AsyncOPD: How Stale Can On-Policy Distillation Be?

On-policy distillation (OPD) trains a student on its own rollouts guided by teacher feedback and is becoming increasingly important for large language model (LLM) post-training. Like reinforcement learning (RL), however, OPD faces an on-policy systems bottleneck, as rollouts can dominate training time for reasoning workloads. Asynchronous training pipelines can alleviate this bottleneck by decoupling rollout generation from learner updates, but doing so introduces stale-policy data. While prior work has studied stale data in asynchronous RL, its effects in OPD remain underexplored. We present the first systematic study of staleness in asynchronous OPD, focusing on a practical setting where teacher feedback is implemented through local KL losses and full-vocabulary teacher logits are too expensive to store or transfer, necessitating finite teacher-score caches. We first show that KL direction changes the stale-data problem: teacher-weighted forward KL is more robust to stale rollouts, whereas student-weighted reverse KL is vulnerable. Second, for this vulnerable reverse-KL case, we study whether methods designed to stabilize asynchronous RL can mitigate OPD staleness. In our experiments, they do not improve over a simpler OPD-specific surrogate: recomputing the reverse-KL signal under the current student at learner time. Third, we analyze how finite teacher-score caches create a bias-variance tradeoff for sparse and sampled reverse-KL OPD estimators. This motivates multi-sample Monte Carlo (MC), which preserves MC correctability while reducing one-sample variance. Finally, we present and open-source AsyncOPD, a fully asynchronous OPD training pipeline built from these estimator choices. Experiments show that AsyncOPD improves training throughput by $1.6\times$ to $3.8\times$ over strict synchronous training while reaching comparable accuracy.

EfficientRollout: System-Aware Self-Speculative Decoding for RL Rollouts

Reinforcement learning (RL) has become a representative post-training paradigm for LLMs, enabling strong reasoning and agentic capabilities. However, rollout generation remains a dominant latency bottleneck because autoregressive sampling decodes responses sequentially and a small number of long-tailed generations often determine completion time. Speculative decoding (SD) offers a natural way to address this bottleneck, as it is a well-established technique for serving fixed LLMs that reduces latency by rapidly drafting tokens and accepting them through parallel verification while preserving the target-model distribution. However, its practical speedups do not directly carry over to RL rollouts: (i) the evolving target policy makes any fixed drafter increasingly mismatched with the policy's output distribution; and (ii) active batch sizes shrink throughout rollout decoding, shifting decoding from compute-bound to memory-bound regimes where parallel verification can exploit underutilized compute. Therefore, accelerating RL rollouts requires both a drafter that remains effective under long, high-temperature generations from an evolving policy and system-aware use of SD that avoids compute-bound regimes. We present EfficientRollout, a system-aware self-SD framework designed to address this gap for RL rollouts. EfficientRollout induces a quantized drafter from the target model (i.e. self-speculative decoding), keeping it coupled to the evolving policy without separate drafter pretraining or online adaptation. It further coordinates a system-aware SD toggle policy with acceptance-aware draft-length adaptation, enabling speculation only in beneficial regimes while matching the drafting budget to evolving drafter quality. EfficientRollout reduces rollout and end-to-end latency by up to 19.6% and 12.7%, respectively, over an accelerated AR rollout baseline, while preserving final model quality.

ACT: Automated CPS Testing for Open-Source Robotic Platforms

Open-source software for cyber-physical systems (CPS) often lacks robust testing involving robotic platforms, resulting in critical errors that remain undetected. This is especially challenging when multiple modules of CPS software are developed by various open-source contributors. To address this gap, we propose Automated CPS Testing (ACT) that performs automated, continuous testing of open-source software with its robotic platforms, integrated with the open-source infrastructure such as GitHub. We implement an ACT prototype and conduct a case study on an open-source CPS with an educational robotic platform to demonstrate its capabilities.

Advancing Multi-Robot Networks via MLLM-Driven Sensing, Communication, and Computation: A Comprehensive Survey

Imagine advanced humanoid robots, powered by multimodal large language models (MLLMs), coordinating missions across industries like warehouse logistics, manufacturing, and safety rescue. While individual robots show local autonomy, realistic tasks demand coordination among multiple agents sharing vast streams of sensor data. Communication is indispensable, yet transmitting comprehensive data can overwhelm networks, especially when a system-level orchestrator or cloud-based MLLM fuses multimodal inputs for route planning or anomaly detection. These tasks are often initiated by high-level natural language instructions. This intent serves as a filter for resource optimization: by understanding the goal via MLLMs, the system can selectively activate relevant sensing modalities, dynamically allocate bandwidth, and determine computation placement. Thus, R2X is fundamentally an intent-to-resource orchestration problem where sensing, communication, and computation are jointly optimized to maximize task-level success under resource constraints. This survey examines how integrated design paves the way for multi-robot coordination under MLLM guidance. We review state-of-the-art sensing modalities, communication strategies, and computing approaches, highlighting how reasoning is split between on-device models and powerful edge/cloud servers. We present four end-to-end demonstrations (sense -> communicate -> compute -> act): (i) digital-twin warehouse navigation with predictive link context, (ii) mobility-driven proactive MCS control, (iii) a FollowMe robot with a semantic-sensing switch, and (iv) real-hardware open-vocabulary trash sorting via edge-assisted MLLM grounding. We emphasize system-level metrics -- payload, latency, and success -- to show why R2X orchestration outperforms purely on-device baselines.

GazeShift: Unsupervised Gaze Estimation and Dataset for VR

Gaze estimation is instrumental in modern virtual reality (VR) systems. Despite significant progress in remote-camera gaze estimation, VR gaze research remains constrained by data scarcity - particularly the lack of large-scale, accurately labeled datasets captured with the off-axis camera configurations typical of modern headsets. Gaze annotation is difficult since fixation on intended targets cannot be guaranteed. To address these challenges, we introduce VRGaze - the first large-scale off-axis gaze estimation dataset for VR - comprising 2.1 million near-eye infrared images collected from 68 participants. We further propose GazeShift, an attention-guided unsupervised framework for learning gaze representations without labeled data. Unlike prior redirection-based methods that rely on multi-view or 3D geometry, GazeShift is tailored to near-eye infrared imagery, achieving effective gaze-appearance disentanglement in a compact, real-time model. GazeShift embeddings can be optionally adapted to individual users via lightweight few-shot calibration, achieving a 1.84-degree mean error on VRGaze. On the remote-camera MPIIGaze dataset, the model achieves a 7.15-degree person-agnostic error, doing so with 10x fewer parameters and 35x fewer FLOPs than baseline methods. Deployed natively on a VR headset GPU, inference takes only 5 ms. Combined with demonstrated robustness to illumination changes, these results highlight GazeShift as a label-efficient, real-time solution for VR gaze tracking. Project code and the VRGaze dataset are released at https://github.com/gazeshift3/gazeshift.

Multimodal Safety Evaluation in Generative Agent Social Simulations

Can generative agents be trusted in multimodal environments? Despite advances in large language and vision-language models that enable agents to act autonomously and pursue goals in rich settings, their ability to reason about safety, coherence, and trust across modalities remains limited. We introduce a reproducible simulation framework for evaluating agents along three dimensions: (1) safety improvement over time, including iterative plan revisions in text-visual scenarios; (2) detection of unsafe activities across multiple categories of social situations; and (3) social dynamics, measured as interaction counts and acceptance ratios of social exchanges. Agents are equipped with layered memory, dynamic planning, multimodal perception, and are instrumented with SocialMetrics, a suite of behavioral and structural metrics that quantifies plan revisions, unsafe-to-safe conversions, and information diffusion across networks. Experiments show that while agents can detect direct multimodal contradictions, they often fail to align local revisions with global safety, reaching only a 55 percent success rate in correcting unsafe plans. Across eight simulation runs with three models - Claude, GPT-4o mini, and Qwen-VL - five agents achieved average unsafe-to-safe conversion rates of 75, 55, and 58 percent, respectively. Overall performance ranged from 20 percent in multi-risk scenarios with GPT-4o mini to 98 percent in localized contexts such as fire/heat with Claude. Notably, 45 percent of unsafe actions were accepted when paired with misleading visuals, showing a strong tendency to overtrust images. These findings expose critical limitations in current architectures and provide a reproducible platform for studying multimodal safety, coherence, and social dynamics.

Modeling and Control Framework for Autonomous Space Manipulator Handover Operations

Autonomous space robotics is poised to play a vital role in future space missions, particularly for In-space Servicing, Assembly, and Manufacturing (ISAM). A key capability in such missions is the Robot-to-Robot (R2R) handover of mission-critical objects. This work presents a dynamic model of a dual-arm space manipulator system and compares various tracking control laws. The key contributions of this work are the development of a cooperative manipulator dynamic model and the comparative analysis of control laws to support autonomous R2R handovers in ISAM scenarios.

Adapting Biological Reflexes for Dynamic Reorientation in Space Manipulator Systems

Robotic arms mounted on spacecraft, known as space manipulator systems (SMSs), are critical for enabling on-orbit assembly, satellite servicing, and debris removal. However, controlling these systems in microgravity remains a significant challenge due to the dynamic coupling between the manipulator and the spacecraft base. This study explores the potential of using biological inspiration to address this issue, focusing on animals, particularly lizards, that exhibit mid-air righting reflexes. Based on similarities between SMSs and these animals in terms of behavior, morphology, and environment, their air-righting motion trajectories are extracted from high-speed video recordings using computer vision techniques. These trajectories are analyzed within a multi-objective optimization framework to identify the key behavioral goals and assess their relative importance. The resulting motion profiles are then applied as reference trajectories for SMS control, with baseline controllers used to track them. The findings provide a step toward translating evolved animal behaviors into interpretable, adaptive control strategies for space robotics, with implications for improving maneuverability and robustness in future missions.

Visually Guided Decoding: Gradient-Free Hard Prompt Inversion with Language Models

Text-to-image generative models like DALL-E and Stable Diffusion have revolutionized visual content creation across various applications, including advertising, personalized media, and design prototyping. However, crafting effective textual prompts to guide these models remains challenging, often requiring extensive trial and error. Existing prompt inversion approaches, such as soft and hard prompt techniques, are not so effective due to the limited interpretability and incoherent prompt generation. To address these issues, we propose Visually Guided Decoding (VGD), a gradient-free approach that leverages large language models (LLMs) and CLIP-based guidance to generate coherent and semantically aligned prompts. In essence, VGD utilizes the robust text generation capabilities of LLMs to produce human-readable prompts. Further, by employing CLIP scores to ensure alignment with user-specified visual concepts, VGD enhances the interpretability, generalization, and flexibility of prompt generation without the need for additional training. Our experiments demonstrate that VGD outperforms existing prompt inversion techniques in generating understandable and contextually relevant prompts, facilitating more intuitive and controllable interactions with text-to-image models.

Study on Real-Time Road Surface Reconstruction Using Stereo Vision

Road surface reconstruction plays a crucial role in autonomous driving, providing essential information for safe and smooth navigation. This paper enhances the RoadBEV [1] framework for real-time inference on edge devices by optimizing both efficiency and accuracy. To achieve this, we proposed to apply Isomorphic Global Structured Pruning to the stereo feature extraction backbone, reducing network complexity while maintaining performance. Additionally, the head network is redesigned with an optimized hourglass structure, dynamic attention heads, reduced feature channels, mixed precision inference, and efficient probability volume computation. Our approach improves inference speed while achieving lower reconstruction error, making it well-suited for real-time road surface reconstruction in autonomous driving.