Searcharxiv⌕ Search

arXiv subjects

Hao Zheng

Publications and source records attributed to Hao Zheng.

At least 37 records · Page 2Linked to original sources
arXiv2026

A Progressive Approach to Synthesizable RTL Design Generation Using LLMs

Large language models can generate register-transfer-level (RTL) designs directly from natural language specifications. Their failures, however, arise mostly from understanding rather than coding \cite{zhang2026understanding, qiu2025towards}. A specification is informal and ambiguous, the model's interpretation stays implicit, and every misreading is committed silently into Verilog, where only simulation can expose it. Intermediate representations make the interpretation partly explicit, yet existing works don't verify the interpretation against the specification, and repair simulation failures at the code level regardless of where the misreading originated. VeriRefine instead treats specification refinement as a verifiable stage of RTL generation. It progressively refines the prose specification into an explicit, schema-constrained account of design intent, expressed as per-signal Abstract Signal Transition Functions (ASTFs) that commit each signal's logic style, clock domain, and reset behavior before any code exists and ground every behavior in a verbatim specification sentence. The refined specification then passes a five-layer audit spanning soundness, completeness, consistency, FSM integrity, and core RTL design rules, so interpretation errors are repaired at the representation level before any Verilog is generated. Once code is generated, each simulation failure is classified as an understanding error or a coding error and routed back to the corresponding stage for targeted repair. Because every signal's hardware class is fixed during refinement, synthesizability becomes a structural property of the pipeline rather than a post-hoc check. With Claude Sonnet 4.6, VeriRefine reaches 94.0\% functional correctness on RTLLM v2.0 and 98.1\% on VerilogEval-Human v2.

cs.AR↗
arXiv2026

Compositional Context Fine-Tuning Vision-Language Model for Complex Assembly Action Understanding from Videos

Assembly action understanding is a key enabler for effective human-robot collaborative assembly, yet it remains challenging due to subtle motions and fine-grained hand-object interactions. We adapt vision-language models (VLMs) to this challenging domain with Compositional Context Fine-Tuning (CCFT), a method that decomposes assembly actions into semantic elements (Verb, Object, Tool) and fine-tunes VLMs to recognize each action element using templated question-answering pairs. This approach ensures near-deterministic outputs. To enable efficient and effective multi-task learning under limited data, a Layer-Partitioned Alternating Training (LP-AT) method is presented, which assigns distinct model layers to recognize specific action elements through element-specific low-rank adapters. LP-AT alternates weight updates across element-specific adapters, reducing cross-task interference while enabling per-adapter hyperparameter optimization. Furthermore, we create HA-ViD-VQA and IKEA-ASM-VQA datasets from existing assembly video datasets. Extensive experiments on these datasets demonstrate that our method consistently outperforms strong action recognition baselines while providing interpretable element-level predictions that can support diverse downstream applications.

cs.CV↗
arXiv2026

MultiFair: Multimodal Balanced Fairness-Aware Medical Classification with Dual-Level Gradient Modulation

Medical decision systems increasingly rely on data from multiple sources to ensure reliable and unbiased diagnosis. However, existing multimodal learning models fail to achieve this goal because they often overlook two critical challenges. First, various data modalities may learn unevenly, thereby converging to a model biased towards certain modalities. Second, the model may emphasize learning on certain demographic groups causing unfair performances. The two aspects can influence each other, as different data modalities may favor respective groups during optimization, leading to both imbalanced and unfair multimodal learning. This paper proposes a novel approach called MultiFair for multimodal medical classification, which addresses these challenges with a dual-level gradient modulation process. MultiFair dynamically modulates training gradients regarding the optimization direction and magnitude at both data modality and group levels. We evaluate MultiFair on three real-world medical classification datasets with diverse demographic attributes,including multiclass classification and missing-modality settings. Experimental results demonstrate its effectiveness.

cs.LG↗
arXiv2026

Large-Norm Solutions and the Relaxation-Time Limit for Quantum Hydrodynamics on the Two-Dimensional Torus

This paper extends to the two-dimensional torus our previous analysis \cite{AMZ3} of weak solutions with large norms for the collisional quantum hydrodynamic (QHD) system in semiconductor modeling. We first establish the global well-posedness of weak solutions with strictly positive density within the functional framework of generalized chemical potential (GCP) solutions introduced in \cite{AMZ1}. Two key ingredients of the analysis are a logarithmic Sobolev-type inequality controlling oscillations of the density and a functional combining a higher-order energy with the physical entropy. This combined functional yields a coercive dissipation mechanism that allows us to establish stability and exponential convergence for solutions with large initial data. As a byproduct of our approach, we also prove the global existence of $H^2$ solutions for a nonlinear Schrödinger--Langevin equation. Finally, for GCP solutions with strictly positive density, we justify the relaxation-time limit and provide an explicit convergence rate. Our analysis relies on compactness techniques that do not require the existence or smoothness of solutions to the limiting system. Moreover, our results impose no well-preparedness assumptions on the initial data, thereby accommodating the possible formation of an initial layer.

math.AP↗
arXiv2026

Initial layer analysis of relaxation-time limit of the collisional QHD

We study the structure of the initial layer arising in the relaxation-time limit of the collisional quantum hydrodynamic (QHD) system. When the initial data are not well prepared, a fast transient regime appears near the initial time, which prevents the uniform-in-time convergence of the momentum density to its limiting value. Using the method of matched asymptotic expansions, we derive a systematic asymptotic description of the solution with respect to the relaxation-time parameter $τ$. In particular, we identify the fast time scale $t/τ^{2}$ governing the initial layer and explicitly construct the corresponding inner expansion for the momentum density together with the outer expansion describing the slow dynamics. The leading-order outer dynamics are shown to coincide with the quantum drift-diffusion equation. The asymptotic expansion is rigorously justified by establishing uniform in $τ$ energy estimates for the remainder terms under suitable regularity assumptions on the solutions. As a consequence, we prove the strong convergence of the momentum density in $L^\infty$ in time after subtracting the leading initial-layer correction. The analysis further shows that the convergence rate of order $τ$ is optimal for general initial data and explains the improved rate in the well-prepared case.

math.AP↗
arXiv2026

Enhanced detectability of axion's electromagnetic response with a RF-excited magnetic field in cavity

Haloscope is one of the typical installations to detect the electromagnetic responses (EMRs) of axion field in radio-frequency (RF) and microwave bands. Given that the detectable signals of the usual Haloscope-type detectors (HTDs), biased only by high stationary magnetic fields, are just the second axion-photon energy and thus are very weak, here we propose a feasible approach to significantly improve their sensitivity by additionally applying a transverse RF- or microwave modulated magnetic field to excite the cavity's magnetic resonant mode to produce the first-order axion-photon energy response signals. Accordingly, it can be argued that the achievable detection sensitivity of the upgrading HTD (i.e., UHTD) could be enhanced by $0.3\sim 1$ orders of magnitude, compared with that achieved by the existing HTDs without the transverse RF-excited magnetic field. The feasibility of the proposed UHTD is also discussed.

hep-ex↗
arXiv2026

Measurement-Based Characterization and Statistical Modeling of 6G Urban Low-Altitude A2G Channels across FR1 and FR3

Unmanned aerial vehicle (UAV) communications have been recognized as a key component of future sixth-generation (6G) space-air-ground-sea integrated networks. Accurate characterization and modeling of air-to-ground (A2G) channels are essential for the design and optimization of low-altitude communication systems. This paper presents a wideband A2G channel measurement campaign in an urban environment at 2.85 and 4.6~GHz in FR1 and 7.25~GHz in the FR3 frequency band, each with a bandwidth of 250~MHz. To enable reliable line-of-sight (LoS) and non-line-of-sight (NLoS) propagation state identification, a weakly supervised method is developed by fusing geometric priors, channel features, and spatial consistency constraints. Furthermore, based on the measured data, A2G channel characteristics are extracted and analyzed under LoS/NLoS conditions across different frequency bands, including path loss (PL), shadow fading (SF), power delay profile, root-mean-square delay spread (RMS-DS), and Rician $K$-factor. The results show that the close-in model fits the measured PL more accurately than the 3GPP reference model, and that NLoS propagation leads to larger path loss exponents and stronger SF than LoS propagation. For channel delay characteristics, higher-frequency channels exhibit fewer effective MPCs and weaker delay dispersion, indicating increased channel sparsity. Specifically, the mean RMS-DS under LoS conditions decreases from 93.11 to 46.84~ns, while the mean Rician $K$-factor increases from 9.16 to 12.88~dB. The statistical results further show that the RMS-DS and the Rician $K$-factor can be well characterized by lognormal and normal distributions, respectively. Moreover, the movement of the receiver in a complex scattering environment intensifies the spatial non-stationarity of the A2G channel.

eess.SP↗
arXiv2026

Neuro-Symbolic Safety Guidance for Vision-Language-Action Models via Constrained Flow Matching

Vision-Language-Action (VLA) models have demonstrated promising generalization capabilities across robotic manipulation tasks, yet their real-world deployment remains limited by the lack of effective safety measures. Specifically, existing safety measures only prevent collisions caused by the robot's next action. In this paper, we propose a neuro-symbolic safety guidance mechanism for flow matching based VLAs that enables predictive collision avoidance. Flow matching based VLAs determine the next actions by predicting a trajectory (a sequence of actions) through an iterative neural flow matching process. Our method formulates safety enforcement as a minimum-norm constrained optimization problem that corrects safety violations during the denoising process of noisy intermediate trajectory predictions. By analyzing predicted trajectories and applying corrections during iterative denoising, our approach anticipates collisions before they become unavoidable. This interleaving of symbolic constraint satisfaction with neural trajectory generation enables predictive collision avoidance rather than reactive intervention. On the SafeLIBERO benchmark, our method achieves 82.8% collision avoidance and 81.6% task success, a 6.3% and 19.8% improvement respectively over single-step methods, with the largest gains on long-horizon tasks where compounding distribution shift is most pronounced. Video demonstrations of our approach are included on our project page at https://willenglish.tech/SafetyGuidedFlowMatching/.

cs.RO↗
arXiv2026

CAMEO: A Conditional and Quality-Aware Multi-Agent Image Editing Orchestrator

Conditional image editing aims to modify a source image according to textual prompts and optional reference guidance. Such editing is crucial in scenarios requiring strict structural control (i.e., anomaly insertion in driving scenes and complex human pose transformation). Despite recent advances in large-scale editing models (i.e., Seedream, Nano Banana, etc), most approaches rely on single-step generation. This paradigm often lacks explicit quality control, may introduce excessive deviation from the original image, and frequently produces structural artifacts or environment-inconsistent modifications, typically requiring manual prompt tuning to achieve acceptable results. We propose \textbf{CAMEO}, a structured multi-agent framework that reformulates conditional editing as a quality-aware, feedback-driven process rather than a one-shot generation task. CAMEO decomposes editing into coordinated stages of planning, structured prompting, hypothesis generation, and adaptive reference grounding, where external guidance is invoked only when task complexity requires it. To overcome the lack of intrinsic quality control in existing methods, evaluation is embedded directly within the editing loop. Intermediate results are iteratively refined through structured feedback, forming a closed-loop process that progressively corrects structural and contextual inconsistencies. We evaluate CAMEO on anomaly insertion and human pose switching tasks. Across multiple strong editing backbones and independent evaluation models, CAMEO consistently achieves 20\% more win rate on average compared to multiple state-of-the-art models, demonstrating improved robustness, controllability, and structural reliability in conditional image editing.

cs.CV↗
arXiv2026

Distributed Persistence Domain for Persistent Memory Pooling

Compute Express Link (CXL) enables memory pooling over disaggregated memory, offering the potential to improve resource utilization in persistent memory systems. However, integrating persistence semantics into CXL-based memory pooling introduces substantial latency, which limits system scalability. This overhead arises because persist operations must traverse the entire CXL fabric, including switches, links, and protocol layers, before reaching remote persistent memory. To this end, we argue that extending CXL switches with persistence support is a promising direction for improving the scalability of persistent memory pooling. However, moving persistence support into the network breaks the traditional correctness assumptions of centralized persistence domains. In particular, enabling persistence within distributed structures, such as CXL switches, can introduce stale reads and writes if not carefully coordinated. In this paper, we propose Distributed Persistence Domain (DPD), a new abstraction for persistent memory pooling that enables persistence support at the CXL switch level. We first formalize the concept of a distributed persistence domain and use DPD as a framework to identify the correctness hazards that arise when persistence structures are distributed across the CXL fabric. Based on this analysis, we derive the design requirements needed to guarantee correctness. Building on these insights, we present Persistent CXL Switch, a CXL switch architecture that incorporates persistence support to significantly reduce persist latency, enable read forwarding, and coalesce writes, while preserving correctness and crash consistency. We evaluated our system design using both SPLASH-4 and YCSB benchmarks. Simulation results show an average speedup of 33% over volatile CXL switches, and up to 36% speedup with read forwarding optimization across all workloads.

cs.ET↗
arXiv2026

Thought-Aware KV Cache Compaction for Reasoning via Adaptive Attention Matching

Reasoning language models generate lengthy chain-of-thought (CoT) sequences whose key-value (KV) cache grows linearly and becomes a memory bottleneck during decoding. Existing compaction methods treat reasoning trajectories as flat token sequences and apply uniform compression, ignoring the hierarchical structure of CoT reasoning where different steps vary drastically in importance. We propose \textbf{Thought-Aware Attention Matching (TAM)}, which exploits this structure through three mechanisms: (i)~thought segmentation that decomposes the trajectory into reasoning blocks, (ii)~adaptive budget allocation that assigns compression budget based on each segment's importance and size, and (iii)~pivotal token protection that preserves high-attention reasoning anchors. We prove that the allocation rule is optimal under a convex error model and that cumulative error under sequential compaction remains bounded. Experiments on AIME 2024 and MATH-500 with Qwen3-4B show that TAM improves accuracy over uniform compaction at the same memory footprint, with periodic compaction bounding peak memory to 3.1--3.2\,GB (a 65\% reduction) while maintaining competitive accuracy.

cs.CL↗
arXiv2026

Kairos: Lightweight Testing Framework for Timing-Induced Interaction Failures in LTE and 5G Core Networks

As cellular core networks evolve toward distributed and cloud-native architectures, control-plane interactions become more intricate and bring new challenges. Among these challenges, we find that introducing specific timing between two control-plane interactions can cause network function crash, which we define as timing-induced interaction failures. Prior research primarily addresses identifying malformed inputs and specification violations, while timing-induced interaction failures remain largely unexplored. In this paper, we conduct a systematic study of timing-induced interaction failures in LTE and 5G core networks. First, we establish a taxonomy of control-plane interaction patterns and analyze the failure modes of each pattern. Then, we design and implement Kairos, a lightweight testing framework to expose timing-induced interaction failures without analyzing cellular standard documents. Evaluating Kairos on two open source and two commercial LTE and 5G core networks, we uncover 20 new vulnerabilities and reproduce 34 existing issues. Our results show that timing-induced interaction failures are prevalent in LTE and 5G core networks and should be explicitly considered in future specifications.

cs.NI↗
arXiv2026

Polyphony: Diffusion-based Dual-Hand Action Segmentation with Alternating Vision Transformer and Semantic Conditioning

Dual-hand action segmentation, densely predicting actions for both hands from untrimmed videos, is essential for understanding complex bimanual activities. However, it poses several unique challenges: complex inter-hand dependencies, visual asymmetry between hands, representation conflicts where the dominant hand monopolizes gradients, and semantic ambiguity in fine-grained actions. We propose Polyphony, a three-stage method to address these challenges through: (1) an Alternating Dual-Hand Vision Transformer that alternates training between left- and right-hand mini-batches to ensure balanced gradient contributions from both hands while sharing a spatio-temporal encoder; (2) Semantic Feature Conditioning that aligns visual features with structured, compositional action descriptions to enhance discrimination of semantically similar actions; and (3) Diffusion-Based Segmentation with cross-hand feature fusion for inter-hand coordination and adaptive loss weighting for balancing performance. Polyphony achieves state-of-the-art on both dual-hand datasets (HA-ViD, ATTACH) with improvements up to 16.8 points, and on the single-stream Breakfast dataset (82.5%), outperforming the prior best method that uses a 12x larger backbone. Notably, our unified model with a single shared backbone surpasses baselines requiring separate per-hand models. Code is at https://github.com/x-labs-xyz/Polyphony-Dual-hand-Action-Segmentation.

cs.CV↗
arXiv2026

FinHarness: An Inline Lifecycle Safety Harness for Finance LLM Agents

Finance LLM agents must simultaneously block prompt-induced unauthorized actions and approve legitimate multi-step business workflows. However, boundary filters often miss irreversible mid-trajectory tool calls, while post-hoc LLM judges perform auditing only after termination -- too late for intervention and at a computational cost that scales linearly with trace length. We present FinHarness, an inline safety harness that wraps a finance agent end-to-end with three components: a Query Monitor that fuses single-turn intent with cross-turn drift, a Tool Monitor that evaluates each prospective tool call, and a Cascade module that integrates per-step risk and adaptively routes verification between a lightweight and an advanced-tier LLM judge. Fired risk factors are re-injected into the agent input as ex-ante evidence, enabling the agent to refuse, re-plan, or approve on its own. On FinVault, routed FinHarness cuts ASR from 38.3% to 15.0% while largely preserving benign approval ($41.1\% \to 39.3\%$), and uses $4.7\times$ fewer advanced-judge calls than an always-advanced ablation.

cs.CL↗
arXiv2026

Research on Intra-Chip Fusion Deployment and Optimization of Embodied Intelligence Business Operator NPU

Embodied intelligent computing integrates perception, computation and control. Traditional separate deployment of the three tasks leads to frequent data transmission, high latency and low hardware efficiency, failing to satisfy millisecond-level real-time requirements in dynamic scenarios. Besides, most operator optimization methods rely on foreign GPU platforms, while full-process collaborative optimization for domestic Phytium-Cambricon heterogeneous architectures is insufficient. This paper builds a domestic heterogeneous computing platform with Phytium FT-2000/4 processor and Cambricon MLU370 acceleration card, and proposes an NPU on-chip fusion deployment and full-process operator collaborative optimization strategy for perception, computation and control pipelines. Targeting embodied robot applications, modular optimization is conducted, including MLU hardware adaptation of motion blur correction operators for high-speed imaging, lightweight inference optimization of ViT models, and customized operator development for multi-DOF inverse kinematics solution. An on-chip data closed-loop and pipeline collaboration-based single-card solution is proposed to implement integrated execution of all perception-computation-control tasks on MLU370. Experimental results show that the proposed method achieves a full-process single-frame latency of 18.7 ms and a speedup of 2.89 compared with NVIDIA Jetson AGX Xavier, with 82.6% MLU utilization and comparable accuracy to mainstream platforms. This work offers a practical reference for domestic engineering applications of integrated embodied intelligent computing services.

cs.DC↗
arXiv2026

ACCORD: Alleviating Concept Coupling through Dependence Regularization for Text-to-Image Diffusion Personalization

Image personalization has garnered attention for its ability to customize Text-to-Image generation using only a few reference images. However, a key challenge in image personalization is the issue of conceptual coupling, where the limited number of reference images leads the model to form unwanted associations between the personalization target and other concepts. Current methods attempt to tackle this issue indirectly, leading to a suboptimal balance between text control and personalization fidelity. In this paper, we take a direct approach to the concept coupling problem through statistical analysis, revealing that it stems from two distinct sources of dependence discrepancies. We therefore propose two complementary plug-and-play loss functions: Denoising Decouple Loss and Prior Decouple loss, each designed to minimize one type of dependence discrepancy. Extensive experiments demonstrate that our approach achieves a superior trade-off between text control and personalization fidelity.

cs.CV↗
arXiv2026

SpecAlign: A Semantic Alignment Framework for SystemVerilog Assertion Generation

Existing Large Language Model (LLM) approaches to SystemVerilog Assertion (SVA) generation primarily focus on syntactic validity and formal verification outcomes, while semantic alignment between generated assertions and natural language specifications remains difficult to quantify. As a result, hallucinated or misaligned SVAs can reduce confidence and increase debugging efforts in the absence of golden RTL. This paper presents SpecAlign, a framework for semantic evaluation and refinement of LLM-generated SVAs. SpecAlign introduces two iterative alignment loops that assess both natural language properties and SVAs against the design specification using entailment-based classification. We improve alignment decisions by generating multiple reasoning paths using chain-of-thought prompting and aggregating them via a self-consistency voting mechanism. Misaligned assertions are analyzed to generate actionable feedback for refinement. We further define a quantitative alignment score to measure semantic consistency across iterations. Experimental results demonstrate that SpecAlign effectively detects semantic inconsistencies and improves assertion alignment without relying on golden RTL, providing a scalable complement to traditional formal verification evaluation metrics.

cs.AI↗
arXiv2026

Hallucination as Exploit: Evidence-Carrying Multimodal Agents

Multimodal agents increasingly choose tool calls from screenshots, documents, and webpages, where a false perceptual claim can turn hallucination from an answer-quality error into an authorization failure. We formalize this failure mode as hallucination-to-action conversion: an unsupported claim supplies the precondition for a privileged action. We propose evidence-carrying multimodal agents (ECA), which treat free-form model text as inadmissible evidence, decompose each tool call into action-critical predicates, obtain typed certificates from constrained DOM/OCR/AX verifiers, and use a deterministic gate to authorize only the privileges those certificates support. Rather than hiding perception error, ECA converts opaque model belief into auditable residuals at the verifier, schema, and implementation levels. Verifier red-teaming across 17 canonical attack categories shows that four targeted hardening steps are each necessary; after hardening, canonical gate bypass is 0/1,700 (Wilson 95% upper bound 0.22%). With content-derived certificates, ECA observes zero unsafe executions on 200 end-to-end tasks (Wilson 95% upper bound 2.67%) and 120 browser tasks (upper bound 4.3%). A HACR audit on 500 stratified task keys shows that unsupported action-critical claims reach unsafe execution for naive agents (100.0%) and prompt-only defenses (49.6%), but not for ECA. Oracle-certificate replay over 7,488 GPT-5.4 traces isolates gate correctness, while neural judge baselines still admit most unsafe actions under the same threat model. The resulting principle is simple: model language may propose tool use, but certified predicates must authorize it.

cs.AI↗