Searcharxiv⌕ Search

arXiv subjects

Yuning Zhang

Publications and source records attributed to Yuning Zhang.

At least 19 recordsLinked to original sources

Project Progress Desk: Know Where All Your Claude Code and Codex Projects Stand at a Glance

Agentic coding tools such as Codex and Claude Code let developers work on many projects in parallel, but keeping track of them still means switching between chats, terminal sessions and providers. We present Project Progress Desk, a local desktop interface that brings overall progress, blockers and pending decisions into one view. Users can see which projects need attention and what remains, and compare earlier approaches before choosing how to continue. Agents report each result in a short receipt; the Desk checks it against the project's checklist and moves the bar, showing earlier estimates in a lighter shade. We describe daily use across 36 projects, including a choice between two implementations based on their recorded outcomes. At the demo, attendees will explore 30 simulated projects, trace completed work to overall progress, and watch a coding agent's new result update the overview.

cs.HC↗

Where to Generate Touch: Do Image Latents Preserve Material Recognition?

Generated touch could extend robot learning beyond recorded contacts. A tactile generator needs a latent space, and we choose one by testing whether a material classifier trained on real GelSight presses still works on encode-decode reconstructions from pretrained image models. On ObjectFolder Real, the classifier loses 6.5 to 9.7 points on reconstructions from three 16x discrete tokenizers and 0.3 points on those from the Stable Diffusion VAE (SD-VAE). The 16x tokenizers again lose far more on Touch and Go and on DIGIT object recognition (YCB-Slide). Retraining the classifier on reconstructions, with no new labels, narrows the discrete gap. SD-VAE reconstructions also keep object identity, with 76.7% object-matching accuracy against 78.9% for real touch and 67.0% for MAGVIT-v2. We therefore select SD-VAE latents and show a photo-to-touch generator trained in them. Code will be released upon acceptance.

cs.CV↗

AgentServe: Algorithm-System Co-Design for Efficient Agentic AI Serving on a Consumer-Grade GPU

Large language models (LLMs) are increasingly deployed as AI agents that operate in short reasoning-action loops, interleaving model computation with external calls. Unlike traditional chat applications, these agentic workloads require inference serving systems to balance low latency, stable token emission, and throughput under multiple request arrivals from different AI agents. Recent deployments highlight a shift toward running small language models (SLMs) locally on consumer-grade GPUs, driven by privacy, compliance, and cost constraints. When heterogeneous requests overlap on a single GPU, long prefills and short decodes contend for resources, creating head-of-line blocking that destabilizes interactive performance. By analyzing agent workloads, we observe that their execution naturally separates into cold prefills, which process long system prompts, resume prefills, which append tool outputs to cached contexts, and short decodes, which are latency-critical. This mix intensifies contention compared to conventional chatbot serving. We present AgentServe, a single-GPU serving system that ensures stable multi-agent execution under such conditions by isolating prefills from decodes, applying dynamic budgeting to resume prefills, and allocating GPU resources through pre-established CUDA Green Context slots with adaptive control. Evaluation results show that AgentServe significantly improves latency stability while sustaining competitive throughput, achieving up to 2.8x TTFT improvement and 2.7x TPOT improvement over state-of-the-art baselines across different settings.

cs.DC↗

HeteroVDiff: Pipelining Video Diffusion across Mobile NPU and GPU with Shared Weights

Video generation is moving from the cloud to phones for privacy and offline use, and applications often ask for several clips at a time, so on-device video generation is a throughput problem before it is a latency problem. Contemporary mobile SoCs carry both a GPU and a neural processing unit (NPU) on shared DRAM, yet existing on-device video diffusion systems run the denoiser on one engine while the other sits idle or only decodes the finished latent afterwards. Heterogeneous designs built for mobile LLMs do not carry over: they synchronise the two engines at every operator and partition weight tensors between them instead of sharing one stored copy. We first measure the platform limits that decide what can run on this class of SoC: how large one buffer may be, whether both runtimes can map the same buffer, and how much activation memory the NPU toolchain lets one graph hold. Guided by these measurements, we store the base weights once in INT8 in the layout the NPU reads natively; the GPU consumes the same bytes directly as integer dot-product operands, so neither runtime rearranges weights on load. We then flatten the denoising trajectory into its full sequence of block executions and cut it once at a block boundary, so the GPU runs the prefix and the NPU the suffix. The cut is placed from measured rather than modelled rates, including the decoder that only the NPU can run, and taken while both engines run together because each engine's speed changes when the other is active. We present HeteroVDiff, which serves a multi-expert video diffusion model across the NPU and GPU with 48% less weight memory than giving each engine its own copy. Generating videos back to back on a Galaxy S25 with decoding inside the pipeline, HeteroVDiff sustains 1.35x the stream throughput of NPU-only execution, and on VBench-2.0 it matches FP32 in the total score over all 18 dimensions.

cs.DC↗

Auditing Action Settlement in LLM Agent Environments: Order, Progress, and Replay

Concurrent actions in large language model (LLM) agent environments require arbitration even when each proposal is individually valid. We implement a typed snapshot-settlement contract and audit three distinct properties: order sensitivity, useful progress, and replay consistency. Five settlement policies are tested in 28,800 exhaustive permutation trials and 2,160 scripted multistep episodes. Joint policies are spatially order-invariant conditional on fixed priorities, yet conservative rejection completes only 31.25% of agents in a six-agent doorway task versus 90.28% for random tickets; the paired improvement is 59.03 percentage points (95% bootstrap interval: 50.00-68.06). All policies preserve the tested spatial constraints, and priority arbitration still misses the independent small-instance optimum. A separate full-state journal audit exactly replays 156 checkpoints and rejects 1,332 constructed corruptions with a retained terminal anchor. The evidence concerns execution semantics, not human realism or long-run fairness.

cs.AI↗

When Does Exercise-Specific Joint Selection Help? An Audit of Evaluation and Control Design

Exercise-specific joint selection can improve skeleton-based correctness classification, but what does that gain establish? We audit 1,057 repetitions from ten REHAB24-6 subjects, separating evaluation aggregation, subset structure, and temporal representation. The manual-subset kNN gain changes from 0.055 for pooled out-of-fold AUROC to 0.020 for equal-weight within-person AUROC; both paired intervals include zero. Among 1,000 dimension-matched random maps, 14 match or exceed the manual pooled result, versus 145 when bilateral structure and trunk inclusion are also matched. RBF-SVM retains a positive within-person gain, whereas logistic regression and a random-convolution comparator have negative point gains under that estimand. Sequence-order and paired-seed controls further qualify the interpretation. This exploratory audit shows why joint-selection claims require explicit estimands and structurally appropriate controls; it does not establish a new algorithm or clinical benefit.

cs.AI↗

WA-SpecDec: World-Aware Speculative Decoding for Vision-Language-Action Models

Vision-language-action (VLA) policies generate robot controls autoregressively, making closed-loop latency dominated by repeated target-model forward passes. Speculative decoding reduces this cost by verifying blocks of draft action tokens in parallel, and recent VLA methods further relax token-level acceptance because small differences in action-token space often map to similar continuous controls. However, this relaxation remains scene-agnostic. A fixed token-distance tolerance treats the same action-token deviation as equally safe across states, although deviations that are harmless in free space can cause collisions or grasp failures near contact. We propose WA-SpecDec, a world-aware speculative decoding framework that injects world-model-derived physical scene awareness during the VLA prefill stage, producing shared world-aware prefill states for draft proposal and target verification without changing the relaxed acceptance rule. Across three state-of-the-art relaxed acceptance schemes, WA-SpecDec preserves higher task success under looser relaxation and enables longer accepted prefixes. At comparable-success operating points, WA-SpecDec achieves a 1.5x matched-success speedup over VLA speculative decoding alone and reduces near-contact failure (NCF) by 18.6% on average relative to the corresponding speculative baselines.

cs.RO↗

HorizonServe: Coordinating Request Scheduling with GPU Sharing for Omni-Model Serving

Omni models unify text, speech, image, and multimodal reasoning in a single serving backend, but this unified deployment exposes a new scheduling problem. Requests with different output modalities may share an initial multimodal backbone and then diverge into downstream generation stages, creating heterogeneous first-response metrics and service-level objective (SLO) targets on the same GPU. Existing large language model (LLM) and multimodal serving systems mainly optimize token progress or input-side processing, and they do not jointly control temporal sharing in the shared stage and spatial sharing among co-running stages. This paper presents HorizonServe, a single-GPU omni-model serving system that coordinates request admission and GPU allocation under heterogeneous SLOs. HorizonServe profiles per-class first-response latency, protects requests with limited slack, rotates shared-stage opportunities across execution paths, and throttles the shared-stage streaming multiprocessor (SM) allocation when downstream stages are active. Across three omni-model workloads and two GPU platforms, HorizonServe improves SLO attainment by up to 4.9$\times$ in arrival-rate sweeps and 7.0$\times$ under downstream-heavy traffic, and reduces per-class first-response latency by 38.4--63.7\%.

cs.DC↗

TacGen: Touch Is a Necessary Dimension of Physical-World Representation -- Addressing Tactile Data Scarcity with Scalable Vision-to-Touch Alignment and Generation

Touch resolves the physical-property ambiguity left by vision: exploratory contact recovers shape, texture, compliance, and material, and visuo-haptic object representations converge in ventral visual cortex. We ask whether representation learning can reproduce this grounding. TacGen mitigates the tactile-data scarcity bottleneck by combining pre-specified V+T contrastive alignment with a latent-space residual-MLP V->T generator that synthesizes tactile latents from RGB for tactile-data scaling. With matched DINOv2 backbones, splits, and probes, V+T improves matched V-only on mass (Delta R^2=+0.570), density (Delta acc=+0.067), hardness (+0.117), and uncertainty-banded force labels (Delta R^2=+0.281); all CIs exclude zero. The same representation lifts matched-capacity TACTO manipulation 0.246->0.979 while V-only capacity scaling accounts for only 4.5% of the gap, preserving 95.5%. The generator reaches cross-seed +0.589, with real tactile +0.585 inside the seed interval; the architecture comparison shows a 13pp downstream gap between reconstruction quality and representation utility. Across five-seed SSVTP/TVL reproductions, YCB-Sight transfer, three-backbone checks, permutation/random-feature controls, hash-verified manifests, and measured-force validation checks, the evidence supports the claim that touch supplies a necessary physical evidence channel for representations of contact-dependent properties.

cs.RO↗

OmniLoc: A Geometry-Aware Foundation Model for Anchor-Free UE Localization Across Diverse Indoor Environments

Indoor localization from wireless measurements remains challenging in large-scale deployments due to substantial variation in building geometry, the set of detectable access points (APs), and the heterogeneity of received signals. Existing learning-based methods often perform well only in limited settings and degrade under environmental shifts, making robust anchor-free localization across diverse indoor environments notoriously difficult. In this paper, we present OmniLoc, an environment-interactive foundation model for anchor-free user equipment localization across diverse indoor environments. To the best of our knowledge, OmniLoc is the first foundation-model-based approach built directly on wireless measurements for this task. OmniLoc is built on three key designs. First, a unified input tokenization module converts heterogeneous wireless measurements into a common representation that is more amenable to learning. Second, a geometry-aware Transformer performs AP-aware feature extraction by emphasizing dominant APs while aggregating complementary evidence from supporting APs. Third, a geometry-aware location estimation module conditions regression on geometric embeddings to produce geometrically consistent location predictions. We evaluate OmniLoc on both a large-scale in-house dataset and a public benchmark dataset. Results show that OmniLoc significantly outperforms existing methods, consistently improves existing backbones when its design components are integrated, and demonstrates strong generalization in cross-environment evaluations.

cs.LG↗

TextFake: Benchmarking AI-Generated Image Detection on Text-Rich Images

Recent AI-generated image (AIGI) detectors perform well on natural-image benchmarks, but their behavior on text-rich forgeries, such as fabricated screenshots, documents, and news pages prevalent in misinformation, remains untested. We introduce TextFake, a 20,000-image benchmark for text-rich AIGI detection spanning 28 languages, 4 topic categories, and 2 scene modalities. Fake images are synthesized via a four-stage pipeline that annotates real images along three controlled dimensions and generates counterparts through distribution-aligned structured prompting, ruling out covariate shortcuts. Zero-shot evaluation of 14 specialized detectors and 3 frontier VLM APIs reveals a large systematic gap: no method exceeds 80% accuracy, with some dropping over 60% from natural-image benchmarks. Diagnostic evaluations identify three failure modes: the Text Density Curse, where dense glyphs overwhelm low-level detectors; Cloaking via Rendering Fidelity, where stronger text rendering suppresses enerative artifacts; and Threshold Collapse, where routine perturbations drive detectors toward chance-level performance.

cs.CV↗

WiLoc: Massive Measured Dataset of Wi-Fi Channel State Information with Application to Machine-Learning Based Localization

Localization is a key component of the wireless ecosystem. Machine learning (ML)-based localization using channel state information (CSI) is one of the most popular methods for achieving high-accuracy localization with low cost. However, to be accurate and robust, ML-based algorithms need to be trained and tested with large amounts of data, covering not only many user equipment (UE)/target locations, but also many different access points (APs) locations to which the UEs connect, in a variety of different environment types. This paper presents a massive-sized CSI dataset, WiLoc (Wi-Fi Localization), and makes it publicly available. WiLoc is obtained by a series of precision measurement campaigns that span three months, and it is massive in all the above-mentioned three dimensions: > 12 million UE locations, > 3,000 APs, covering 16 buildings for indoor localization, and > 30 streets for outdoor use. The paper describes the dataset structure, measurement environments, measurement protocols, and the dataset validations. Comprehensive case studies validate the advantages of large datasets in ML-driven localization strategies for both "standard" and transfer learning. We envision this dataset, which is by far the largest of its kind, to become a standard resource for researchers in the field of ML-based localization.

eess.SP↗

When Agents Look the Same: Quantifying Distillation-Induced Similarity in Tool-Use Behaviors

Model distillation is a primary driver behind the rapid progress of LLM agents, yet it often leads to behavioral homogenization. Many emerging agents share nearly identical reasoning steps and failure modes, suggesting they may be distilled echoes of a few dominant teachers. Existing metrics, however, fail to distinguish mandatory behaviors required for task success from non-mandatory patterns that reflect a model's autonomous preferences. We propose two complementary metrics to isolate non-mandatory behavioral patterns: \textbf{Response Pattern Similarity (RPS)} for verbal alignment and \textbf{Action Graph Similarity (AGS)} for tool-use habits modeled as directed graphs. Evaluating 18 models from 8 providers on $τ$-Bench and $τ^2$-Bench against Claude Sonnet 4.5 (thinking), we find that within-family model pairs score 5.9 pp higher in AGS than cross-family pairs, and that Kimi-K2 (thinking) reaches 82.6\% $S_{\text{node}}$ and 94.7\% $S_{\text{dep}}$, exceeding Anthropic's own Opus 4.1. A controlled distillation experiment further confirms that AGS distinguishes teacher-specific convergence from general improvement. RPS and AGS capture distinct behavioral dimensions (Pearson $r$ = 0.491), providing complementary diagnostic signals for behavioral convergence in the agent ecosystem. Our code is available at https://github.com/Syuchin/AgentEcho.

cs.CL↗

DuoServe-MoE: Dual-Phase Expert Prefetch and Caching for LLM Inference QoS Assurance

Large Language Models (LLMs) are increasingly deployed as Internet/Web services (LLM-as-a-Service) with strict latency Service-Level Objectives (SLOs) under tight GPU memory budgets. Mixture-of-Experts (MoE) models improve quality and throughput via sparse expert activation, but serving them efficiently is challenging because expert weights dominate memory footprint and incur costly host--device transfers when offloaded. Moreover, MoE serving exhibits a phase disparity: the prefill phase tends to activate experts densely across many tokens, while the decode phase activates only a few experts per step. A uniform expert loading/caching policy across phases leads to either peak-memory blowup (prefill) or tail-latency inflation (decode). We present DuoServe-MoE, a QoS-oriented MoE serving system that decouples prefill and decode and applies phase-specialized expert scheduling. For prefill, DuoServe-MoE uses a two-stream CUDA pipeline to overlap expert prefetching with non-MoE computation, reducing expert residency time and peak GPU memory. For decode, it employs a lightweight layer-level predictor trained offline from activation traces to prefetch only likely experts without model changes. Experiments on representative MoE LLMs show that DuoServe-MoE improves TTFT by up to $5.34\times$ and end-to-end latency by up to $7.55\times$ over representative baselines, while maintaining low runtime GPU memory usage under resource-constrained deployment.

cs.DC↗

High Performance 4H-SiC Optically Controlled MOS Transistor

This paper introduces an optically controlled 4H-SiC MOSFET designed to avoid the gate-oxide interface unreliability and electromagnetic interference (EMI) susceptibility inherent in conventional voltage-driven devices. By replacing the conventional gate electrode with a semi-transparent optical window, the device enables direct modulation of channel conductivity through ultraviolet illumination. Electrical and optical characterization demonstrates that under an optical power density above 0.1 W/cm^2, the device achieves an on/off current ratio exceeding 10^6 between illuminated and dark states. Notably, at an optical power density of 0.031 W/cm^2, the photogenerated current density exceeds that obtained under a gate bias of 15 V in magnitude. Energy band analysis confirms that the optical switching mechanism operates through direct photogenerated carrier generation and transport, fundamentally differing from conventional gate voltage control and thus circumventing interface-trap and EMI-related limitations. Dynamic measurements further reveal fast switching capability, with a rise time of 1.44 ns. These results validate the feasibility of optically driven switching in SiC-based devices and highlight their potential for high-speed logic applications.

physics.app-ph↗

Cognibit: From Digital Exhaustion to Real-World Connection Through Gamified Territory Control and LLM-Powered Twin Networking

We present an LLM-powered social discovery platform that uses digital twins to autonomously evaluate interpersonal compatibility through behavioral simulation. The platform unifies three key pillars: (1) digital twins that engage in autonomous multi-turn conversations on behalf of users to estimate compatibility, (2) gamified territory conquest mechanics that incentivize real-world exploration and create organic settings for in-person encounters, and (3) AI companions that preserve persistent shared memory across devices. Built upon CogniPair's cognitive architecture (Ye et al., 2026), validated on the Columbia Speed Dating dataset (551 participants), our system extends prior simulation-only matching into a fully deployed social discovery environment. Through deployment, we derive empirical cost-quality baselines and identify fundamental scaling bottlenecks that remain hidden in component-level testing alone.

cs.HC↗

AI assisted optimization of integrated waveguide polarizers containing 2D reduced graphene oxide

Reduced graphene oxide (rGO) exhibits strong anisotropic light absorption and high compatibility with photonic integrated chips, making it a promising material for implementing high performance onchip polarization selective devices. The performance of rGO integrated waveguide polarizers is highly dependent on the waveguide geometry, and achieving optimal performance requires exploring a large parameter space, making conventional mode simulation methods computationally demanding. Here, we propose and demonstrate a machine learning framework based on fully connected neural networks (FCNNs) to map the dependence of the polarizer figure of merit (FOM) on the waveguide geometry. Once trained by using a small dataset of low resolution mode simulation results, the FCNN framework can rapidly and accurately predict FOM values across a large structural parameter space with high resolution. Results show that this method can reduce overall computing time by more than 4 orders of magnitude as compared to the mode simulation methods, and achieve high prediction accuracy with an average deviation (AD) below 0.05. These results highlight the FCNN based machine learning framework as an efficient tool for the design and optimization of rGO integrated waveguide polarizers.

physics.optics↗

AI based design of 2D material integrated optical polarizers

On-chip integration of highly anisotropic two-dimensional (2D) materials offers new opportunities for realizing high performance polarization selective devices. Obtaining optimized designs for such devices requires extensively sweeping large parameter spaces, which in conventional approaches relies on massive mode simulations that demand considerable computational resources. Here, we address this limitation by developing a machine learning (ML) model based on fully connected neural networks (FCNNs). Trained by using mode simulation results for low resolution structural parameters, the FCNN model can accurately predict polarizer figures of merits (FOMs) for high resolution parameters and rapidly map the global variation trend across the entire parameter space. We test the performance of the FCNN model using two types of polarizers with 2D graphene oxide (GO) and molybdenum disulfide (MoS2). Results show that, compared to conventional mode simulation approach, our approach can not only reduce the overall computing time by about 4 orders of magnitude, but also achieve highly accurate FOM predictions with an average deviation of less than 0.04. In addition, the measured FOM values for the fabricated devices show good agreement with the predicted ones, with discrepancies remaining below 0.2. These results validate artificial intelligence (AI) as an effective approach for designing and optimizing 2D-material based optical polarizers with high efficiency.

physics.optics↗