TALSC: Timeliness-Aware Large-Small VLM Collaboration for Infrastructure-Assisted Autonomous Driving

The deployment of Vision-Language Models (VLMs) in autonomous driving (AD) systems is constrained by on-board computing power, restricting vehicles to small VLMs (SVLMs) with limited perception and reasoning capabilities. Infrastructure-assisted AD alleviates this resource constraint by enabling collaboration with large VLMs (LVLMs) at edge servers. However, in dynamic vehicular environments, the utility of sensory data for downstream tasks decays rapidly, making timeliness of information a critical concern. To balance the accuracy gains of LVLMs with their latency-induced timeliness degradation, we develop a Timeliness-Aware Large-Small VLM Collaboration (TALSC) framework. Specifically, we first model the Age of Information (AoI) evolution for VLM inference and characterize the coupling among AoI, token length, and task performance to formulate a general timeliness metric. Building on this, we propose the TALSC online scheduling algorithm. Since scheduling decisions have a delayed impact on future timeliness metric and the output token number is unknown at scheduling time, we design a Lyapunov drift-plus-estimated-penalty algorithm and provides a guaranteed performance. In simulation, we first conduct a case study to derive a fitted timeliness metric based on nuScenes dataset, and further show that TALSC outperforms baselines under various communication and computing settings, achieving up to a 12.6\% normalized improvement in Micro-F1 score compared with the best-performing baseline.

Rapid Embodiment Adaptation for Quadrupedal Locomotion

Humans readily adapt their movements as their bodies change through aging, injury, or load carrying, but learning-based robot policies often break when hardware properties shift. We introduce an online embodiment adaptation framework for quadrupedal locomotion that infers embodiment parameters from short interaction histories and conditions control on the inferred hardware state. Our method pairs a generalist policy trained under embodiment randomization with a lightweight adaptation module that identifies physical changes within half a second. We evaluate two representative forms of embodiment variation: joint-range constraints and trunk-mass changes, corresponding to joint-level kinematic degradation and body-level dynamic variation. In simulation, the module accurately estimates these changes and enables closed-loop control that substantially outperforms policies conditioned directly on interaction history. On a real Unitree Go2 robot, our system maintains stable locomotion under severe instances of the evaluated changes, including a fully locked leg and a 5 kg payload, where non-adaptive methods fail. These results demonstrate the practicality of explicit online embodiment identification for rapid adaptation to joint-limit and payload-mass changes, and provide a step toward handling broader forms of uncertain, degraded, or changing robot hardware.

ComHymba: Low-Complexity Domain-Informed Foundation Model for Wireless Communications

Wireless foundation models are a promising route to unify channel reconstruction, sensing, and beam management in future wireless communication systems, but existing designs often inherit LLM-style Transformers with quadratic token complexity and weak integration of propagation priors. This paper proposes ComHymba, a domain-informed wireless foundation model built on an asymmetric masked autoencoder for large-scale self-supervised pre-training on Channel State Information (CSI). ComHymba introduces (i) 3D spatio-temporal-frequency patchification with rotary positional embedding, (ii) domain-informed masking strategies that emulate realistic CSI sparsity and fading patterns, and (iii) a decoupled amplitude--phase weighted objective tailored to channel statistics. Architecturally, we employ Hymba blocks that fuse windowed self-attention with state space models (SSMs), enabling linear-time modeling with respect to the overall channel input size. Experiments on eight downstream tasks spanning channel state information reconstruction, environmental sensing, and beam management show consistent accuracy gains over strong task-specific baselines, together with up to a $3.3\times$ inference speedup versus Transformer backbones. Overall, ComHymba provides a scalable and efficient backbone for AI-native physical-layer intelligence.

GELATO: Generative Entropy- and Lyapunov-based Adaptive Token Offloading for Device-Edge Speculative LLM Inference

The recent growth of on-device Large Language Model (LLM) inference has driven significant interest in device-edge collaborative LLM inference. As a promising architecture, Speculative Decoding (SD) is increasingly adopted where a lightweight draft model rapidly generates candidate tokens to be verified by a powerful target model. However, a fundamental challenge lies in achieving per-token resource scheduling to effectively adapt SD paradigm to resource-constrained edge environment. This paper proposes a Generative Entropy- and Lyapunov-based Adaptive Token Offloading framework, named GELATO, to maximize decoding throughput under energy constraints in a device-edge collaborative SD system. Specifically, an outer drift-plus-penalty loop makes online decisions to establish a reference drafting budget, managing long-term energy-throughput trade-off. Further, a nested entropy-driven generation mechanism executes early exiting to adapt to per-token dynamic generative uncertainty. Theoretical analysis establishes a rigorous performance bound on long-term throughput for GELATO. Extensive evaluations demonstrate that GELATO achieves a globally optimal tradeoff, outperforming state-of-the-art distributed SD architectures by 64.98% in token throughput and reducing energy consumption by 47.47% under resource-constrained environments, while preserving LLM decoding quality.

Mobile Communications in Intelligent Rail Transit: From LCX to PASS

Wireless communications in intelligent rail transit face harsh propagation conditions, including severe penetration loss, frequent blockages, and amplified large-scale fading. Existing leaky coaxial cables (LCX) provide wired-to-wireless conversion and stable coverage, but can be energy- and spectrum-inefficient, particularly at high carrier frequencies. Motivated by the growing demand for high-capacity and high-reliability rail services, this article introduces pinching-antenna systems (PASS), which are flexible waveguide-based architectures that enable reconfigurable radiation points with low deployment overhead and a natural fit to predominantly straight track geometries. We discuss the key benefits and deployment flexibility of PASS, evaluate their performance relative to LCX via representative simulations, and present a deep learning (DL)-enabled channel-estimation framework to cope with mobility-induced channel dynamics. Finally, we summarize the major open challenges for practical deployment and outline promising research directions.

Deep Learning Based Site-Specific Channel Inference Using Satellite Images

Site-specific channel inference plays a critical role in the design and evaluation of next-generation wireless communication systems by considering the surrounding propagation environment. However, traditional methods are unscalable, while existing AI-based approaches using satellite image are confined to predicting large-scale fading parameters, lacking the capacity to reconstruct the complete channel impulse response (CIR). To address this limitation, we propose a deep learning-based site-specific channel inference framework using satellite images to predict structured Tapped Delay Line (TDL) parameters. We first establish a joint channel-satellite dataset based on measurements. Then, a novel deep learning network is developed to reconstruct the channel parameters. Specifically, a cross-attention-fused dual-branch pipeline extracts macroscopic and microscopic environmental features, while a recurrent tracking module captures the long-term dynamic evolution of multipath components. Experimental results demonstrate that the proposed method achieves high-quality reconstruction of the CIR in unseen scenarios, with a Power Delay Profile (PDP) Average Cosine Similarity exceeding 0.96. This work provides a pathway toward site-specific channel inference for future dynamic wireless networks.

AnyHand: A Large-Scale Synthetic Dataset for RGB(-D) Hand Pose Estimation

We present AnyHand, a large-scale synthetic dataset designed to advance the state of the art in 3D hand pose estimation from both RGB-only and RGB-D inputs. While recent works with foundation approaches have shown that an increase in the quantity and diversity of training data can markedly improve performance and robustness in hand pose estimation, existing real-world-collected datasets on this task are limited in coverage, and prior synthetic datasets rarely provide occlusions, arm details, and aligned depth together at scale. To address this bottleneck, our AnyHand contains 2.5M single-hand and 4.1M hand-object interaction RGB-D images, with rich geometric annotations. In the RGB-only setting, we show that extending the original training sets of existing baselines with AnyHand yields significant gains on multiple benchmarks (FreiHAND and HO-3D), even when keeping the architecture and training scheme fixed. More impressively, the model trained with AnyHand shows stronger generalization to the out-of-domain HO-Cap dataset, without any fine-tuning. We also contribute a lightweight depth fusion module that can be easily integrated into existing RGB-based models. Trained with AnyHand, the resulting RGB-D model achieves superior performance on the HO-3D benchmark, showing the benefits of depth integration and the effectiveness of our synthetic data.

Two-Stage Heterogeneous Graph Neural Network for RIS-Aided Physical-Layer Security

This paper investigates physical-layer security (PLS) enabled by graph neural networks (GNNs). We propose a two-stage heterogeneous GNN (HGNN) to maximize the secrecy energy efficiency (SEE) of a reconfigurable intelligent surface (RIS)-assisted multi-input-single-output (MISO) system that serves multiple legitimate users (LUs) and eavesdroppers (Eves). The first stage formulates the system as a bipartite graph involving three types of nodes-RIS reflecting elements, LUs, and Eves-with the goal of generating the RIS phase shift matrix. The second stage models the system as a fully connected graph with two types of nodes (LUs and Eves), aiming to produce beamforming and artificial noise (AN) vectors. Both stages adopt an HGNN integrated with a multi-head attention mechanism, and the second stage incorporates two output methods: beam-direct and model-based approaches. The two-stage HGNN is trained in an unsupervised manner and designed to scale with the number of RIS reflecting elements, LUs, and Eves. Numerical results demonstrate that the proposed two-stage HGNN outperforms state-of-the-art GNNs in RIS-aided PLS scenarios. Compared with convex optimization algorithms, it reduces the average running time by three orders of magnitude with a performance loss of less than $4\%$. Additionally, the scalability of the two-stage HGNN is validated through extensive simulations.

A Hybrid Model-Assisted Approach for Path Loss Prediction in Suburban Scenarios

Accurate path loss prediction is crucial for wireless network planning and optimization in suburban environments with complex terrain variation and diverse land cover. This paper proposes a model assisted hybrid path loss prediction method that introduces an environment adaptive compensation on top of the classic close-in free-space reference distance (CI) path loss model. By jointly predicting the path loss exponent and a compensation term, the proposed approach dynamically adjusts the empirical trend. To improve the effectiveness of environmental representation, three environmental image organization schemes are constructed and evaluated. Experiments on measurement data collected in Pingtan Island show that the proposed method outperforms the CI model and a conventional model assisted baseline, achieving a test root mean square error of 4.04 dB.

Statistics Approximation-Enabled Distributed Beamforming for Cell-Free Massive MIMO

We study a distributed beamforming approach for cell-free massive multiple-input multiple-output networks, referred to as Global Statistics \& Local Instantaneous information-based minimum mean-square error (GSLI-MMSE). The scenario with multi-antenna access points (APs) is considered over three different channel models: correlated Rician fading with fixed or random line-of-sight (LoS) phase-shifts, and correlated Rayleigh fading. With the aid of matrix inversion derivations, we can construct the conventional MMSE combining from the perspective of each AP, where global instantaneous information is involved. Then, for an arbitrary AP, we apply the statistics approximation methodology to approximate instantaneous terms related to other APs by channel statistics to construct the distributed combining scheme at each AP with local instantaneous information and global statistics. With the aid of uplink-downlink duality, we derive the respective GSLI-MMSE precoding schemes. Numerical results showcase that the proposed GSLI-MMSE scheme demonstrates performance comparable to the optimal centralized MMSE scheme, under the stable LoS conditions, e.g., with static users having Rician fading with a fixed LoS path.