Search arXiv⌕ Search

arXiv subjects

Wei Wang

Publications and source records attributed to Wei Wang.

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

Beyond Discrete Tokens: Preserving Numerical Continuity in Medical Vision-Language Models

Objective: To develop a unified vision-language model for prompt-conditioned quantitative ultrasound measurement that addresses the mismatch between continuous measurement targets and discrete token-based numerical prediction. Methods: This article presents HAD-VLM, a hybrid autoregressive-diffusion framework that combines autoregressive text generation with continuous measurement prediction within a shared backbone. Visual prompt tuning adapts the vision encoder to ultrasound images. A diffusion process predicts a five-component representation comprising caliper endpoint coordinates and measurement length. A post-processing strategy combining the proposed coordinate-length consistency mechanism with ensemble prediction is introduced to refine measurement estimates. We evaluate four echocardiographic measurement tasks on EchoNet-LVH, with additional assessments of medical question answering and X-ray vertebral measurement. Results: HAD-VLM achieves mean absolute errors of 1.36, 2.77, 3.05, and 1.16 mm for IVSd, LVIDd, LVIDs, and LVPWd, respectively, reducing errors by 21.4-37.6\% relative to the Vicuna-based autoregressive baseline. It also outperforms fine-tuned Qwen3-VL and direct MLP regression across all four tasks. Medical question answering performance is preserved on the evaluated dataset. For X-ray vertebral measurement, HAD-VLM reduces mean absolute error from 11.4 to 6.7 pixels. Conclusion: Integrating autoregressive language generation with diffusion-based continuous prediction improves quantitative medical image measurement while retaining prompt-conditioned task flexibility. Significance: HAD-VLM provides a unified approach to combining language interaction with anatomically grounded numerical prediction for quantitative medical image analysis.

cs.CV↗
arXiv2026

PreCoMem: Predictive Cognitive Memory for Self-Evolving Long-Term Dialogue Agents

Long-horizon dialogue agents need to keep user memories stable enough for personalization while adapting when the user genuinely changes. Existing memory systems often append every observation or overwrite old entries by heuristic rules, which makes transient noise hard to distinguish from real belief drift. We introduce PreCoMem, a memory consolidation framework inspired by predictive coding that represents memory as confidence-weighted beliefs about the user's state. Its core signal, Effective Surprisal, combines semantic deviation and contradiction evidence, then downweights this signal when retrieval is diffuse and unreliable. According to effective surprisal, a two-threshold gate maps each turn to one of three deterministic updates: MAINTAIN reinforces supported beliefs, PROFILE stores ambiguous cues as low-confidence hypotheses, and CORRECT softly decays reliably contradicted facts. The PROFILE stage acts as an explicit waiting room, preventing premature commitment to noisy evidence. Experiments on LoCoMo, LongMemEval-S, and PersonaMem-v2 show consistent state-of-the-art accuracy; our new DiMoBench confirms that PreCoMem adapts to non-stationary belief drift without over-reacting to noise. The code is available at https://github.com/ZhongWishing/PreCoMem.

cs.AI↗
arXiv2026

Uni-Mem: Learning a Long-Horizon Agentic Memory Manager with Unified Online Policy Optimization

Long-term agent memory is shifting from predefined workflows toward learnable policies for memory construction and retrieval. Existing learning-based methods either optimize construction alone or combine retrieval-side online reinforcement learning (RL) with offline construction supervision, limiting co-adaptation between the two stages. In this paper, we explore end-to-end online RL for a unified memory manager (Uni-Mem) that learns both capabilities under a shared policy. However, joint optimization with existing RL algorithms such as GRPO faces confounded group comparisons and sparse feedback over long memory-management trajectories. To this end, we propose Stage-Decoupled GRPO, which alternates construction and retrieval optimization while controlling variation from the other stage within each rollout group, yielding stage-specific learning signals. We further introduce Event-Wise Credit Assignment (ECA), which localizes feedback to responsible events and propagates event-level advantages to their constituent operations. Experiments on LoCoMo and LongMemEval show that Uni-Mem outperforms the strongest RL-based memory, demonstrating the effectiveness of jointly learning memory construction and retrieval via Stage-Decoupled GRPO and ECA.

cs.CL↗
arXiv2026

Finite Combinatorics and Fragments of Arithmetic

In fragments of first order arithmetic, definable maps on finite domains could behave very differently from finite maps. Here combinatorial properties of $Σ_{n+1}$-definable maps on finite domains are compared in the absence of $BΣ_{n+1}$. It is shown that $\mathrm{GPHP}(Σ_{n+1})$ (the $Σ_{n+1}$-instance of Kaye's General Pigeonhole Principle) lies strictly between $\mathrm{CARD}(Σ_{n+1})$ and $\mathrm{WPHP}(Σ_{n+1})$ (Weak Pigeonhole Principle for $Σ_{n+1}$-maps), and also that $\mathrm{FRT}(Σ_{n+1})$ (Finite Ramsey's Theorem for $Σ_{n+1}$-maps) does not imply $\mathrm{WPHP}(Σ_{n+1})$.

math.LO↗
arXiv2026

Yuvion VL: A Multimodal Foundation Model for Adversarial Content and AI Safety

General-purpose models often struggle to reliably identify and understand real-world multimodal risks, largely due to the inherent multimodal adversarial nature of content and AI safety. We present Yuvion VL, a family of multimodal large language models purpose-built for content and AI safety, with both instruction-tuned and reasoning-oriented variants. Yuvion VL addresses this gap by treating safety as an inherently adversarial and multimodal problem and designing the entire pipeline around adversarial robustness. For data construction, we develop an automated pipeline integrating adversarial-aware data synthesis with multi-stage quality control, producing large-scale, high-quality multimodal samples augmented with domain knowledge and reasoning annotations. For training, we adopt a three-stage pipeline that includes continued pretraining for risk-concept cross-modal alignment, instruct post-training for production-grade safety tasks, and reasoning post-training for enhanced interpretability and performance in complex tasks. We further introduce Confuse-then-Contrast Fine-Tuning, a contrastive framework that mines model-specific confusions and constructs multi-image contrastive groups to enforce explicit discrimination of fine-grained visual-semantic elements, enabling the model to distinguish between visually similar cases with different safety implications in adversarial safety tasks. To support rigorous evaluation, we further introduce Yuvion VL RiskEval (YVRE), a collection of benchmarks covering diverse open and internal evaluations, with a focus on content and AI safety, adversarial robustness, and real-world capability requirements. Experiments show that Yuvion VL-32B achieves industry-leading safety performance, surpassing comparably sized open-source models and best closed-source commercial models, while maintaining comparable general capabilities.

cs.CV↗
arXiv2026

The Complexity of Weak Partition Connectivity in Hedgegraphs

We prove that deciding whether the weak partition connectivity (WPC) of an unweighted hedgegraph is at most a given integer is NP-complete, resolving an open question about its computational complexity. Hardness holds even for connected full-support split systems, where each hedge consists of two nonempty, vertex-disjoint hyperedges covering all vertices. We also obtain hardness on connected edge-colored simple graphs in which each color class is a spanning forest with exactly two nontrivial trees. On both classes, exact computation of fractional WPC is NP-hard, whereas hedge connectivity is computable exactly in polynomial time. A binary matrix formulation yields the reduction and a deterministic polynomial-time exact algorithm under a linear intersection condition on row supports after normalization and column compression. In particular, both integer and fractional WPC are computable in polynomial time when these supports have size at most two, whereas NP-hardness holds already when their size is bounded by three. For every fixed $b$, we give deterministic and randomized efficient polynomial-time approximation schemes on hedgegraphs in which each hedge has at most $b$ connected components, counting isolated vertices. Both schemes return one partition that simultaneously approximates integer and fractional WPC. Unless $\mathrm P=\mathrm{NP}$, neither objective admits a deterministic FPTAS, even on the two hard instance classes.

cs.DS↗
arXiv2026

SLM-SS: Token-Based Speech Separation with Speech Language Models

Speech separation (SS) has witnessed substantial progress in recent years with the advent of neural network-based methods, achieving notable improvements on signal-level evaluation metrics. Existing speech separation systems are mostly optimized with signal-level objectives, and often yield outputs that are acceptable in waveform metrics but suboptimal for intelligibility-sensitive downstream tasks. To address this gap, we propose SLM-SS, a token-based generative speech separation framework that models multi-talker mixtures with speech language models. We first discretize clean speech into multi-codebook token sequences with neural codecs and serialize multi-speaker targets via Serialized Output Training (SOT). An encoder-decoder autoregressive speech language model predicts the zero-order sequence to establish speaker ordering and coarse linguistic content, and a non-autoregressive module then refines higher-order residual codebooks for efficient reconstruction. We further introduce a per-talker non-autoregressive strategy that simplifies high-order modeling and improves the stability. On Libri2Mix, SLM-SS with per-talker NAR achieves $6.92\%$ WER, approaching codec-only reconstruction. The WavTokenizer variant eliminates NAR refinement and achieves the highest mean subjective quality score among evaluated separation systems. Its UTMOS and DNSMOS advantages over discriminative baselines persist on unseen noisy and Chinese mixtures without fine-tuning.

cs.SD↗
arXiv2026

AutoTailor: Automatic and Efficient Adaptive Model Deployment for Diverse Edge Devices

On-device machine learning (ML) has become a fundamental component of emerging mobile applications. Adaptive model deployment delivers efficient inference for heterogeneous device capabilities and performance requirements through customizing neural architectures. SuperNet-based approaches offer a promising solution by generating numerous model variants from a pre-trained ML model. However, applying SuperNet in existing frameworks requires tedious model-aware development and time-consuming hardware-aware profiling, which limits their practical adoption. We present AutoTailor, the first framework to enable automated, end-to-end SuperNet-based adaptive model deployment for edge devices. Unlike manual SuperNet construction, AutoTailor employs a computation graph-guided compilation approach to automatically transform user-provided ML models into SuperNets. To support efficient specialization, AutoTailor incorporates learning-free latency and accuracy predictors, enabling low-cost yet accurate performance prediction. Our extensive evaluations demonstrate that AutoTailor reduces the lines of code for SuperNet construction by 10.8--34.5$\times$, reduces predictor profiling time by 2.2--11.5$\times$, and achieves up to 15.60\% absolute accuracy improvement and 60.03\% latency reduction compared to state-of-the-art approaches across diverse models and devices. AutoTailor's source code is publicly available at \url{https://github.com/liu-mengyang/AutoTailor}.

cs.LG↗
arXiv2026

From Solvers to Research: Large Language Model-Driven Formal Mathematics at the Research Frontier

Recent developments in AI for Mathematics (AI4Math), especially Large Language Model (LLM)-driven theorem provers, has achieved remarkable success in formal proof generation for well-defined mathematical problems through Interactive Theorem Proving (ITP) languages. However, current systems remain fundamentally limited in tackling frontier research mathematics, such as discovering new theorems or resolving open conjectures, which are often open-ended, under-specified, and involve multiple layers of abstraction. We argue that the next leap in AI4Math systems requires a decisive shift from predefined problem-solvers to research agents that can address frontier mathematical challenges with rigorous formal mathematical reasoning. In this position paper, we provide a systematic review of the field, covering datasets, auto-formalization, and proof synthesis. More importantly, we identify core limitations of existing systems in serving as mathematical research agents, examining issues across datasets, relational structure, mathematical exploration, tool ecosystem, and human-AI collaboration, outlining a strategic road-map for the future of AI4Math.

cs.CL↗
arXiv2026

REATS: LLM Reasoning-based Ensemble Learning for Adaptive Time Series Forecasting

Due to the diversity of real-world time series, no single forecasting model consistently dominates across all samples. Ensemble learning addresses this by combining complementary model strengths, yet existing methods rely on fixed rules or black-box models based solely on numerical inputs, failing to leverage LLM reasoning for interpretable weighting decisions. We propose REATS, which leverages LLM reasoning capabilities as an intelligent ensemble router that jointly processes textual temporal pattern descriptions and numerical features to produce interpretable, sample-adaptive ensemble weights through chain-of-thought reasoning. To enable effective LLM-based ensembling, we study its key design choices and propose: (i) a structured input pipeline that transforms raw time series into hybrid textual--numerical representations with fixed token cost, enabling rule-based chain-of-thought construction without API dependency, augmented with retrieved similar-sample priors; (ii) a diverse multi-row weight supervision scheme coupled with a token-efficient percentage-table format that reduces numerical complexity and mitigates LLM hallucinations; and (iii) a two-stage fine-tuning framework combining SFT with GRPO, where a reciprocal reward mapping transforms the continuous unbounded MSE gap into bounded signals with amplified near-oracle sensitivity, addressing the uniform sensitivity and outlier-dominated advantage compression inherent in naive reward designs for regression-based GRPO. Experiments on eight benchmarks demonstrate that REATS outperforms competitive ensemble baselines while providing natural language explanations and demonstrating strong transfer learning and out-of-domain generalization to unseen candidate models.

cs.LG↗
arXiv2026

Jumps and Iterated Jumps of Perfect Subtrees of Positive Trees

An infinite (binary) tree is \emph{positive} iff its set of paths is non-null. As an elementary observation, positives trees always have positive perfect subtrees. Given a positive tree of certain complexity, here we present several theorems concerning the distribution of the Turing jumps or iterated Turing jumps of (positive) perfect subtrees of the given tree in the universe of Turing degrees.

math.LO↗
arXiv2026

Simultaneous Sharp-Interface Limits for a Diffuse-Interface Landau--de Gennes Droplet Model

Nematic droplet models based on diffuse interfaces involve competing microscopic transition layers whose interaction is lost if the small parameters are sent to zero successively. We study this interaction for the diffuse-interface Landau--de Gennes model of Shi et al., with phase-field thickness $\varepsilon$ and exterior $Q$-tensor relaxation length $κ_\varepsilon$, and show that their relative scale selects the effective sharp-interface theory. Within a regular-interface class, we establish the corresponding sharp-interface limit for every limiting ratio $r=\lim_{\varepsilon\to0}κ/\varepsilon\in[0,+\infty]$. Three qualitatively different regimes emerge. When $r=+\infty$, the two transition layers separate asymptotically and the interfacial costs decouple, recovering the previously obtained iterated limit. When $r\in(0,+\infty)$, the layers overlap and produce a genuinely coupled, $r$-dependent anisotropic surface density. When $r=0$, finite energy forces the interior trace of $Q$ to vanish on the limiting interface. The comparable-scale regime requires a genuinely coupled interfacial analysis. Thus the exterior relaxation length is not merely a regularization parameter: together with the phase-field thickness, it determines the macroscopic surface energy and anchoring law of the sharp-interface limit.

math.AP↗
arXiv2026

Commutators of a class of $Ψ$-bounded oscillation operators on abstract weighted Orlicz-Morrey spaces endowed with ball-basis and applications

In our previous work \cite{SZ2026}, we established a class of abstract Orlicz-Morrey spaces endowed with a ball-basis and introduced the notion of \(Ψ\)-bounded oscillation operators. The present paper further develops this theory by investigating sparse domination and weighted estimates for such operators and their commutators within the framework of abstract Orlicz-Morrey spaces endowed with a ball-basis. We obtain pointwise sparse domination results for the operators and establish their boundedness on weighted Orlicz-Morrey spaces. The hypotheses of these conclusions differ from those in the classical Orlicz-Morrey setting and do not rely on dyadic decompositions of Euclidean structure. Moreover, we study the duality of these spaces and prove that the dual of the associated block space is exactly the Orlicz-Morrey space under consideration.

math.FA↗
arXiv2026

Elliptic closed geodesics on non-negatively curved Finsler manifolds

In this paper, we prove that if a compact simply-connected manifold $(M,F)$ with a bumpy irreversible Finsler metric $F$ whose flag curvature $K$ satisfies $K\ge 0$ and $H^*(M;\Q)\cong T_{d,n+1}(x)$ for some $d\in2\N$ and $n\in\N$ carries exactly $\frac{dn(n+1)}{2}$ (the number of closed geodesics in Katok's example) closed geodesics, then there are at least $2$ irrationally elliptic closed geodesics on $(M, F)$.

math.DG↗
arXiv2026

Recent Advances in Transverse-Momentum--Dependent Distributions in Lattice QCD

Unraveling the three-dimensional structure of hadrons is a central objective in modern nuclear and particle physics. Transverse Momentum Dependent parton distributions (TMDs) provide crucial insights into the interplay between partonic transverse motion and spin, and are indispensable for describing a wide range of high-energy processes at facilities such as JLab, COMPASS, the LHC, and the upcoming Electron-Ion Collider. In recent years, significant progress has been made in accessing TMDs from lattice QCD using novel approaches. Methods based on ratios of Mellin moments, large-momentum effective theory (LaMET), and Coulomb-gauge techniques have enabled first-principles investigations of the Collins-Soper kernel, intrinsic soft functions, TMD wave functions, and TMDPDFs. Recent lattice results for both unpolarized and spin-dependent distributions, including the Boer-Mulders function, demonstrate the feasibility of extracting transverse-momentum--dependent information directly from QCD. This paper reviews recent advances in lattice-QCD studies of TMDs, outlines remaining challenges such as continuum and infinite-momentum extrapolations, and discusses future directions toward precision determinations of hadron structure in three dimensions.

hep-lat↗
arXiv2026

Harnessing Trust in Directed Graphs: Redefining Robustness of Graph Learning

Existing research on robust Graph Neural Networks (GNNs) focuses predominantly on undirected graphs, overlooking the trustworthiness inherent in the link directions of directed graphs. In this work, we analyze the limitations of current approaches from both the attack and defense perspectives and redefine the robustness of GNNs in directed graphs. From the attack side, we introduce a more realistic directed graph attack setting that addresses the shortcomings of existing attacks. From the defense side, we propose a simple yet effective message-passing framework that serves as a plug-in layer to strengthen GNN robustness while avoiding a false sense of robustness. Our findings show that the trust signals encoded in directed graphs can be harnessed to substantially improve the robustness of GNNs. When combined with existing defense strategies, the framework achieves state-of-the-art robust performance alongside strong clean accuracy against both transfer and adaptive attacks. These results point to a novel and promising direction for strengthening the robustness of graph deep learning.

cs.LG↗
arXiv2026

Neural Block-wise Auto-exposure for Single-Shot High Dynamic Range Imaging

Block-wise varying exposures are effective for cap- turing appearance of a real-world high dynamic range (HDR) scene via a single-shot. However, brightness order reversal (BOR) artifacts, blocking artifacts and loss of details in highlight/shadow regions are possible issues for block-wise varying exposures. In this paper, a novel neural block-wise auto-exposure algorithm is proposed to address the issues. The proposed algorithm maps a temporary image and its exposure time to the exposure times of all non-overlapping 16 X 16 blocks. The temporary image can be either an sRGB or a raw image. Although the sRGB image is widely adopted in existing auto-exposure algorithms, exposure times of blocks in the highlight and shadow regions might not be accurately learned due to the limited radiance measurement capability of the sRGB image. This paper thus also explores the feasibility of temporary image as a raw image. Due to the better radiance measurement capability of the raw image, the exposure times of blocks are learned more accurately, and more details in the highlight/shadow regions of the real-world HDR scene are captured by the resultant HDR imaging system. Experiment results validate the proposed neural block-wise auto-exposure algorithm.

cs.CV↗