arXiv subjects

Jia Xie

Publications and source records attributed to Jia Xie.

3 recordsLinked to original sources

arXiv2026

AM-Bench: A Modular Simulation Suite and Benchmark for Aerial Manipulation Policy Learning

Standardized benchmarks have played a central role in advancing robot manipulation learning, yet most focus on ground-supported manipulation systems, which limits their applicability to dynamics-critical domains such as aerial manipulation (AM). AM presents distinct system-level challenges, including environmental disturbances, coupled dynamics between the manipulator and floating base, and constrained degrees of freedom. Consequently, task performance depends jointly on robot embodiment, low-level control, and high-level policy design. We introduce AM-Bench, a modular simulation suite and benchmark for multirotor-based AM policy learning. AM-Bench includes representative embodiments spanning underactuated, fully actuated, and overactuated systems, 12 tasks across contact, transport, and constrained interaction, configurable aerodynamic disturbances and actuator saturation, standard low-level controllers, and baseline policy-learning algorithms. Unlike prior manipulation benchmarks that primarily emphasize end-to-end policy performance, AM-Bench enables system-level evaluation of how embodiment, control, disturbances, and policy choices interact. We demonstrate its diagnostic value through three simulation studies spanning high-level policies, policy--control interfaces, and embodiments, together with real-world validation of modeled effects and a hardware test of the learning pipeline.

cs.RO↗

arXiv2026

ENPIRE: Agentic Robot Policy Self-Improvement in the Real World

Achieving dexterous robotic manipulation in the real world heavily relies on human supervision and algorithm engineering, which becomes a central bottleneck in the pursuit of general physical intelligence. Although emerging coding agents can generate code to automate algorithm search, their successes remain largely confined in digital environments. We conjecture that the missing abstraction to automate robotics research is a repeatable feedback loop for real-world policy improvement: reset the scene, execute a policy, verify the outcome, and refine the next iteration. To bridge this gap, we introduce ENPIRE, a harness framework for coding agents that instantiates this physical feedback routine with four core modules: an Environment module (EN) for automatic reset and verification, a Policy Improvement module (PI) that launches policy refinement, a Rollout module (R) to evaluate policies with one or multiple physical robots operating in parallel, and an Evolution module (E) in which coding agents analyze logs, consult literature, improve training infrastructure and algorithm code to address failure modes. This closed-loop system transforms real-world manipulation learning into a controllable optimization procedure, minimizing human effort while allowing fair ablations across training recipe and agent variants. Powered by ENPIRE, frontier coding agents can autonomously train a policy to achieve a 99% success rate on challenging, dexterous manipulation tasks, such as organizing a pin box, fastening a zip tie, and tool use, a process that further accelerates when we dispatch an agent team on a robot fleet. Our results suggest a practical and scalable path toward deploying coding agents to autonomously advancing robotics in the physical world.

cs.AI↗

arXiv2026

A dual-system approach for epilepsy diagnosis: integrating mamba-Bi-LSTM architecture with SHAP-based verification

This study develops a medical AI-assisted diagnosis system based on deep learning, which provides intelligent diagnostic solutions for epilepsy, a disease that seriously threatens the life and health of patients. Epilepsy has sudden and unpredictable seizures. Traditional diagnostic methods mainly rely on doctors' manual interpretation of EEG, which is time-consuming and dependent by experience. In response to the above challenges, this study designed a dual-system intelligent diagnosis framework, which includes two core components: the main discrimination system and the verification system. The main discrimination system uses a deep learning model that combines the innovative Mamba architecture with the Bi-LSTM structure to integrate and analyze heterogeneous data to achieve extremely high diagnostic accuracy; the verification system provides an explainable diagnostic basis through the SHAP method to enhance the credibility of the results. This system establishes a cross-modal database to realize intelligent analysis of multi-source heterogeneous data-fusion EEG signals and clinical text data for epilepsy. The system outputs results based on diagnostic consistency and confidence levels, and high-confidence predictions can also be used as automatic feedback sources to optimize the model. The experimental results show that the accuracy of the main discriminant model of the intelligent diagnosis system for epilepsy has increased from 92.6% to 98.7% and the F1 score has increased from 0.895 to 0.992, all of which have exceeded the existing optimal methods; the average processing time for verification system feedback integration is only 220 ms, which increases the overall diagnostic accuracy by 5.1%.

eess.SP↗