Searcharxiv⌕ Search

arXiv subjects

Jian Huang

Publications and source records attributed to Jian Huang.

At least 19 recordsLinked to original sources
arXiv2026

Black-Box Membership Inference via Word-Level Probability Estimation

Membership inference attacks (MIAs) have emerged as critical tools for auditing privacy risks in large language models (LLMs), aiming to determine whether a given text was included in a model's training corpus. However, most existing MIAs require access to per-token logits or probabilities, making them inapplicable in practice to proprietary LLMs that expose only textual continuations. To address this underexplored setting, we propose Word-level Probability MIA (WPMIA), a statistically principled MIA for strict black-box privacy auditing. WPMIA estimates word-level generation probabilities via Monte Carlo sampling with local kernel smoothing, then aggregates these estimates into a sequence-level likelihood estimator. Furthermore, WPMIA constructs the likelihood conditioned on different prefixes, thereby amplifying the distributional differences between members and non-members. We evaluate WPMIA across various open-source LLMs and find that it consistently outperforms existing black-box baselines. Importantly, we also evaluate WPMIA on modern proprietary LLMs, including GPT-5-Chat, Gemini-2.5-Flash, and Claude-4.5-Haiku, achieving an average TPR@5\%FPR of 42.0 across these models. These results offer a sound foundation for future research on strict black-box membership inference. Code is available at \href{https://github.com/niusj03/WPMIA}{https://github.com/niusj03/WPMIA}.

cs.CR↗
arXiv2026

TACIT-Switch: Cost-Aware Model Escalation for LLM Agents from Censored Supervision

Agents with smaller language-model backbones are less expensive but can drift into persistent failure modes, whereas those with larger backbones are generally more reliable but more costly. This reliability-cost trade-off motivates routing methods that decide when to invoke an agent with a larger backbone: before execution, after a fixed trajectory prefix, or locally at individual steps. Our method, TACIT-SWITCH, learns permanent handoff policies from accumulated trajectory evidence and Teacher-Annotated Censored Intervention Times (TACIT). It represents each annotation as an interval-censored observation on a cumulative-risk scale. The resulting mixture-cure threshold model estimates the probability that the paired Strong rollout succeeds and, conditional on success, the handoff threshold; no teacher is required at deployment. In a mechanism-based multi-step simulation, TACIT-SWITCH improves success by 7.4-11.1 percentage points over task-level, step-level, and fixed-prefix routing baselines at comparable cost. Within that controlled simulation, ablations show that task features and cumulative trajectory risk provide complementary information. With operating points selected on development data, TACIT-SWITCH achieves the highest held-out success among learned policies on both ALFWorld (48.5% with 4B Cheap; 45.5% with 9B Cheap) and DABench (73.1%).

cs.LG↗
arXiv2026

Wasserstein Filtering: A Sample Selection Method for Robust Distribution Learning

Given a dataset where a portion of the samples are contaminated, our goal is to recover the underlying clean population distribution. To this end, we propose Wasserstein Filtering (WF), a novel sample selection framework that discards a fraction of suspicious samples and estimates the target distribution using the empirical measure of the remaining data. The core insight is to select a subset of samples whose empirical distribution maximizes its Wasserstein distance to the fully contaminated empirical distribution, thereby preferentially isolating and removing geometrically influential outliers. To render this optimization computationally tractable, we introduce three algorithms: a marginal screening scheme, SinkMarg, and two joint optimization algorithms, SinkWF and SlicedWF, leveraging entropic optimal transport and sliced Wasserstein approximations, respectively. On the theoretical front, we introduce the Far Exclusion and Local Projection (FELP) contamination model, which characterizes corruptions consisting of well-separated outliers and locally indistinguishable perturbations. Under this model, we prove that the WF estimator achieves minimax optimality over distribution families with bounded covariance. Extensive numerical experiments on synthetic datasets, benchmark anomaly detection suites, and robust generative learning with diffusion models demonstrate that WF serves as a highly practical, model-agnostic preprocessing tool. It delivers competitive outlier detection performance and provides substantial downstream benefits for generative modeling under heavy contamination.

stat.ML↗
arXiv2026

Benchmarking LLMs on File System Design and Implementation

Large Language Models (LLMs) are fundamentally transforming computer system research and development. As we employ LLMs in file system (fs) development, it is essential to understand their capabilities, limitations, and operational efficiency for domain-specific tasks. We present \phi-Bench, an LLM benchmarking framework for fs-specific tasks. To facilitate benchmarking, we develop six types of tasks in \phi-Bench: basic understanding, basic implementation, performance modeling, debugging, optimization, and new feature development. Each type emphasizes different LLM capabilities: instruction following, knowledge recall, reasoning, or coding. To create high-quality tasks while achieving broad coverage with minimal human effort, we develop a new AI-assisted task generation pipeline in addition to expert-written and textbook-adapted tasks. With 505 tasks in \phi-Bench, we conduct an empirical study with both open source (DeepSeek-V4-Flash, GLM-5.1, and MiniMax-M2.7) and proprietary (Claude-Opus-4.7, GPT-5.2, and Gemini-3.1-Pro) LLMs. Our study discloses the model efficiency for different tasks, causes of failed fs tasks, and techniques for mitigating LLM failures. We will open source \phi-Bench to facilitate public research on using LLMs for fs development.

cs.OS↗
arXiv2026

Harness-G: A Graph-Structured Harness for Search Agents

Reinforcement learning (RL) search agents commonly model retrieval as free-form natural-language query generation and optimize multi-turn interactions using final-answer rewards. Current studies mainly improve training with denser or more structured credit signals, but rarely examine whether retrieval is properly formulated at the policy-environment interface. We observe pronounced retrieval aliasing during Search-R1 training: rollouts for the same question continue to generate distinct query strings, yet their accumulated evidence sets increasingly overlap. We call this phenomenon retrieval-equivalence collapse; in this regime, trajectories approach utility equivalence with respect to retrieval decisions, leaving within-group returns with little effective retrieval contrast. To address this problem, we propose Harness-G, a graph-structured retrieval framework that redesigns this interface. It reformulates free-form query generation as finite action selection: the policy selects an evidence sentence or entity, or chooses to answer, while the environment constructs the menu, tracks retrieval state, and validates and executes each choice. This interface reduces linguistic aliasing and makes same-state alternatives directly comparable. Building on this interface, we introduce Structured Non-myopic Credit (SNC), which uses a frozen answer scorer to compare the selected action with its alternatives and assigns downstream gains to the earlier actions that enabled them. Across six QA benchmarks, Harness-G achieves the highest average F1 at both evaluated model scales, outperforming the strongest baseline, Graph-R1, by 10.74 points at 1.5B and 3.98 points at 3B.

cs.CL↗
arXiv2026

Semi-Supervised Conditional Diffusion via Label Augmentation

Conditional diffusion models have become a powerful and flexible framework for learning complex conditional distributions from labeled data. In practice, however, acquiring high-quality labels is costly and time-consuming, leaving large volumes of unlabeled data unused. To address this, we introduce label-augmented conditional diffusion (LACD), a simple and effective approach that incorporates unlabeled examples by assigning them a designated trivial label and performing joint denoising score matching over the augmented dataset. We provide sufficient conditions guaranteeing population-level identifiability of the target conditional distribution under this scheme. Moreover, we establish rigorous statistical guarantees: when sufficiently many unlabeled samples are available, the sampling distribution produced by LACD converges strictly faster than the purely supervised estimator in total variation distance, and at least as fast in Wasserstein-1 distance. Extensive experiments on synthetic, image, and tabular benchmarks corroborate our theory and show substantial gains in sample efficiency and generative performance compared with the purely supervised estimator.

stat.ML↗
arXiv2026

Semi-Supervised Conditional Generative Learning through Stochastic Interpolation and Sufficient Representations

Conditional generative modeling remains a challenging problem in semi-supervised settings where labeled data is scarce but unlabeled samples are abundant. To effectively leverage structural information embedded within the unlabeled dataset and compensate for sparse conditioning signals, we propose a semi-supervised framework combining conditional stochastic interpolation with low-dimensional latent representations. RepG decomposes generation into two stages: label-dependent latent sampling and high-dimensional reconstruction. This isolates the supervised learning of conditional dependencies to a low-dimensional space, requiring few labels while utilizing the abundant unlabeled data purely for reconstruction. Theoretically, we establish an error decomposition showing that the Kullback-Leibler divergence of RepG comprises stage-wise estimation errors and a structural bias quantified by conditional mutual information. For deep neural network estimators, we derive non-asymptotic convergence rates proving that RepG significantly improves sample complexity. By confining the supervised estimation burden to the low intrinsic dimension of the latent representation, RepG achieves a strictly faster convergence rate. Complemented by a minimax lower bound, our theoretical results demonstrate that this method effectively mitigates the curse of dimensionality inherent in direct ambient-space generative modeling.

stat.ML↗
arXiv2026

Event3R: Asynchronous-to-Global 3D Reconstruction from Event Camera via Spatial-Temporal Feature Aggregation

Robust 3D reconstruction is essential for robotics and embodied perception. Recent feed-forward approaches such as DUSt3R have demonstrated impressive progress in dense 3D reconstruction from RGB images, achieving global geometric consistency and strong generalization. However, extending such dense 3D reconstruction to event cameras remains challenging due to their asynchronous, sparse, and highly dynamic nature, as well as the lack of large-scale, well-labeled datasets. In this work, we introduce Event3R, a feed-forward framework that directly maps asynchronous event streams to globally consistent 3D point clouds. Event3R represents incoming events as spatial-temporal voxels, enabling time-aware feature integration through a temporal attention module that enhances the module's temporal feature learning. To further strengthen temporal representation learning and reduce reliance on labeled data, we propose a Masked Bin Modeling (MBM) strategy for self-supervised pre-training, enabling robust temporal representation learning with minimal labeled data, and retain it as an auxiliary fine-tuning objective. In addition, contrastive alignment and consistency regularization losses are incorporated during fine-tuning to reinforce structural correspondence and temporal coherence across views. Extensive experiments on both synthetic and real-world benchmarks demonstrate that Event3R achieves robust, temporally consistent, and globally aligned 3D reconstructions, significantly outperforming existing event-based methods.

cs.CV↗
arXiv2026

Enabling Spatially Fine-Grained DVFS in Neural Processing Units for Energy-Efficient LLM Serving

As neural processing units (NPUs) evolve rapidly to accommodate the ever-increasing compute demand of large language models (LLMs), their power consumption is becoming a limiting factor. Our study shows that using dynamic voltage and frequency scaling (DVFS) to exploit the service-level objective (SLO) slacks is a promising way to improve NPU energy efficiency for LLM services. And as tensor operators in LLMs exhibit diverse bottlenecks across NPU components, it is desirable to configure the frequency separately for each component to maximize their energy efficiency. In this paper, we develop eNPU that enables hardware and software support for spatially fine-grained, component-level DVFS on NPUs. eNPU refactors the NPU core pipeline to partition components into separate V/$f$ domains. It introduces lightweight cross-domain communication mechanisms to mitigate synchronization overheads across components, and extends the NPU ISA for sub-$\mu$s DVFS control. eNPU uses a compiler-driven two-level greedy search to co-optimize instruction scheduling and per-component V/$f$ selection under SLO constraints. We implement eNPU's pipeline design on an open-source NPU core to verify its functionality and evaluate the energy savings with a production-level NPU simulator with various LLMs using production traces. eNPU reduces energy consumption of LLM services by 25.8%--35.2% with 3.45% area overhead on a TPUv4 chip, while preserving strict SLO guarantees.

cs.AR↗
arXiv2026

Auto-Scaling Heterogeneous Neural Processing Units for Energy and Cost-Efficient LLM Serving

To meet the ever-increasing computing demands of large language model (LLM) services, modern cloud platforms have widely deployed neural processing units (NPUs). These NPU chips have been developed and evolved at an incredibly fast pace, this inevitably produces heterogeneous compute pools backed by different versions of NPU chips. Unfortunately, due to the lack of system and architecture support for managing NPU heterogeneity in the cloud, it is unclear how to best utilize heterogeneous NPUs to maximize the energy and cost efficiency for LLM services. In this paper, we first conduct a characterization study of various generations of real NPU chips to demonstrate the potential benefits on energy/cost efficiency and performance by utilizing heterogeneous NPU chips. To realize these benefits, we present NeuScale, an auto-scaling framework to automatically exploit heterogeneous NPUs for cloud platforms. NeuScale manages heterogeneous NPU resources with a new vPod abstraction, which abstracts the core hardware parameters of different NPU versions and provides compatibility with existing ML frameworks. It makes the best-fit vPod allocations for different LLM inference requests using an intuitive and lightweight roofline-based analysis. It supports fine-grained dynamic NPU resource provisioning by adjusting both the vPod configuration (i.e., scaling up/down) and the number of vPods (e.g., scaling in/out). To validate the benefits of NeuScale at scale, we implement it with a production-level NPU simulator. Our evaluation with popular LLMs shows that NeuScale can significantly improve cost efficiency and service-level objective (SLO) satisfaction rate by best utilizing heterogeneous NPU resources.

cs.AR↗
arXiv2026

Slope-Guided Mamba and Angular-Refined Transformer for Light Field Super-Resolution

Light Field Super-Resolution (LFSR) necessitates accurate modeling of spatial-angular correlations while preserving intrinsic 4D ray coherence. However, maintaining such high-dimensional consistency remains challenging, primarily due to two inherent limitations in prevailing modeling paradigms. First, spatial and angular dimensions are often modeled in a decoupled manner, restricting early cross-dimensional interaction and leading to geometric inconsistencies. Moreover, although continuous sequence modeling paradigms show promise in representing epipolar structures, their rigid scanning mechanisms fundamentally conflict with epipolar geometry, limiting geometry-aware feature aggregation. To address these challenges, we propose a hybrid light field super-resolution network, termed SMART, which integrates a Slope-Guided Mamba and an Angular-Refined Transformer to effectively overcome these limitations. Specifically, we introduce an angular-modulated spatial module to bridge the decoupling gap, incorporating angular priors to strengthen spatial-angular correlation modeling. To mitigate the scan-geometry mismatch, we propose a manifold-aligned trajectory module that enables geometry-consistent sequence modeling along epipolar structures. Experiments on five benchmarks demonstrate that SMART achieves state-of-the-art performance, surpassing previous methods by 0.42 dB (PSNR) with significantly reduced artifacts.

cs.CV↗
arXiv2026

Benchmarking LLM Serving Systems for Agentic AI Workloads with XPerf

We present XPerf, a benchmarking framework that load-tests LLM serving systems with diverse agentic AI workloads. It provides detailed profiling of the serving system and hardware, enabling users to identify performance bottlenecks introduced by agentic workloads. Benchmarking LLM serving systems under agentic workloads is challenging - agentic applications rely on nondeterministic LLM outputs to guide their control flow; therefore, workload patterns vary unpredictably from run to run. XPerf minimizes this workload variation with a fine-grained trace replay approach: it enables users to easily collect traces from real agentic applications, synthesize new workloads with various patterns if needed, and reproducibly replay them on different LLM serving systems. XPerf includes eight agentic applications across diverse use cases (e.g., coding, deep research, and Q&A) by default. Our empirical study using these workloads shows that XPerf accurately replays agentic workloads, provides detailed performance breakdowns, scales to larger serving systems, and assists in serving system debugging. We will open-source XPerf on GitHub.

cs.DC↗
arXiv2026

The Curious Case of PHL 1811: Heavy Obscuration Versus Intrinsic X-ray Weakness

We present a systematic X-ray analysis of the narrow-line Type 1 quasar PHL 1811, which has long been regarded as the prototype of intrinsically X-ray weak quasars. A critical breakthrough came with the first detection of a bright X-ray flare from this source by the Einstein Probe (EP) in 2024. We utilize archival X-ray observations spanning 2001-2024, including the post-flare EP and Swift data. We confirm that PHL 1811 shows X-ray weakness factors $f_{\rm weak} \approx 23$-179 across all epochs before 2024. The 2024 EP flare marks the first detection of an X-ray nominal state with $f_{\rm weak} \approx 0.63$, followed by a rapid flux decline. We identify three key observational signatures that strongly support heavy obscuration: (1) a significant hard X-ray excess above $\approx5$ keV in the 2015 XMM-Newton spectrum; (2) relatively flat spectral shapes in two Swift observations; and (3) transitions between X-ray nominal and multiple X-ray weak states without corresponding optical/infrared variability, consistent with expectations from obscuration by a clumpy dust-free absorber. Fitting with a partial-covering obscuration model reproduces all multi-epoch spectra well. The observed steep spectra are dominated by a small leaked/scattered fraction of the intrinsic continuum, and variability is driven by changes in the leakage fraction and column density. Our results strongly favor the scenario where PHL 1811 is obscured by a radiatively driven accretion-disk wind from super-Eddington accretion, unifying PHL 1811 with the broader population of super-Eddington accreting AGNs under a single obscuration framework.

astro-ph.HE↗
arXiv2026

LAP: An Agent-to-Instrument Protocol for Autonomous Science

Autonomous science is moving from demonstration to infrastructure. Large language model agents now plan experiments, and self-driving laboratories execute them. Yet every such system rebuilds the link between the reasoning agent and the physical instrument from scratch, against fragmented vendor SDKs and standards built for deterministic software clients rather than probabilistic, goal-directed agents. Recent agent-interoperability protocols clarify two of the three edges of an agentic ecosystem (Anthropic's Model Context Protocol (MCP) standardizes the agent-to-tool edge, and Google's Agent2Agent (A2A) the agent-to-agent edge), but neither models the agent-to-instrument edge, where operations are stateful, safety-critical, exclusively owned, physically embodied, and produce measurements with units, calibration, and uncertainty. We present the Lab Agent Protocol (LAP), a protocol design that fills this gap. LAP retains A2A's peer-to-peer, discovery-first, task-lifecycle structure and adds four physical-world primitives: (i) the InstrumentCard, a signed capability and physical-limit description; (ii) first-class reservation for exclusive instrument and sample locking; (iii) a safety-fence handshake with operator-confirmation tokens cryptographically bound to a specific task and its parameters, gating hazardous and irreversible operations; and (iv) a MeasurementResult schema that makes every result physically typed (QUDT/UCUM), calibration-anchored, uncertainty-bearing, and reproducible by construction. We specify roles, a six-layer architecture, the JSON-RPC method set, the task and safety state machines, the error model, and cross-laboratory federation, and walk a closed-loop autonomous campaign through the protocol end-to-end. LAP is transport-compatible with the A2A/MCP ecosystem and encapsulates rather than replaces existing device standards such as SiLA 2 and OPC-UA.

cs.AI↗
arXiv2026

Vegas: Self-Speculative Decoding with Verification-Guided Sparse Attention

Long-context large language model (LLM) inference has become the norm for today's AI applications. However, it is severely bottlenecked by the increasing memory demands of its KV cache. Previous works have shown that self-speculative decoding with sparse attention, where tokens are drafted using a subset of the KV cache and verified in parallel against the full KV cache, speeds up inference in a lossless manner. However, they rely on a standalone KV selection algorithm to select the KV entries used for drafting and overlook the fact that the criticality of each KV entry is inherently computed during verification. In this paper, we propose Vegas, a self-speculative decoding method with verification-guided sparse attention. Vegas identifies critical KV cache entries as a byproduct of verification and computes attention only over these entries when drafting subsequent tokens. This not only improves the draft token acceptance rate but also incurs low KV selection overhead, thereby improving decoding throughput. Vegas achieves a 1.25$\times$-2.81$\times$ speedup in decoding throughput over default vLLM and a 1.15$\times$-1.29$\times$ speedup over state-of-the-art sparse attention-based self-speculative decoding methods. Our code is available at https://github.com/platformxlab/vegas.

cs.LG↗
arXiv2026

Selective Fermi-Level Pinning: A Design Strategy for Giant Rectification in Molecular Junctions

Molecular rectifiers are key functional components of molecular-scale integrated circuits, yet achieving high rectification ratios remains a longstanding challenge due to the intrinsic symmetry of resonant tunneling and the complexity of interfacial energy-level alignment. Here, we propose a rectifier design strategy based on selective Fermi-level pinning that breaks transport symmetry via pinning interactions between molecular frontier orbitals and electrodes. This framework enforces tunneling transport to be predominantly governed by unoccupied molecular orbitals, while substantially suppressing contributions from occupied states, thereby establishing a simplified and highly controllable rectification mechanism. The resulting cyclo[n]carbon-based molecular junctions exhibit giant rectification ratios exceeding 103, while retaining exceptional structural robustness against variations in both donor chain length and carbon ring size. This work reveals the critical role of selective Fermi-level pinning in molecular junctions and provides a general design principle for engineering functional single-molecule electronic devices.

cond-mat.mes-hall↗
arXiv2026

Phase-tunable remote nonreciprocal charging in waveguide QED

Remote quantum batteries require directional and controllable energy transfer between spatially separated quantum nodes, yet most existing protocols rely on direct charger-battery Hamiltonian couplings. Here we propose a phase-tunable waveguide-QED architecture for remote quantum-battery charging, in which a driven charger and a remote battery are coupled solely via engineered waveguide-mediated interference, without any direct local interaction. We systematically compare four configurations: two-giant-emitter and giant-small-emitter hybrids, each with open or mirror-terminated waveguides. By engineering the propagation and coupling phases, the waveguide-mediated coherent exchange interaction and collective dissipation can be balanced to suppress the backward channel while retaining a finite forward channel, thereby realizing cascaded-like unidirectional charging. Our analysis shows that nonreciprocity and storage efficiency can be independently engineered, offering design flexibility for different quantum network scenarios. The giant-small-emitter mirror-terminated configuration simultaneously achieves perfect nonreciprocity and battery-dominated storage, while both giant-small-emitter configurations exhibit distance-insensitive directionality. Extending the scheme to quadratic driving, we show that anomalous second moments render the battery state non-passive, making ergotropy a performance metric distinct from stored energy. These results establish phase-tunable waveguide networks as a versatile platform for remote quantum-energy transfer and provide design principles for directional and work-extractable energy storage in quantum networks.

quant-ph↗
arXiv2026

Strong X-ray Variability of I Zwicky 1: Obscuration from Clumpy Accretion-Disk Winds

Obscuration from clumpy accretion-disk winds has been invoked to explain the extreme X-ray weakness and X-ray variability observed in a substantial fraction of super-Eddington accreting quasars. We present a comprehensive study of the strong X-ray variability of the super-Eddington accreting active galactic nucleus (AGN) I Zwicky 1 (I Zw 1), a prototypical narrow-line} Seyfert 1 galaxy (NLS1), to test the disk-wind obscuration scenario as the underlying mechanism and characterizing the disk-wind absorber properties. We focus on spectral and temporal analyses of simultaneous XMM-Newton and NuSTAR observations in 2020, and a 100-day NICER monitoring campaign in 2022. Despite strong X-ray variability by factors of $\approx3$ and $\approx6$ on short-term and long-term timescales, respectively, the XMM-Newton Optical Monitor observations do not show contemporaneous significant UV variability, and archival data reveal only mild long-term optical/infrared variability ($\approx30\%$), indicating a stable accretion process in I Zw 1. The strong X-ray variability thus likely arises from variable absorption of relatively stable coronal emission. We perform time-resolved X-ray spectroscopy utilizing a partial-covering absorption model with a stable corona and varying ionized absorbers. We identify three distinct absorbers whose variations in the column density and covering factor successfully explain the observed X-ray ``flares'' in 2020 and the longer-term spectral evolution in 2022. Our results support a unified scenario in which obscuration from clumpy disk winds produces the strong X-ray variability observed in super-Eddington accreting AGNs. This scenario may be applicable to other NLS1s exhibiting strong X-ray variability to better characterize the disk winds driven by super-Eddington accretion.

astro-ph.HE↗