arXiv subjects

Yujie Zhao

Publications and source records attributed to Yujie Zhao.

At least 19 recordsLinked to original sources

arXiv2026

GML-Based Optimization for Movable Antenna Wireless Networks: Challenges and Opportunities

Movable antenna (MA) is proposed as an emerging technology for future wireless networks. By leveraging the additional spatial degrees of freedom, MA can proactively reshape the wireless propagation environment, thereby enhancing network performance.However, fully unlocking the potential of MA networks necessitates the joint optimization of MA antenna positioning and beamforming. For this non-convex and highly coupled problem, existing solutions exhibit significant limitations. Therefore, this paper proposes a gradient-based meta learning (GML) optimization framework. Specifically, we first elaborate on the hardware architecture and channel characteristics of MA, based on which we analyze the primary challenges in optimizing MA wireless networks. Subsequently, we introduce the fundamental logic of the GML framework and compare it with existing methods. Furthermore, we discuss the constraint handling strategies for applying the proposed optimization framework to MA networks. A specific case is studied to show the performance of proposed framework based on numerical simulation. Finally, this paper outlines future research directions for both the GML framework and MA wireless networks.

eess.SP↗

arXiv2026

Zero-WAM: In-Context World-Action Modeling from Human Videos for Open-Ended Task Generalization

Zero-shot cross-task generalization, where a policy must execute manipulation tasks never seen during training, remains a central challenge in robot learning. In large language models, a novel task can be performed simply by specifying it in the context, without any parameter update. This form of in-context learning (ICL) turns generalization into a problem of task specification. To achieve cross-task generalization, we bring this paradigm to robotic manipulation, and argue that the natural task specification for manipulation is a human video: unlike language, it provides rich visual cues about the intended task evolution. We present Zero-WAM, a causal video-action model that executes unseen tasks by following in-context human video guidance. To address the scarcity of task-rich paired human-robot data, we propose an automatic pipeline that converts task-sampled robot trajectories into semantically matched human videos, yielding HumanGen, a dataset of 74.2K human-robot ICL pairs across 8.6K tasks. For model training, we further introduce an in-context future chunk prediction (IFP) objective that suppresses shortcuts learned from seen tasks and forces the policy to draw task information from the video prompt. On seven unseen tasks in RoboTwin 2.0 simulation, Zero-WAM achieves a 47.0% average success rate, an absolute improvement of 29.5 percentage points over the strongest video-action baseline. In real-world evaluations, it follows human video guidance to generalize to unseen task configurations involving multi-object scenes, long-horizon manipulation, and fine-grained insertion.

cs.RO↗

arXiv2026

Antenna Positioning and Beamforming Optimization in MA Enabled Secure ISAC Systems: A Gradient-Based Meta Learning Approach

Integrated sensing and communications (ISAC) significantly improves spectral efficiency but introduces security risks regarding the interception of embedded communication signals. This paper proposes an movable antenna (MA)-enabled secure ISAC system that utilizes the spatial degrees of freedom of MA to mitigate these risks. Then, a problem is formulated to maximize the system secrecy rate by jointly optimizing antenna positioning, transmit beamforming, and artificial noise. However, the principal challenge arises from the non-convexity of the optimization problem and the strong coupling of the optimization variables. Generally, traditional optimization methods for this problem suffer from complex mathematical derivations, while existing deep learning approaches rely heavily on the training data distribution. To address these issues, we introduce a gradient-based meta learning (GML) algorithm, which works without pre-training and demonstrates favorable performance. Specifically, the algorithm establishes a neural network for each optimization variable, where the gradient of the objective function with respect to the variable serves as the input, and the output of the network determines the variable's update step. By handling the constraints and constructing penalty terms, the global loss function is used to guide the optimization process. Extensive numerical simulations confirm that the proposed algorithm achieves satisfactory performance in terms of both communication security and sensing capabilities.

eess.SP↗

arXiv2026

Native Video-Action Pretraining for Generalizable Robot Control

The advent of video-action models offers a promising path for robot control. Nevertheless, we argue that repurposing video generative models designed for digital content creation is inherently inadequate for physical environments. To bridge this gap, we present LingBot-VA 2.0, a video-action foundation model built from the ground up for embodiment. Four core design principles showcase its evolution from LingBot-VA. (1) Departing from traditional reconstruction-focused VAEs, we introduce a semantic visual-action tokenizer, which aligns visual representations with both semantics and actions, improving instruction following and action precision in subsequent policy learning. (2) Given the strictly causal nature of temporal dynamics, we adopt a causal pretraining paradigm, training from scratch to circumvent the catastrophic forgetting that frequently occurs when adapting bidirectional architectures. (3) To meet the demands of high-frequency inference, our model employs a sparse MoE backbone, expanding model capacity without compromising efficiency. (4) Real-time closed-loop control is realized through an enhanced asynchronous inference scheme, which predicts future latents in parallel with action execution while re-grounding each rollout on the latest observation via learned forward dynamics. Real-world deployment validates LingBot-VA 2.0 as a robust foundation model, as evidenced by its few-shot generalization across complex manipulation tasks.

cs.RO↗

arXiv2026

JetSpec: Breaking the Scaling Ceiling of Speculative Decoding with Parallel Tree Drafting

Speculative decoding (SD) accelerates autoregressive Large Language Models (LLMs) by drafting multiple tokens and verifying them in parallel, but it faces a scaling limitation: increasing the draft budget improves speed only when acceptance remains high and drafting overhead stays low. This ceiling has been difficult to break because prior head-based SD methods face a causality-efficiency dilemma. Autoregressive drafters produce path-conditioned candidates that are effective for tree speculative decoding with higher acceptance length, but their drafting cost grows with tree depth. Bidirectional block-diffusion drafters generate all positions in one pass, but their branch-agnostic marginals can form individually plausible yet mutually inconsistent trees, wasting budget and reducing acceptance. We propose JetSpec, a head-based SD framework that combines one-forward drafting efficiency with branch-wise causal conditioning. JetSpec trains a causal parallel draft head over fused hidden states from the frozen target model, producing candidate trees whose scores align with the target model's autoregressive factorization. This enables JetSpec to convert larger draft budgets into longer accepted prefixes and higher end-to-end speedup. Across math, coding, and chat benchmarks on dense and MoE Qwen3 models, JetSpec consistently outperforms bidirectional-head and tree-based SD baselines. On H100 GPUs, JetSpec achieves up to 9.64x speedup on MATH-500 and 4.58x on open-ended conversational workloads, with further latency gains demonstrated through vLLM integration under realistic serving loads. Our code and models are available at https://github.com/hao-ai-lab/JetSpec.

cs.CL↗

arXiv2026

Giant and Continuous Ionic Current Oscillation Induced by Dynamic Surface Charge Regulation in Cylindrical Mesopores

Nanofluidic ionic oscillators based on the dynamic regulation of surface charges hold great promise for neuromorphic computing, biosensing, and ionic circuits. Here, by dynamically adjusting the local charge inversion on pore walls, we present a simple and effective strategy to achieve periodic current oscillations by harnessing the transient adsorption and desorption of Ca2+ ions in cylindrical mesopores under concentration gradients. Based on the combined precision current measurements and multiphysics simulations, we demonstrate that local overadsorption of Ca2+ ions may induce asymmetric bipolar charge distributions along the pore axis, which periodically reverses the direction of electroosmotic flow and modulates the local ion concentration inside the pore, generating highly regular current oscillations. Notably, both the oscillation frequency and the open-state probability of the pore vary nearly linearly with the applied voltage. Moreover, under dynamic voltage scanning, the system exhibits typical memristive hysteresis, and the switching between the open and closed states is highly reproducible. This work not only reveals the dynamic, heterogeneous surface charge regulation by divalent ions, but also provides a simple, material agnostic method for constructing ionic oscillators and memristors based on dynamic adsorption/desorption of multivalent ions.

physics.chem-ph↗

arXiv2026

Quantum optimal control of steady orbits

Periodically driven dissipative systems can settle into steady orbits - fixed loops on their dynamical manifolds. In quantum mechanics, steady orbits occur in cooling engines (used to initialise quantum devices), coherent oscillators (such as lasers and masers), precision metrology devices (atomic clocks, optical and spin magnetometers), and magnetic resonance (steady state free precession, dynamic nuclear polarisation). Steady orbits and stroboscopic steady states are a promising target for quantum optimal control, but the numerical complexity is prohibitive: the infinite loop defeats gradient ascent pulse engineering (GRAPE) which relies on explicit numerical propagation in the time domain. Here we propose an efficient quantum control strategy for stroboscopic steady states and limit cycles that are approached asymptotically when a control sequence is repeated infinitely many times. The formalism is different from Floquet-Lindblad state engineering and effective Hamiltonian theories: it finds control sequences that drive a dissipative quantum system towards a steady orbit passing through user-specified waypoints. The software implementation (same numerical complexity scaling as GRAPE) is done for the Spinach library.

quant-ph↗

arXiv2026

AMA-Bench: Evaluating Long-Horizon Memory for Agentic Applications

Large Language Models (LLMs) are increasingly used as autonomous agents in complex, long-horizon applications, where effective memory is critical for sustained performance. Yet existing memory benchmarks are largely dialogue-centric, while real agent memory consists of continuous agent-environment interaction trajectories composed of states, actions, observations, and tool outputs. To address this gap, we introduce **AMA-Bench** (**A**gent **M**emory with **A**ny length), a benchmark for evaluating long-horizon memory in realistic agentic settings. AMA-Bench combines real-world agent trajectories from representative applications with expert-curated QA, as well as synthetic trajectories that scale to arbitrary horizons with rule-based QA. Our study shows that existing memory systems underperform because they fail to capture causal and objective information and rely heavily on lossy similarity-based retrieval. We further propose **AMA-Agent**, a memory system based on causality-graph construction and tool-augmented retrieval. AMA-Agent achieves **57.22%** accuracy on AMA-Bench, outperforming the strongest baseline by **11.16%**. Resources are available at: [https://ama-bench.github.io/](https://ama-bench.github.io/).

cs.AI↗

arXiv2026

AgentKVShift: Efficient KV Cache Reuse for Agentic Memory Systems

Memory-augmented LLM agents maintain context across hundreds of interactions through agentic memory systems that actively curate retrieved content with LLM-generated metadata such as summaries, keywords, and tags. From an inference cost standpoint, every retrieval triggers a full re-encoding of these structured memory units into Key-Value (KV) states, which dominates prefill latency. Existing training-free KV reuse methods mitigate this by selectively recomputing a small fraction of tokens, but were designed for RAG-style raw passages and degrade on structured agentic memories. We present AgentKVShift, a training-free, probe-guided KV residual correction method that operates per retrieved memory unit. One of the crucial insights we demonstrate is that the per-memory KV reuse residual decomposes into a shared memory-level offset plus small token-wise fluctuations. Estimating this offset from a small probe set allows us to correct every reused token by a single weighted correction. Unlike prior reuse methods which decide which tokens to recompute and leave the rest of the cache stale, AgentKVShift also corrects the tokens it does not recompute, turning the refresh budget into useful signal across the entire chunk. Across four open-source LLMs spanning 3B to 32B parameters and two long-horizon agentic memory benchmarks (long-term dialogue and agentic applications), AgentKVShift achieves near full recompute performance while refreshing only 10-30% of the cache, outperforming baselines at the same recompute ratio. It requires up to 5x lower recompute to reach this near-full performance, which prior reuse methods only attain at 45-55% refresh. In this regime, AgentKVShift delivers prefill speedups of 2-3.5x over no-KV-reuse on a single A100. AgentKVShift orthogonally composes with KV cache quantization, retaining over 2x the F1 of prior reuse methods under aggressive 2- and 4-bit settings.

cs.AI↗

arXiv2026

MetaAgent-X : Breaking the Ceiling of Automatic Multi-Agent Systems via End-to-End Reinforcement Learning

Automatic multi-agent systems aim to instantiate agent workflows without relying on manually designed or fixed orchestration. However, existing automatic MAS approaches remain only partially adaptive: they either perform training-free test-time search or optimize the meta-level designer while keeping downstream execution agents frozen, which creating a frozen-executor ceiling and leaving the end-to-end training of self-designing and self-executing agentic models unexplored. To address this, we introduce MetaAgent-X, an end-to-end reinforcement learning framework that jointly optimizes automatic MAS design and execution. MetaAgent-X enables script-based MAS generation, execution rollout collection, and credit assignment for both designer and executor trajectories. To support stable and scalable optimization, we propose Executor Designer Hierarchical Rollout and Stagewise Co-evolution to improve training stability and expose the dynamics of designer-executor co-evolution. MetaAgent-X consistently outperforms existing automatic MAS baselines, achieving up to 21.7% gains. Comprehensive ablations show that both designer and executor improve throughout training, and that effective automatic MAS learning follows a stagewise co-evolution process. These results establish end-to-end trainable automatic MAS as a practical paradigm for building self-designing and self-executing agentic models.

cs.AI↗

arXiv2026

ChipMATE: Multi-Agent Training via Reinforcement Learning for Enhanced RTL Generation

Existing API-based agentic systems for RTL code generation are fundamentally misaligned with industrial practice: they assume a golden testbench is available at generation time, rely on closed-source APIs incompatible with chip vendors' air-gapped security requirements, and cannot be trained on vendors' proprietary RTL codebases, leaving valuable internal data unused. Recent self-trained models address the deployment constraint but remain single-turn generators that overlook the critical role of verification in real industrial flows. To bridge these gaps, we present ChipMATE, the first self-trained multi-agent framework for RTL generation. Inspired by industrial practice where correctness emerges from cross-comparison between independently written RTL modules and reference models, ChipMATE pairs a Verilog agent with a Python reference-model agent that mutually verify each other's outputs without any golden oracle. We design a backtrack-based inference workflow to prevent error propagation across turns, and a two-stage training pipeline that first trains each agent individually to saturate its code-generation capability, then trains the team jointly to collaborate effectively. To support the training, we further build a hybrid data-generation framework that produces 64.4K high-quality reference model training samples. ChipMATE achieves 75.0\% and 80.1\% pass@1 on VerilogEval V2 with 4B and 9B base models, outperforming all existing self-trained models and even DeepSeek V4 with 1600B parameters. Our code and model weights are publicly available in https://github.com/zhongkaiyu/ChipMATE.

cs.MA↗

arXiv2026

Multi-Agent Memory from a Computer Architecture Perspective: Visions and Challenges Ahead

As LLM agents evolve into collaborative multi-agent systems, their memory requirements grow rapidly in complexity. This position paper frames multi-agent memory as a computer architecture problem. We distinguish shared and distributed memory paradigms, propose a three-layer memory hierarchy (I/O, cache, and memory), and identify two critical protocol gaps: cache sharing across agents and structured memory access control. We argue that the most pressing open challenge is multi-agent memory consistency. Our architectural framing provides a foundation for building reliable, scalable multi-agent systems.

cs.AR↗

arXiv2026

Real2Edit2Real: Generating Robotic Demonstrations via a 3D Control Interface

Recent progress in robot learning has been driven by large-scale datasets and powerful visuomotor policy architectures, yet policy robustness remains limited by the substantial cost of collecting diverse demonstrations, particularly for spatial generalization in manipulation tasks. To reduce repetitive data collection, we present Real2Edit2Real, a framework that generates new demonstrations by bridging 3D editability with 2D visual data through a 3D control interface. Our approach first reconstructs scene geometry from multi-view RGB observations with a metric-scale 3D reconstruction model. Based on the reconstructed geometry, we perform depth-reliable 3D editing on point clouds to generate new manipulation trajectories while geometrically correcting the robot poses to recover physically consistent depth, which serves as a reliable condition for synthesizing new demonstrations. Finally, we propose a multi-conditional video generation model guided by depth as the primary control signal, together with action, edge, and ray maps, to synthesize spatially augmented multi-view manipulation videos. Experiments on four real-world manipulation tasks demonstrate that policies trained on data generated from only 1-5 source demonstrations can match or outperform those trained on 50 real-world demonstrations, improving data efficiency by up to 10-50x. Moreover, experimental results on height and texture editing demonstrate the framework's flexibility and extensibility, indicating its potential to serve as a unified data generation framework. Project website is https://real2edit2real.github.io/.

cs.RO↗

arXiv2026

Extracting and analyzing 3D histomorphometric features related to perineural and lymphovascular invasion in prostate cancer

Diagnostic grading of prostate cancer (PCa) relies on the examination of 2D histology sections. However, the limited sampling of specimens afforded by 2D histopathology, and ambiguities when viewing 2D cross-sections, can lead to suboptimal treatment decisions. Recent studies have shown that 3D histomorphometric analysis of glands and nuclei can improve PCa risk assessment compared to analogous 2D features. Here, we expand on these efforts by developing an analytical pipeline to extract 3D features related to perineural invasion (PNI) and lymphovascular invasion (LVI), which correlate with poor prognosis for a variety of cancers. A 3D segmentation model (nnU-Net) was trained to segment nerves and vessels in 3D datasets of archived prostatectomy specimens that were optically cleared, labeled with a fluorescent analog of H&E, and imaged with open-top light-sheet (OTLS) microscopy. PNI- and LVI-related features, including metrics describing cancer-nerve and cancer-vessel proximity, were then extracted based on the 3D nerve/vessel segmentation masks in conjunction with 3D masks of cancer-enriched regions. As a preliminary exploration of the prognostic value of these features, we trained a supervised machine learning classifier to predict 5-year biochemical recurrence (BCR) outcomes, finding that 3D PNI-related features are moderately prognostic and outperform 2D PNI-related features (AUC = 0.71 vs. 0.52). Source code is available at https://github.com/sarahrahsl/SegCIA.git.

cs.CV↗

arXiv2026

Stronger-MAS: Multi-Agent Reinforcement Learning for Collaborative LLMs

Multi-agent systems (MAS) and reinforcement learning (RL) are widely used to enhance the agentic capabilities of large language models (LLMs). MAS improves task performance through role-based orchestration, while RL uses environmental rewards to learn stronger policies, such as GRPO-style optimization. However, applying on-policy RL to MAS remains underexplored and presents unique challenges. Algorithmically, standard GRPO grouping assumptions break down because prompts vary by role and by turn. System-wise, the training stack must support MAS-workflow rollouts and on-policy updates for both single-policy and multi-policy models. We propose AT-GRPO, which includes (i) an agent- and turn-wise grouped RL algorithm tailored to MAS and (ii) a training system that supports both single- and multi-policy regimes. Across game, planning, coding, and math tasks, AT-GRPO delivers substantial gains. On long-horizon planning, it increases accuracy from a 14.0 to 47.0 percent single-agent RL baseline to 96.0 to 99.5 percent. It also improves reasoning performance, with average gains of 3.87 to 7.62 percent on coding tasks and 9.0 to 17.93 percent on math. Code and environments are available at: https://github.com/pettingllms-ai/PettingLLMs.

cs.LG↗

arXiv2026

The Impact of Interference Cognition on the Reliability and Capacity of Industrial Wireless Communications

Interference significantly impacts the performance of industrial wireless networks, particularly n severe interference environments with dense networks reusing spectrum resources intensively. Although delicate interference information is often unavailable in conventional networks, emerging interference cognition techniques can compensate this critical problem with possibly different precision. This paper investigates the relationship between precision of interference cognition and system performance. We propose a novel performance analysis framework that quantifies the impact of varying interference information precision on achievable rate. Specifically, leveraging the Nakagami-$\mathbf{m}$ fading channel model, we analytically and asymptotically analyze the average achievable rate in the finite blocklength regime for different precision levels of signal and interference information. Our findings reveal the critical importance of identifying per-link interference information for achieving optimal performance. Additionally, obtaining instantaneous information is more beneficial for signal links.

eess.SY↗

arXiv2025

TwinAligner: Visual-Dynamic Alignment Empowers Physics-aware Real2Sim2Real for Robotic Manipulation

The robotics field is evolving towards data-driven, end-to-end learning, inspired by multimodal large models. However, reliance on expensive real-world data limits progress. Simulators offer cost-effective alternatives, but the gap between simulation and reality challenges effective policy transfer. This paper introduces TwinAligner, a novel Real2Sim2Real system that addresses both visual and dynamic gaps. The visual alignment module achieves pixel-level alignment through SDF reconstruction and editable 3DGS rendering, while the dynamic alignment module ensures dynamic consistency by identifying rigid physics from robot-object interaction. TwinAligner improves robot learning by providing scalable data collection and establishing a trustworthy iterative cycle, accelerating algorithm development. Quantitative evaluations highlight TwinAligner's strong capabilities in visual and dynamic real-to-sim alignment. This system enables policies trained in simulation to achieve strong zero-shot generalization to the real world. The high consistency between real-world and simulated policy performance underscores TwinAligner's potential to advance scalable robot learning. Code and data will be released on https://twin-aligner.github.io

cs.RO↗

arXiv2025

PRO-V-R1: Reasoning Enhanced Programming Agent for RTL Verification

Register-Transfer Level (RTL) verification is a primary bottleneck, consuming 60-70% of development time. While Large Language Models (LLMs) show promise for RTL automation, their performance and research focus have overwhelmingly centered on RTL generation rather than verification. Current methods for RTL verification rely on large scale proprietary models (e.g., GPT-4o) to generate Python-based functional references, incurring a high cost and raising data-privacy risks. To date, an end-to-end open-source solution for autonomous verification remains absent. We introduce PRO-V-R1, the first trainable open-source agentic framework for autonomous RTL verification. Our contributions are threefold: (1) we design PRO-V sys, a modular agentic system that couples LLM-based reasoning with programmatic tool use for RTL verification; (2) we establish a data construction pipeline that leverages existing RTL datasets to build simulation-validated, expert-level trajectories tailored for supervised fine-tuning (SFT) RTL verification agents; and (3) we implement an efficient reinforcement learning (RL) algorithm that uses verification-specific rewards derived from program-tool feedback to optimize the end-to-end verification workflow. Our empirical evaluation demonstrates PRO-V-R1 achieves a 57.7% functional correctness rate and 34.0% in robust fault detection, significantly outperforming the base model's 25.7% and 21.8% (respectively) from the state-of-the-art (SOTA) automatic verification system. This configuration also outperforms large-scale proprietary LLMs in functional correctness and shows comparable robustness for fault detection.

cs.AI↗