arXiv subjects

Yan Zhang

Publications and source records attributed to Yan Zhang.

At least 55 records · Page 3Linked to original sources

arXiv2026

Multimodal Aspect-Level Sentiment Analysis Based on Gated Noise Filtering and Emotion-Relevance Interaction

Multimodal Aspect-Based Sentiment Analysis (MABSA) infers fine-grained sentiment polarity toward specific aspects by jointly modeling text and images. Despite progress in cross-modal fusion, two challenges remain in multi-aspect settings: (1) multimodal noise, where aspect-irrelevant content distracts sentiment learning; and (2) weak cross-modal sentiment alignment, as visual evidence can be ambiguous and textual--visual sentiments may conflict, limiting multimodal complementarity. To address these issues, we propose a Gated Noise-filtered Sentiment-Relevance Interaction (GNSRI) framework. It employs a gated noise-filtering module to suppress sentiment-irrelevant features and enhance aspect-aware sentiment cues, and a sentiment-relevance interaction module to capture consistent and conflicting cross-modal signals at micro and macro levels. Finally, a learnable decision fusion mechanism adaptively combines predictions from textual, visual, and cross-modal branches at the aspect level. Experiments on public MABSA benchmarks show that GNSRI outperforms state-of-the-art methods, improving accuracy by 1.94\% and 2.06\% on Twitter-2015 and Twitter-2017, respectively.

cs.MM↗

arXiv2026

Chamaileon: Cross-Context Binder Design with Contextualized Modeling and Mixed Sampling

The rapid evolution of generative models has unlocked new potentials in protein binder design, a pivotal task in structural biology, by facilitating end-to-end generation via joint sequence-structure modeling or hallucination. However, existing approaches are predominantly implemented under a single-target, single-state assumption, limiting their ability to model multi-target or multi-state interactions required for advanced function-oriented protein design. Here, we introduce Chamaileon, which unifies multi-target and multi-state binder design by formulating the problem as cross-context binding landscape modeling. The framework is underpinned by a training paradigm termed In-Context Complex Co-Design (I3CD) for context-aware sequence-structure co-modeling. During inference, we employ Mixture-of-Paths Sampling (MoPS), a scalable strategy that optimizes a single sequence across contexts while alleviating the scarcity of high-quality multi-conformational paired data. Extensive evaluation on our newly constructed benchmark, CROSS, demonstrates that Chamaileon effectively generates sequences adaptable to diverse conformational landscapes and multi-target requirements. The code is available on https://github.com/caohengyuan/Chamaileon.

cs.LG↗

arXiv2026

Varifocal Displays Reduce the Impact of the Vergence-Accommodation Conflict on 3D Pointing Performance in Augmented Reality Systems

This paper investigates whether a custom varifocal display can improve 3D pointing performance in augmented reality (AR), where the vergence-accommodation conflict (VAC) is known to impair interaction. Varifocal displays have been hypothesized to alleviate the VAC by dynamically matching the focal distance to the user's gaze-defined target depth. Following prior work, we conducted a within-subject study with 24 participants performing an ISO 9241-411 pointing task under varifocal and fixed-focal viewing. Overall, varifocal viewing yielded significantly higher performance than the fixed-focal baseline across key interaction metrics, although the magnitude and even the direction of the benefit varied across individuals. In particular, participants' responses exhibited a baseline-dependent pattern, with smaller improvements (or occasional degradation) observed for those with better baseline performance. Our findings suggest that varifocal technology can improve AR pointing performance relative to fixed-focal viewing, while highlighting substantial individual differences that should be considered in design and evaluation.

cs.HC↗

arXiv2026

More Perspectives, Stronger Signals: Multi-Perspective Enhancement and Progressive Fusion for Multimodal Entity Representation Learning

Learning effective multimodal entity representations is fundamental for reasoning tasks such as multimodal knowledge graph completion (MMKGC). However, existing methods often suffer from semantic over-smoothing within modalities and ineffective noise filtration across modalities, particularly under sparse or ambiguous conditions. To overcome these limitations, we propose PrismF, a unified framework that synergizes multi-perspective enhancement with progressive fusion to extract stronger signals from diverse inputs. PrismF enhances fine-grained intra-modal semantics through a multi-perspective mechanism that decomposes each modality into complementary views and constrains them with a decoupling loss to reduce representation collapse. Furthermore, it improves cross-modal integration through a progressive fusion strategy that dynamically calibrates inter-modal interactions, enabling the model to emphasize informative signals while suppressing noisy or unreliable ones. Extensive experiments on three public benchmarks show that PrismF achieves the strongest overall performance, including relative improvements of 4.04% in MRR and 11.17% in Hits@1 on KVC16K. Our code can be found at https://github.com/HubuKG/PrismF.

cs.AI↗

arXiv2026

Multiple superconducting phases and order-parameter evolution in pressurized UTe$_2$

The recently discovered heavy-fermion spin-triplet superconductor candidate UTe$_2$ provides a rich platform for unconventional pairing and topological phenomena. However, limited has been known about its superconducting order parameters and their evolution with control parameters, largely due to the lack of appropriate symmetry-sensitive detections. Here, we report comprehensive point-contact spectroscopy measurements of pressurized UTe$_2$ on the (0~0~1) surface. The observation of Andreev bound states strongly suggests the presence of a $p_z$ component in the superconducting order parameters. Quantitative analysis based on an extended Blonder-Tinkham-Klapwijk model unveils the superconducting order parameters with a finite odd-$k_z$ component (e.g. $B_{2u}$ or $B_{3u}$) for both ambient and pressurized UTe$_2$. Remarkably, the multiple superconducting phases can be distinguished by a single parameter $\langle Δ_{z}\rangle/\langleΔ_{x(y)}\rangle$, the relative weight between the $p_z$-wave and $p_{x(y)}$-wave pairings. These findings place stringent constraints on the pairing symmetry and provide essential spectroscopic signatures for distinguishing pressure-induced multiple superconducting phases in UTe$_2$.

cond-mat.str-el↗

arXiv2026

Spectroscopy of hidden-heavy tetraquark states with $J^{PC}=0^{--}$ in a color-octet configuration

Within the QCD sum-rule framework, we investigate hidden-heavy tetraquark channels with the exotic quantum number $J^{PC}=0^{--}$ using four representative local color-octet--octet interpolating currents. The currents include both vector--axialvector and scalar--pseudoscalar Dirac structures. The operator product expansion is carried out up to dimension-eight condensates. The four diagonal sum rules yield mutually consistent mass estimates in the range $10.8$--$11.1~\mathrm{GeV}$ for the hidden-bottom sector and around $4.3$--$4.6~\mathrm{GeV}$ for the corresponding hidden-charm sector, with the bottom sector exhibiting the clearest Borel stability. Since local tetraquark currents with the same quantum numbers are related by Fierz rearrangements, the current-dependent results do not by themselves imply four distinct states or uniquely defined internal color structures. We also discuss quantum-number-allowed decay channels and emphasize the absence of the lowest pseudoscalar--pseudoscalar heavy-meson modes for a neutral $0^{--}$ state. The results provide theoretical guidance for future experimental searches at Belle II, LHCb, and BESIII.

hep-ph↗

arXiv2026

Velocity Index Modulation for Movable Antenna Systems

In movable antenna (MA) systems, antenna movement induces Doppler frequency shifts that are conventionally treated as an impairment requiring mitigation. In this paper, we propose \emph{Velocity Index Modulation for Movable Antennas} (VIM-MA), which reframes this Doppler effect as an additional information-bearing degree of freedom. The transmitter selects the antenna movement velocity from a pre-designed discrete codebook, so that the resulting Doppler shift conveys extra index bits beyond those carried by the conventional modulation symbol. Codebook design is formulated as a spectral efficiency maximization over the velocity spacing $δ$ and codebook size $N_v$, subject to an average-information Cramér--Rao-type bound (AIF-CRB) on velocity estimation accuracy, a physical track length constraint, and a spatial channel decorrelation constraint. A logarithmic change of variables renders the problem convex and yields a closed-form solution. We further establish that the peak codebook velocity equals $D_{\max}/T_s$, and that the decorrelation-limited spacing always lies below the Rayleigh Doppler resolution, so that VIM-MA is intrinsically a super-resolution scheme. A covariance-matched detector is derived that requires neither per-path angle knowledge nor channel state information. Simulation results show that the decorrelation-limited codebook, which carries five index bits over a $10λ$ aperture, is attainable only with oracle angle knowledge, whereas the channel-state-free detector is limited to three bits but reaches that payload approximately $10$~dB earlier than position-domain indexing charged a realistic pilot budget.

eess.SP↗

arXiv2026

CRAFT: Constrained Reward via Attention Fine-Tuning for Subject Personalization without Composed Targets

Subject-driven image personalization---generating new images that preserve the identity of one or several reference subjects in novel scenes---is a foundational capability for modern visual content creation. It is currently dominated by generalized methods that fine-tune a pretrained multimodal diffusion transformer (MMDiT) on hundreds of thousands to millions of paired \emph{(reference, composed-target)} examples, where each composed target is a synthesized image of the subject in a novel scene. Producing such targets demands a costly multi-stage curation pipeline---LLM-based prompt generation, T2I-based composed-target synthesis, reference-subject extraction, VLM-based quality filtering, and correspondence labeling---and tightly couples each method to a particular target synthesizer and curation choice. We introduce \emph{CRAFT} (Constrained Reward via Attention Fine-Tuning), a single-step ReFL framework that fine-tunes a pre-trained \emph{reference-aware} MMDiT via LoRA adapters using a compact reference-only data construction---$10$K reference images and subject masks, with no composed-target supervision. CRAFT realizes a \emph{Where to look} principle: attention-level rewards align noise- and phrase-token attention with the correct reference subject, and the resulting per-subject attention masks gate a pixel-level identity reward to keep image-space supervision consistent with the learned attention routing. Applied to FLUX.2-klein-9B, CRAFT achieves state-of-the-art performance on XVerseBench \rev{while using no composed-target supervision---only $10$K reference-only samples, whereas prior generalized methods require $150$K to over $2$M composed-target pairs}. The same recipe transfers to other reference-aware backbones, consistently improving performance. Project page: https://jihun999.github.io/projects/CRAFT/.

cs.CV↗

arXiv2026

Graph-MambaNav: Spatial-Temporal Graph Mamba Leveraging Object-Relation Knowledge for Object-Goal Navigation

Object-goal navigation requires an agent to reason over object relationships and prioritize target-relevant objects for efficient decision making in unseen environments. While existing graph-based methods incorporate target-awareness at the feature or attention level, they remain permutation-invariant and lack an explicit mechanism to control information propagation order, limiting their ability to model target-dependent importance and long-range dependencies. In contrast, Graph-Mamba highlights that node prioritization through sequence ordering is critical for effective global reasoning. In this work, we investigate the node prioritization mechanism in Graph-Mamba and study its role in object navigation. We propose Graph-MambaNav, a target-aware spatial-temporal graph encoding framework that introduces a heuristic ordering over objects based on their relevance to the target, allowing more informative objects to be processed later to aggregate richer context. Both node ordering and edge weights are initialized from LLM-derived commonsense object relationships, providing a unified prior for structured reasoning. A spatial module integrates local message passing with global GraphMamba-based selective scanning, while a temporal module applies Mamba-based sequence modeling over object-wise temporal orders, allowing selective aggregation of historical context for long-range temporal reasoning. Experiments on AI2-THOR and RoboTHOR demonstrate improved navigation performance with generalization, and additional real-world robot deployment further validates the effectiveness of our proposed approach.

cs.RO↗

arXiv2026

Learning from Consensus and Disagreement: Unsupervised On-Policy Self-Distillation with Minority-Trajectory Contrast

On-policy self-distillation improves language-model reasoning by querying a teacher on states actually visited by the student. Recent methods create a powerful information asymmetry by exposing the teacher to privileged context, yet they fundamentally rely on external supervision---such as gold solutions or verifiers---to construct this advantage. We introduce CoDA (Consensus and Disagreement Alignment), a fully unsupervised framework that creates reliable privileged information entirely from the latent uncertainty structure of a model's own unlabeled rollouts. CoDA extracts two complementary signals. In the positive branch, answer-level consensus identifies a stable reasoning mode, which conditions a frozen self-teacher to provide dense distributional guidance on fresh student trajectories. However, because agreement does not guarantee correctness, positive-only distillation risks amplifying correlated errors into a false consensus. To break this harmful feedback loop, CoDA incorporates a negative branch that exploits disagreement: minority trajectories are treated as unstable alternatives and gently penalized via a reference-anchored, KTO-style calibration objective. This unpaired binary feedback provides robust regularization without requiring the strong assumption that the consensus is the absolute ground truth. Empirical evaluations on competition-level mathematical benchmarks demonstrate that CoDA significantly improves reasoning, outperforming self-generated baselines and effectively stabilizing training against erroneous consensus.

cs.LG↗

arXiv2026

StatCite: A Large-scale Citation Network Dataset for Statistics and Data Science

In this paper, we introduce StatCite, a large-scale citation network dataset covering publications in statistics and data science from 1981 to 2025. The dataset contains 189,101 research articles collected from 62 representative journals and provides bibliographic metadata, including title, author list, publisher, published year, abstract, keywords, and reference list. Based on the collected publications, we construct four complementary citation-based networks, namely the paper citation network, the co-citation network, the bibliographic coupling network, and the journal citation network. To illustrate the utility of the dataset, we present descriptive analyses of the constructed networks and investigate the community structure of the paper citation network. The results show that StatCite preserves key structural characteristics commonly observed in large-scale citation networks and captures several major research areas in statistics and data science. By integrating multiple network representations with rich textual metadata, StatCite provides a valuable resource for statistical analysis, knowledge discovery, and data-driven studies of scientific literature.

cs.DL↗

arXiv2026

CofactVLA: Deconfounding Vision-Language-Action Models via Counterfactual Intervention

Vision-Language-Action (VLA) models have driven significant progress in robotic manipulation, yet they fundamentally struggle with the vision-override phenomenon. Driven by the severe modality imbalance between dense visual streams and sparse linguistic instructions, VLAs frequently fall prey to causal confusion. Instead of treating language as the primary causal driver, the policy entirely bypasses the original instruction by overfitting to spurious visual confounders, such as prominent objects or familiar layouts. To systematically alleviate this bias, we formalize the process of action generation as a Dual-path Deconfounding Graph (DDG) and propose CofactVLA, a novel causal intervention framework. By dynamically constructing a language-masked counterfactual branch within a single forward pass, CofactVLA isolates and neutralizes visual confounders through two synergistic mechanisms. First, Action-Level Orthogonal Projection Guidance (OPG) geometrically projects the factual velocity field away from the counterfactual visual bias during continuous flow matching, extracting the pure semantic intent. Second, Feature-Level Counterfactual Covariance Reduction (CCR) mathematically deconfounds latent representations by penalizing the positive eigenspace of the covariance difference, explicitly suppressing dominant visual shortcuts while preserving the causal language intent. Extensive experiments demonstrate that CofactVLA establishes a new state-of-the-art across diverse simulation benchmarks. Beyond simulation, real-world robot experiments demonstrate the causal efficacy of our method in bridging the generalization gap, yielding a 52.3\% absolute success rate gain under out-of-distribution scenarios.

cs.CV↗

arXiv2026

Agentic Reinforcement Learning with Observation-Calibrated Self-Distillation

Large language model agents are commonly trained through reinforcement learning with sparse trajectory-level rewards, which offer limited guidance on how strongly individual tokens should be updated. On-Policy Self-Distillation (OPSD) addresses this by re-scoring generated tokens under a privileged replay view to obtain dense, token-level supervision. However, we identify a confounding issue: the resulting support may reflect both the privileged information contained in the replay view and score shifts induced by the replay scaffold, making it difficult to attribute the support specifically to that information. This issue is especially pronounced when future environment observations serve as privileged information, since replaying them requires reconstructing an extended scaffold that itself perturbs token scores. To resolve this confounding, we propose Observation-Calibrated Self-Distillation (OCSD), which contrasts two structurally matched replay views, Full and Observation-Ablated, differing only in whether the actual future observation is present, to derive an observation residual that discounts score changes shared by the replay scaffold. OCSD then applies this residual to modulate token-level GRPO updates at high-uncertainty steps, while preserving the trajectory-level update direction. Experiments on ALFWorld, WebShop, and Search-QA across three Qwen3 model scales show that OCSD consistently outperforms strong baselines. Diagnostic analyses further confirm that the calibrated residual aligns better with local environment feedback. Our code is publicly available at https://github.com/yiy1x/OCSD.

cs.LG↗

arXiv2026

Calibrate Before Reason: Robust Visual Token Reduction against Semantic Drift in VLMs

Large Vision-Language Models (VLMs) suffer from prohibitive inference overhead due to long sequences of visual tokens. However, existing visual token reduction methods mainly improve efficiency by pruning or compressing redundant tokens without examining whether the resulting representation remains semantically consistent with the original representation. Mapping the original N-token visual sequence to K tokens may discard, dilute, or misassign critical visual cues, triggering severe semantic drift that deviates the VLM's understanding. In this paper, we first introduce the principle of 'Calibrate Before Reason' to visual token reduction and propose CaRe, a training-free robust framework that calibrates compact visual representations before reasoning to preserve semantic fidelity in VLMs. CaRe consists of two mutually complementary modules: 1) Perturbation-Robust Calibration Anchoring, which identifies calibration anchors with stable model-side influence under multi-directional perturbations; 2) Confidence-Gated Token Calibration, which extracts reliable calibration signals from unselected tokens and injects them into anchors. Extensive evaluations across diverse VLM architectures and benchmarks verify that CaRe outperforms state-of-the-art token reduction baselines. While pruning 94.4% of visual tokens, our method retains 96.4% of the original full-token performance, delivering up to 2.30 times faster end-to-end inference speed relative to unpruned vanilla models.

cs.CV↗

arXiv2026

Scores Are Not Decisions: Cost-Aware Stopping for Tool Acquisition in LLM Agents

As LLM agents increasingly depend on diverse external services such as search engines, databases, and connectors, agent harnesses face a fundamental tool-selection challenge: acquiring too few tools leaves the task under-informed, while too many adds cost, context load, and privacy exposure. Routers and retrievers can rank candidate tools by relevance, but a ranking alone does not determine how many are worth selecting. Existing approaches leave acquisition under heterogeneous costs unaddressed. We formulate this decision as cost-aware marginal decision-focused stopping (CAM-DF) over ranked tool prefixes, with CAM-DF-lite as a compact interpretable variant. We train directly on the offline gap between stopping now and the best continuation: its sign labels the decision, its magnitude weights each error by the payoff at stake. We prove this objective is Bayes-aligned with the stopping target and that score-only rules are suboptimal under heterogeneous costs. We evaluate on 1,343 tasks across five tool-use domains. On $τ$-bench Retail, CAM-DF attains the highest payoff among deployable methods, with gains over a predict-then-threshold baseline across all five ranking sources and two cost regimes. Our approach is state-of-the-art under heterogeneous costs and high cost pressure, with larger gains under weaker rankings. In live execution, CAM-DF exposes the agent to 37\% fewer tools than full access while maintaining comparable task success. The CAM-DF family is a lightweight pre-execution plugin that turns existing tool rankings into lower-cost acquisition decisions without fine-tuning the underlying LLM.

cs.LG↗

arXiv2026

Regime-Adaptive Bayesian Optimization via Dirichlet Process Mixtures of Gaussian Processes

Standard Bayesian Optimization (BO) assumes uniform smoothness across the search space an assumption violated in multi-regime problems such as molecular conformation search through distinct energy basins or drug discovery across heterogeneous molecular scaffolds. A single GP either oversmooths sharp transitions or hallucinates noise in smooth regions, yielding miscalibrated uncertainty. We propose RAMBO, a Dirichlet Process Mixture of Gaussian Processes that automatically discovers latent regimes during optimization, each modeled by an independent GP with locally-optimized hyperparameters. We derive collapsed Gibbs sampling that analytically marginalizes latent functions for efficient inference, and introduce adaptive concentration parameter scheduling for coarse-to-fine regime discovery. Our acquisition functions decompose uncertainty into intra-regime and inter-regime components. Experiments on synthetic benchmarks and real-world applications, including molecular conformer optimization, virtual screening for drug discovery, and fusion reactor design, demonstrate consistent improvements over state-of-the-art baselines on multi-regime objectives.

cs.LG↗

arXiv2026

A Catalogue of Topological Moiré Bands in Twisted Semiconductors

Twisted two-dimensional semiconductors provide a route to flat and topological moiré minibands, but systematic principles for organizing their material dependence have remained unclear. Here, we establish a high-throughput framework that integrates structural relaxation, first-principles electronic structure calculations, and moiré band topology. We apply this framework to 43 experimentally realized monolayers and 91 symmetry-inequivalent bilayer prototypes, yielding over 1,000 angle-resolved moiré electronic band structures. This database reveals that the low-energy moiré electronic structure is organized primarily by the valley character of the parent band edge together with stacking symmetry. In $Γ$-valley systems, the miniband width usually follows a nearly quadratic twist-angle scaling, consistent with a folding-dominated kinetic-energy scale. In $K$-valley systems, stacking-controlled interlayer hybridization governs whether parent Berry curvature is redistributed into isolated valley Chern minibands. By contrast, $M$-valley systems form a more material-specific class associated with anisotropic and symmetry-constrained band folding. The same valley-and-stacking hierarchy rationalizes the emergence or suppression of $\mathbb{Z}_2$ minibands, and surface termination in Janus bilayers provides a microscopic knob for changing the relevant valley character. These results establish a materials-level organizing principle for designing flat and topological moiré bands in twisted semiconductors.

cond-mat.mtrl-sci↗

arXiv2026

Verbalized Particle Posterior: Bayesian Inference over Natural Language Hypotheses

Verbalized Machine Learning (VML) parameterizes a model as a natural-language prompt that an LLM evaluates as f(x; theta). The framework is interpretable, but it commits to a single hypothesis with no measure of uncertainty, and that hypothesis varies substantially across optimization runs on the same data. We propose the Verbalized Particle Posterior (VPP), which treats verbalized learning as a Bayesian inference problem: maintain a population of natural-language hypotheses as particles, update them with Metropolis-Hastings (VPP-MH) or Sequential Monte Carlo (VPP-SMC), and predict by Bayesian model averaging. Both algorithms treat the LLM as a black box, requiring no access to logits or gradients. A distinctive consequence follows. In classical Bayesian learning, model selection sits outside the posterior; in VPP both model structure and parameters share a single language space, and the posterior ranges over both. We evaluate VPP on regression, classification, and rule-discovery benchmarks. It improves over a single VML run on every benchmark and matches or exceeds an oracle-best ensemble of independent VML runs on most, while eliminating the catastrophic single-run failures that VML occasionally produces. Because each particle is a human-readable hypothesis, the posterior is itself something a reader can inspect, seeing in plain text which explanations the data supported and which it ruled out.

cs.LG↗