-collected…
Neural CDEs as Correctors for Learned Time Series Models
Muhammad Bilal Shahid, Zhanhong Jiang, Prajwal Koirala +2
Learned time-series models, whether continuous or discrete, are widely used for forecasting the states of dynamical systems but suffer from error accumulation in multi-step forecas…
Balancing Utility and Privacy: Dynamically Private SGD with Random Projection
Zhanhong Jiang, Md Zahid Hasan, Nastaran Saadati +3
Stochastic optimization is a pivotal enabler in modern machine learning, producing effective models for various tasks. However, several existing works have shown that model paramet…
Decentralized Relaxed Smooth Optimization with Gradient Descent Methods
Zhanhong Jiang, Aditya Balu, Soumik Sarkar
-smoothness, which has been pivotal to advancing decentralized optimization theory, is often fairly restrictive for modern tasks like deep learning. The recent advent of relax…
Data-driven Kinematic Modeling in Soft Robots: System Identification and Uncertainty Quantification
Zhanhong Jiang, Dylan Shah, Hsin-Jung Yang +1
Precise kinematic modeling is critical in calibration and controller design for soft robots, yet remains a challenging issue due to their highly nonlinear and complex behaviors. To…
DeCAF: Decentralized Consensus-And-Factorization for Low-Rank Adaptation of Foundation Models
Nastaran Saadati, Zhanhong Jiang, Joshua R. Waite +4
Low-Rank Adaptation (LoRA) has emerged as one of the most effective, computationally tractable fine-tuning approaches for training Vision-Language Models (VLMs) and Large Language…
Bidirectional Linear Recurrent Models for Sequence-Level Multisource Fusion
Qisai Liu, Zhanhong Jiang, Joshua R. Waite +3
Sequence modeling is a critical yet challenging task with wide-ranging applications, especially in time series forecasting for domains like weather prediction, temperature monitori…
FAWAC: Feasibility Informed Advantage Weighted Regression for Persistent Safety in Offline Reinforcement Learning
Prajwal Koirala, Zhanhong Jiang, Soumik Sarkar +1
Safe offline reinforcement learning aims to learn policies that maximize cumulative rewards while adhering to safety constraints, using only offline data for training. A key challe…