Institution profile

Amazon

Industry researchnorthamerica · us
Official website
Research library1,530linked papers
Opportunities1,135open roles
Selected work

Representative Papers

Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfaces

Jan 17, 2026

This work addresses the challenge that existing AI agent benchmarks inadequately evaluate performance on real-world, complex, and long-horizon command-line tasks. To bridge this gap, the authors introduce a novel evaluation benchmark comprising 89 high-difficulty terminal tasks, all derived from authentic workflows and accompanied by isolated execution environments, human-authored reference solutions, and automated verification tests. The benchmark is designed to ensure realism, verifiability, and diversity, substantially narrowing the disparity between practical scenarios and current model evaluation paradigms. Experimental results demonstrate that even state-of-the-art agents achieve success rates below 65% on this benchmark. The paper further provides comprehensive error analysis and publicly releases the dataset and evaluation toolchain to support future research in this domain.

9 citations1 influentialRead paper

Agentic Reasoning for Large Language Models

Jan 18, 2026

This work addresses the limited capacity of large language models (LLMs) to plan, act, and learn through sustained interaction in open, dynamic environments. To overcome this, the authors propose a three-tiered reasoning framework that treats LLMs as autonomous agents, unifying single-agent foundational reasoning, self-evolution, and multi-agent collaboration within a coherent paradigm. The framework orchestrates structured interactions, incorporates memory mechanisms, enables tool use, and integrates both reinforcement learning and supervised fine-tuning. It explicitly distinguishes between in-context reasoning and post-training optimization pathways, systematically coupling cognition with action. Empirical validation across diverse domains—including scientific discovery, robotics, healthcare, autonomous research, and mathematics—demonstrates its effectiveness and highlights promising future directions such as personalized interaction, long-horizon engagement, world modeling, and scalable multi-agent training.

7 citations1 influentialRead paper

Root Cause Analysis of Outliers with Missing Structural Knowledge

Jun 07, 2024arXiv.org

Real-world root cause analysis (RCA) faces a critical challenge: post-intervention distributions often contain only a few—or even a single—sample, rendering distribution-dependent or low-density-region regression methods statistically ill-posed. This paper proposes a lightweight root cause identification framework that requires neither counterfactual reasoning nor a fully specified structural causal model (SCM). It operates either given a causal DAG or, in the absence of one, solely from an anomaly score ranking. We theoretically prove that low-scoring anomalies rarely trigger high-scoring ones and derive a probabilistic upper bound on non-monotonic propagation paths. By abandoning Shapley-value-based attribution and density-sensitive regression, our method achieves linear time complexity O(n). It eliminates SCM fitting and counterfactual computation while providing rigorous theoretical guarantees and strong empirical performance.

4 citations1 influentialRead paper

Ordering-based Conditions for Global Convergence of Policy Gradient Methods

Apr 02, 2025Neural Information Processing Systems

This paper investigates the global convergence of policy gradient (PG) methods in finite-armed bandits under linear function approximation. It identifies a fundamental limitation of conventional convergence analyses—namely, their reliance on approximation error—and reveals that convergence fundamentally hinges on the ability of the policy update and function representation to preserve the relative ordering of action rewards. The work establishes, for the first time, necessary and sufficient convergence conditions based on reward-order preservation: natural policy gradient (NPG) converges globally if and only if the linear projection preserves the relative order of optimal actions; softmax PG additionally requires non-dominance and order preservation. These results are rigorously proved via projection geometry and policy gradient theory, and empirically validated. The core contribution is a reformulation of PG convergence criteria that decouples representational capacity from optimization dynamics, providing a novel theoretical foundation for trustworthy function-approximation-based reinforcement learning.

4 citationsRead paper
Recent publications

Latest Papers