Searcharxiv⌕ Search

arXiv subjects

Zixu Hao

Publications and source records attributed to Zixu Hao.

10 recordsLinked to original sources

SteerDistill: Quantizing LLM Agents via Closed-Loop Teacher Steering

Low-bit quantization reduces the inference cost of LLM agents but can compromise task success. Examining failed rollouts, we find that quantized agents overlook task requirements or environmental conditions, producing environment transitions that diverge from successful full-precision references. Correcting these behavioral divergences remains difficult: offline data may miss the student's failure contexts, while its own rollouts may provide few correct responses in those contexts. We introduce SteerDistill, a quantization-aware distillation method in which a full-precision teacher guides the student to revise failed responses and iteratively refines its guidance using feedback from student execution. This closed-loop teacher steering produces locally verified behavioral corrections. We internalize these corrections by training the student without guidance to match the steered student's predictions at corrected turns and the full-precision teacher's at other turns, enabling unassisted inference. Across five agentic benchmarks and three models, SteerDistill exceeds the strongest quantized baseline by 1.67--2.18 points in average score under W4A16 and stays within one point of BF16. On $τ^2$-Bench, it outperforms QeRL by 1.84 percentage points in success rate using 67% fewer training GPU hours and produces shorter successful trajectories than the quantized baselines.

cs.AI↗

Replicate What Varies: Resource Sharing for Domain-Randomized Embodied Simulation

Simulation provides a low-cost sandbox where embodied agents can attempt tasks, observe outcomes, and learn from experience. Running many diverse simulation environments concurrently can consume substantial CPU and GPU memory because independently constructed scenes duplicate resources that remain unchanged across instances. We present Patchwork, a resource-sharing system for domain-randomized simulation, built around the representation $E=T+Δ$: a shared template contains common resources, and a patch records the mutations needed to construct a target environment. Patchwork separates randomization sampling from replay and selects reusable templates together with patches that preserve each target environment's intended configuration. Its layered runtime shares compiled model structure and immutable assets while keeping mutable parameters and simulation state private to each instance. We implement Patchwork on robosuite and MuJoCo and evaluate it on LIBERO-plus and RoboCasa365 workloads. At 32 replicas, shared-template execution reduces CPU RSS by 88-92% relative to process-per-environment execution. Controlled experiments show how sharing shifts memory growth toward template diversity and unique asset content rather than instance count, within the supported sharing boundaries.

cs.DC↗

EBRL: Asynchronous Embodied RL by Multi-Grained Resource Management

Embodied reinforcement learning (RL) improves model capabilities with a pipeline of environment simulation, action generation, and model updates. These stages show heterogeneous CPU and GPU demands, making efficient resource utilization difficult. Recent systems overlap rollout (simulation and generation) with training for efficiency, but exclusive GPU allocation and synchronized barrier in rollout still leave substantial hardware resource waste. In this paper, we present EBRL, an asynchronous embodied RL training system with two core techniques. The asynchronous pipelined scheduler overlaps rollout and training, pipelines simulation and generation across environment groups, and carries out each environment independently, eliminating synchronization stalls. The fine-grained resource manager pools CPU cores and GPU streaming multiprocessors, and uses stage profiles and runtime feedback to adjust resource quotas and batch sizes to meet the shifting demands among stages. We implement EBRL on RLinf and evaluate it with four embodied policies and four simulation benchmarks across heterogeneous GPU testbeds. Experiments show that EBRL achieves 1.30-3.47 times the end-to-end rollout throughput and 2.5 times of training convergency compared to the SOTA embodied RL systems.

cs.LG↗

Zetta $ζ$: An Efficient Closed-Loop Embodied Harness for Self-Evolving Physical Intelligence

Embodied agents are increasingly used to close the gap left by end-to-end policy models. Yet the agentic path has not realized closed-loop learning in physical execution: existing harnesses remain largely open-loop, following fixed skills during rollout and reflecting only after an episode completes. Such post-hoc reflection cannot govern execution as it unfolds, because physical interaction requires decisions to track rapidly changing robot-environment states at a frequency beyond today's large agentic models. We present Zetta, a closed-loop embodied harness that evolves code-based runtime critics and recovery skills online while keeping the base policy frozen. Through three timescale-separated loops, Zetta provides action-frequency governance, rollout-level critic-recovery proposal, and validation-gated skill updates. Together with Z-Infra, a rollout infrastructure decoupling agent logic from heterogeneous execution resources, Zetta achieves state-of-the-art success on LIBERO-Pro and RoboCasa under our current rollout budget, reaching 90.8% and 93.6%, with an 11.1x inference speedup; success continues to scale with self-exploration experience; learned skills transfer zero-shot, and clear robotic "Aha Moments" emerge. These results show that closed-loop harness self-evolution opens a scaling path for reliable physical intelligence.

cs.RO↗

ActProbe: Action-Space Probe for Early Failure Detection of Generative Robot Policies

Generative robot policies fail unpredictably at deployment: they hesitate at critical moments, drift off-task, or commit to unrecoverable actions. Existing online failure detectors either require white-box access to policy internals or add runtime overhead through resampling and observation-side signals. Our empirical analysis shows that emitted action chunks themselves already carry strong predictive signal for impending failures in generative robot policies. Motivated by this observation, we introduce ActProbe, a lightweight, pure action-space detector that uses two compact signals available from a single forward pass: Temporal Consistency Error (TCE) between consecutive action chunks and Action Chunk Magnitude (ACM) of the current chunk. ActProbe maps these signals to per-step failure probabilities with a task-conditioned LSTM-MLP architecture. Across a diverse suite of generative robot policies and benchmarks, ActProbe raises alerts before failures become visually recognizable, improving the accuracy (F1)-timeliness Pareto frontier of failure detection by an average hypervolume gain of +12.7% over both internal- and external-feature baselines, with a +9.0% early-detection ROC-AUC lead on unseen tasks. ActProbe further transfers to deployment, predicting failures on unseen real-robot pick tasks and accelerating RL fine-tuning (PPO) with 2.9x fewer environment interactions.

cs.RO↗

T-MAN: Enabling End-to-End Low-Bit LLM Inference on NPUs via Unified Table Lookup

Large language models (LLMs) are increasingly deployed on customer devices. To support them, current devices are adopting SoCs (System on Chip) with NPUs (Neural Processing Unit) installed. Although high performance is expected, LLM inference on NPUs is slower than its CPU counterpart. The reason is that NPUs have poor performance on computations other than GEMM, like dequantization. Current works either disaggregate prefill on the NPUs and decoding on the CPUs, or put both on the NPUs but with an accuracy loss. To solve this issue, based on the insight that low-bit can enable target computation encoded within an acceptably sized table, we propose table lookup to subsume hardware operations otherwise unsupported. To realize this, we overcome the conflicting hardware behavior of prefill and decoding to design a unified table layout and tiling through (1) fused two-level table-based dequantization and (2) concurrency-hierarchy-guided tiling. Based on that, we implement the prefill phase by three-stage pipeline and map the table-lookup-based decoding to NPU's vector units. Results show 1.4x and 3.1x speedup for prefill and decoding respectively, and 84% energy savings compared to the baseline NPU methods. The code is available at https://github.com/microsoft/T-MAC/tree/main/t-man.

cs.AR↗

Long Exposure: Accelerating Parameter-Efficient Fine-Tuning for LLMs under Shadowy Sparsity

The adaptation of pre-trained large language models (LLMs) to diverse downstream tasks via fine-tuning is critical for numerous applications. However, the inefficiency of parameter-efficient fine-tuning (PEFT) techniques presents significant challenges in terms of time investments and operational costs. In this paper, we first introduce a nuanced form of sparsity, termed Shadowy Sparsity, which is distinctive in fine-tuning and has not been adequately addressed for acceleration. Under Shadowy Sparsity, we propose Long Exposure, an efficient system to accelerate PEFT for LLMs. Long Exposure comprises three key components: Shadowy-sparsity Exposer employs a prolonged sensing range to capture more sparsity details under shadowy sparsity; Sequence-oriented Predictor provides efficient yet accurate predictions to handle large sequence inputs and constantly-evolving parameters; and Dynamic-aware Operator facilitates more structured computational patterns and coalesced memory accesses, addressing dynamic sparse operations. Extensive evaluations show that Long Exposure outperforms state-of-the-arts with up to a $2.49\times$ speedup in end-to-end fine-tuning, offering promising advancements in accelerating PEFT for LLMs.

cs.LG↗

Neuralink: Fast LLM Inference on Smartphones with Neuron Co-Activation Linking

Large Language Models (LLMs) have achieved remarkable success across various domains, yet deploying them on mobile devices remains an arduous challenge due to their extensive computational and memory demands. While lightweight LLMs have been developed to fit mobile environments, they suffer from degraded model accuracy. In contrast, sparsity-based techniques minimize DRAM usage by selectively transferring only relevant neurons to DRAM while retaining the full model in external storage, such as flash. However, such approaches are critically limited by numerous I/O operations, particularly on smartphones with severe IOPS constraints. In this paper, we propose Neuralink, a novel approach that accelerates LLM inference on smartphones by optimizing neuron placement in flash memory. Neuralink leverages the concept of Neuron Co-Activation, where neurons frequently activated together are linked to facilitate continuous read access and optimize I/O efficiency. Our approach incorporates a two-stage solution: an offline stage that reorganizes neuron placement based on co-activation patterns, and an online stage that employs tailored data access and caching strategies to align well with hardware characteristics. Evaluations conducted on a variety of smartphones and LLMs demonstrate that Neuralink achieves on average $1.49\times$ improvements in end-to-end latency compared to the state-of-the-art. As the first solution to optimize storage placement under sparsity, Neuralink explores a new optimization space at the intersection of sparsity-driven algorithm and storage-level system co-design for LLM inference.

cs.LG↗

Scaling LLM Test-Time Compute with Mobile NPU on Smartphones

Deploying Large Language Models (LLMs) on mobile devices faces the challenge of insufficient performance in smaller models and excessive resource consumption in larger ones. This paper highlights that mobile Neural Processing Units (NPUs) have underutilized computational resources, particularly their matrix multiplication units, during typical LLM inference. To leverage this wasted compute capacity, we propose applying parallel test-time scaling techniques on mobile NPUs to enhance the performance of smaller LLMs. However, this approach confronts inherent NPU challenges, including inadequate hardware support for fine-grained quantization and low efficiency in general-purpose computations. To overcome these, we introduce two key techniques: a hardware-aware tile quantization scheme that aligns group quantization with NPU memory access patterns, and efficient LUT-based replacements for complex operations such as Softmax and dequantization. We design and implement an end-to-end inference system that leverages the NPU's compute capability to support test-time scaling on Qualcomm Snapdragon platforms. Experiments show our approach brings significant speedups: up to 19.0 for mixed-precision GEMM and 2.2 for Softmax. More importantly, we demonstrate that smaller models using test-time scaling can match or exceed the accuracy of larger models, achieving a new performance-cost Pareto frontier.

cs.DC↗

Design, Modeling and Dynamic Compensation PID Control of a Fully-Actuated Aerial Manipulation System

This paper addresses design, modeling and dynamic-compensation PID (dc-PID) control of a novel type of fully-actuated aerial manipulation (AM) system. Firstly, design of novel mechanical structure of the AM is presented. Secondly, kinematics and dynamics of AM are modeled using Craig parameters and recursion Newton-Euler equations respectively, which give rise to a more accurate dynamic relationship between aerial platform and manipulator. Then, the dynamic-compensation PID control is proposed to solve the problem of fully-actuated control of AM. Finally, uniform coupled matrix equations between driving forces/moments and rotor speeds are derived, which can support design and analysis of parameters and decoupling theoretically. It is taken into account practical problems including noise and perturbation, parameter uncertainty, and power limitation in simulations, and results from simulations shows that the AM system presented can be fully-actued controlled with advanced control performances, which can not achieved theoretically in traditional AM. And with compared to backstepping control dc-PID has better control accuracy and capability to disturbance rejection in two simulations of aerial operation tasks with motion of joint. The experiment of dc-pid proves the availability and effectiveness of the method proposed.

cs.RO↗