arXiv subjects

Jian Wu

Publications and source records attributed to Jian Wu.

At least 109 records · Page 6Linked to original sources

arXiv2025

WST: Weakly Supervised Transducer for Automatic Speech Recognition

The Recurrent Neural Network-Transducer (RNN-T) is widely adopted in end-to-end (E2E) automatic speech recognition (ASR) tasks but depends heavily on large-scale, high-quality annotated data, which are often costly and difficult to obtain. To mitigate this reliance, we propose a Weakly Supervised Transducer (WST), which integrates a flexible training graph designed to robustly handle errors in the transcripts without requiring additional confidence estimation or auxiliary pre-trained models. Empirical evaluations on synthetic and industrial datasets reveal that WST effectively maintains performance even with transcription error rates of up to 70%, consistently outperforming existing Connectionist Temporal Classification (CTC)-based weakly supervised approaches, such as Bypass Temporal Classification (BTC) and Omni-Temporal Classification (OTC). These results demonstrate the practical utility and robustness of WST in realistic ASR settings. The implementation will be publicly available.

cs.CL↗

arXiv2025

Hulu-Med: A Transparent Generalist Model towards Holistic Medical Vision-Language Understanding

Real-world clinical decision-making requires integrating heterogeneous data, including medical text, 2D images, 3D volumes, and videos, while existing AI systems fail to unify all these signals, limiting their utility. In this paper, we introduce Hulu-Med, a transparent, generalist medical Vision-Language Model (VLM) designed to unify language-only, 2D/3D vision-language, and video understanding within a single architecture. Hulu-Med is trained on a curated corpus of 16.7 million samples, comprising exclusively public or synthetic data, spanning 12 major anatomical systems and 14 medical imaging modalities. Hulu-Med employs a medical-aware token-reduction strategy that prunes redundant visual tokens, achieving up to a 55% reduction for 3D and video inputs, improving cross-modal efficiency, and enabling training at 7B-32B parameter scales in approximately 4,000-40,000 GPU hours. Across 30 public in-domain and out-of-domain medical benchmarks-covering text reasoning, visual question answering, report generation, multilingual dialogue, video understanding, and rare disease diagnosis-Hulu-Med surpasses existing open-source models on 27 of 30 benchmarks and outperforms proprietary systems such as GPT-4o on 16 benchmarks. Despite being a VLM, Hulu-Med outperforms GPT-4o and matches GPT-o1 on the text-only HealthBench. For the first time in the community, we provide a fully transparent, reproducible and cost-effective pipeline for holistic medical vision-language understanding by releasing our end-to-end data curation, training procedures, and model parameters. Code and models are available at https://github.com/ZJUI-AI4H/Hulu-Med.

cs.CV↗

arXiv2025

LongWeave: A Long-Form Generation Benchmark Bridging Real-World Relevance and Verifiability

Generating long, informative, and factual outputs remains a major challenge for Large Language Models (LLMs). Existing benchmarks for long-form generation typically assess real-world queries with hard-to-verify metrics or use synthetic setups that ease evaluation but overlook real-world intricacies. In this paper, we introduce \textbf{LongWeave}, which balances real-world and verifiable assessment with Constraint-Verifier Evaluation (CoV-Eval). CoV-Eval constructs tasks by first defining verifiable targets within real-world scenarios, then systematically generating corresponding queries, textual materials, and constraints based on these targets. This ensures that tasks are both realistic and objectively assessable, enabling rigorous assessment of model capabilities in meeting complex real-world constraints. LongWeave supports customizable input/output lengths (up to 64K/8K tokens) across seven distinct tasks. Evaluation on 23 LLMs shows that even state-of-the-art models encounter significant challenges in long-form generation as real-world complexity and output length increase.

cs.CL↗

arXiv2025

3D-RAD: A Comprehensive 3D Radiology Med-VQA Dataset with Multi-Temporal Analysis and Diverse Diagnostic Tasks

Medical Visual Question Answering (Med-VQA) holds significant potential for clinical decision support, yet existing efforts primarily focus on 2D imaging with limited task diversity. This paper presents 3D-RAD, a large-scale dataset designed to advance 3D Med-VQA using radiology CT scans. The 3D-RAD dataset encompasses six diverse VQA tasks: anomaly detection, image observation, medical computation, existence detection, static temporal diagnosis, and longitudinal temporal diagnosis. It supports both open- and closed-ended questions while introducing complex reasoning challenges, including computational tasks and multi-stage temporal analysis, to enable comprehensive benchmarking. Extensive evaluations demonstrate that existing vision-language models (VLMs), especially medical VLMs exhibit limited generalization, particularly in multi-temporal tasks, underscoring the challenges of real-world 3D diagnostic reasoning. To drive future advancements, we release a high-quality training set 3D-RAD-T of 136,195 expert-aligned samples, showing that fine-tuning on this dataset could significantly enhance model performance. Our dataset and code, aiming to catalyze multimodal medical AI research and establish a robust foundation for 3D medical visual understanding, are publicly available at https://github.com/Tang-xiaoxiao/3D-RAD.

cs.CV↗

arXiv2025

Experimental Multipartite Entanglement Detection With Minimal-Size Correlations

Multiparticle entanglement is a valuable resource for quantum technologies, including measurement based quantum computing, quantum secret sharing, and a variety of quantum sensing applications. The direct way to detect this resource is to observe correlations arising from local measurements performed simultaneously on all particles. However, this approach is increasingly vulnerable to measurement imperfections when the number of particles grows, and becomes unfeasible for large-scale entangled states. It is therefore crucial to devise detection methods that minimize the number of simultaneously measured particles. Here we provide the first experimental demonstration of multipartite entanglement detection with minimal-size correlations, showing that our setup is robust to misalignment of the local measurement bases and enables the certification of genuine multipartite entanglement in a regime where the direct approach fails. Overall, our results indicate a promising route to the experimental detection of genuine multipartite entanglement in large-scale entangled states.

quant-ph↗

arXiv2025

Switching between Skyrmions and Yoshimori Spin Spirals via Li Absorption in Janus Magnets

Chiral magnetic textures have attracted considerable attention owing to their topological properties and potential applications in spintronic devices. Here, we employ first-principles calculations together with atomic spin dynamics simulations to explore the switching between skyrmions and Yoshimori-type spin spirals induced by Li adsorption in Janus two-dimensional (2D) CrTeSe. We show that selective Li adsorption on either the Se- or Te-terminated surface stabilizes distinct magnetic phases: Li adsorption on the Se side favors a Yoshimori-type spin spiral, whereas adsorption on the Te side stabilizes the skyrmionic state. This contrast originates from site-dependent modifications of exchange interactions, magnetic anisotropy (MA), and the Dzyaloshinskii-Moriya interaction (DMI). In addition, the response of magnetic textures to out-of-plane magnetic fields differs strongly between the two systems. These results demonstrate that surface adsorption provides an effective strategy for reversible control of chiral magnetic states in 2D magnets, while also offering fundamental insights into the competing interactions that govern the stability of skyrmions and Yoshimori spin spirals. Our findings highlight the potential of Janus 2D materials as a versatile platform for engineering tunable spintronic devices.

cond-mat.mtrl-sci↗

arXiv2025

Deep Literature Survey Automation with an Iterative Workflow

Automatic literature survey generation has attracted increasing attention, yet most existing systems follow a one-shot paradigm, where a large set of papers is retrieved at once and a static outline is generated before drafting. This design often leads to noisy retrieval, fragmented structures, and context overload, ultimately limiting survey quality. Inspired by the iterative reading process of human researchers, we propose \ours, a framework based on recurrent outline generation, in which a planning agent incrementally retrieves, reads, and updates the outline to ensure both exploration and coherence. To provide faithful paper-level grounding, we design paper cards that distill each paper into its contributions, methods, and findings, and introduce a review-and-refine loop with visualization enhancement to improve textual flow and integrate multimodal elements such as figures and tables. Experiments on both established and emerging topics show that \ours\ substantially outperforms state-of-the-art baselines in content coverage, structural coherence, and citation quality, while producing more accessible and better-organized surveys. To provide a more reliable assessment of such improvements, we further introduce Survey-Arena, a pairwise benchmark that complements absolute scoring and more clearly positions machine-generated surveys relative to human-written ones. The code is available at https://github.com/HancCui/IterSurvey\_Autosurveyv2.

cs.CL↗

arXiv2025

TranSUN: A Preemptive Paradigm to Eradicate Retransformation Bias Intrinsically from Regression Models in Recommender Systems

Regression models are crucial in recommender systems. However, retransformation bias problem has been conspicuously neglected within the community. While many works in other fields have devised effective bias correction methods, all of them are post-hoc cures externally to the model, facing practical challenges when applied to real-world recommender systems. Hence, we propose a preemptive paradigm to eradicate the bias intrinsically from the models via minor model refinement. Specifically, a novel TranSUN method is proposed with a joint bias learning manner to offer theoretically guaranteed unbiasedness under empirical superior convergence. It is further generalized into a novel generic regression model family, termed Generalized TranSUN (GTS), which not only offers more theoretical insights but also serves as a generic framework for flexibly developing various bias-free models. Comprehensive experimental results demonstrate the superiority of our methods across data from various domains, which have been successfully deployed in two real-world industrial recommendation scenarios, i.e. product and short video recommendation scenarios in Guess What You Like business domain in the homepage of Taobao App (a leading e-commerce platform with DAU > 300M), to serve the major online traffic.

cs.IR↗

arXiv2025

Tunable Chern Insulator States with Coexisting Magnonic and Electronic Topology in 2D Honeycomb Kitaev Ferromagnets

The coexistence of topological magnons and electrons in magnetic materials presents a compelling route toward developing low-dissipation, multifunctional spintronic devices. However, material systems enabling their simultaneous realization and control remain largely unexplored. Here, we propose the coexistence and concurrent tunability of magnonic and electronic Chern insulator phases in Kitaev magnets and use MnBr$_{3}$ monolayer as a prototype. We find the significant Kitaev interaction in MnBr$_{3}$ induces the magnonic Chern insulator phase, manifesting as the magnon thermal Hall effect. Concurrently, MnBr$_{3}$ exhibits the quantum anomalous Hall effect driven by its electronic Chern insulator phase. Crucially, we demonstrate that these dual topological phases can be simultaneously controlled by reorienting the in-plane spins with an external magnetic field. Our findings not only deepen the fundamental understanding of spin excitations in Kitaev magnets but also provide a promising platform for exploring the interplay between electronic and magnonic topology.

cond-mat.mes-hall↗

arXiv2025

League: Leaderboard Generation on Demand

This paper introduces Leaderboard Auto Generation (LAG), a novel and well-organized framework for automatic generation of leaderboards on a given research topic in rapidly evolving fields like Artificial Intelligence (AI). Faced with a large number of AI papers updated daily, it becomes difficult for researchers to track every paper's proposed methods, experimental results, and settings, prompting the need for efficient automatic leaderboard construction. While large language models (LLMs) offer promise in automating this process, challenges such as multi-document summarization, leaderboard generation, and experiment fair comparison still remain under exploration. LAG solves these challenges through a systematic approach that involves the paper collection, experiment results extraction and integration, leaderboard generation, and quality evaluation. Our contributions include a comprehensive solution to the leaderboard construction problem, a reliable evaluation method, and experimental results showing the high quality of leaderboards.

cs.CL↗

arXiv2025

Interactive Recommendation Agent with Active User Commands

Traditional recommender systems rely on passive feedback mechanisms that limit users to simple choices such as like and dislike. However, these coarse-grained signals fail to capture users' nuanced behavior motivations and intentions. In turn, current systems cannot also distinguish which specific item attributes drive user satisfaction or dissatisfaction, resulting in inaccurate preference modeling. These fundamental limitations create a persistent gap between user intentions and system interpretations, ultimately undermining user satisfaction and harming system effectiveness. To address these limitations, we introduce the Interactive Recommendation Feed (IRF), a pioneering paradigm that enables natural language commands within mainstream recommendation feeds. Unlike traditional systems that confine users to passive implicit behavioral influence, IRF empowers active explicit control over recommendation policies through real-time linguistic commands. To support this paradigm, we develop RecBot, a dual-agent architecture where a Parser Agent transforms linguistic expressions into structured preferences and a Planner Agent dynamically orchestrates adaptive tool chains for on-the-fly policy adjustment. To enable practical deployment, we employ simulation-augmented knowledge distillation to achieve efficient performance while maintaining strong reasoning capabilities. Through extensive offline and long-term online experiments, RecBot shows significant improvements in both user satisfaction and business outcomes.

cs.IR↗

arXiv2025

DentVLM: A Multimodal Vision-Language Model for Comprehensive Dental Diagnosis and Enhanced Clinical Practice

Diagnosing and managing oral diseases necessitate advanced visual interpretation across diverse imaging modalities and integrated information synthesis. While current AI models excel at isolated tasks, they often fall short in addressing the complex, multimodal requirements of comprehensive clinical dental practice. Here we introduce DentVLM, a multimodal vision-language model engineered for expert-level oral disease diagnosis. DentVLM was developed using a comprehensive, large-scale, bilingual dataset of 110,447 images and 2.46 million visual question-answering (VQA) pairs. The model is capable of interpreting seven 2D oral imaging modalities across 36 diagnostic tasks, significantly outperforming leading proprietary and open-source models by 19.6% higher accuracy for oral diseases and 27.9% for malocclusions. In a clinical study involving 25 dentists, evaluating 1,946 patients and encompassing 3,105 QA pairs, DentVLM surpassed the diagnostic performance of 13 junior dentists on 21 of 36 tasks and exceeded that of 12 senior dentists on 12 of 36 tasks. When integrated into a collaborative workflow, DentVLM elevated junior dentists' performance to senior levels and reduced diagnostic time for all practitioners by 15-22%. Furthermore, DentVLM exhibited promising performance across three practical utility scenarios, including home-based dental health management, hospital-based intelligent diagnosis and multi-agent collaborative interaction. These findings establish DentVLM as a robust clinical decision support tool, poised to enhance primary dental care, mitigate provider-patient imbalances, and democratize access to specialized medical expertise within the field of dentistry.

cs.CV↗

arXiv2025

RecIS: Sparse to Dense, A Unified Training Framework for Recommendation Models

In this paper, we propose RecIS, a unified Sparse-Dense training framework designed to achieve two primary goals: 1. Unified Framework To create a Unified sparse-dense training framework based on the PyTorch ecosystem that meets the training needs of industrial-grade recommendation models that integrated with large models. 2.System Optimization To optimize the sparse component, offering superior efficiency over the TensorFlow-based recommendation models. The dense component, meanwhile, leverages existing optimization technologies within the PyTorch ecosystem. Currently, RecIS is being used in Alibaba for numerous large-model enhanced recommendation training tasks, and some traditional sparse models have also begun training in it.

cs.IR↗

arXiv2025

Femtosecond low-threshold all-optical switching enabled by giant broadband optical nonlinearity from heteroatom doping

Ultrafast all-optical switching (AOS) is pivotal for advancing integrated photonic devices, from high-speed photonic information processing to next generation all-optical computing and communication networks. However, conventional nonlinear materials suffer from sluggish response time, high power threshold, weak and narrow-bandwidth optical nonlinearities, critically limiting their viability. Here, we report a heteroatom engineering strategy to overcome these limitations by designing zero-dimensional nitrogen-doped carbon quantum dots (N-CQDs) with nonlinear optical performance far exceeding the state-of-the-art. Leveraging spatial self-phase modulation (SSPM) and ultrafast pump-probe technique, we first demonstrate an all-in-one AOS platform, where femtosecond laser pulses serve dual roles as control and signal beams. The AOS simultaneously realizes ultrafast response time (520 fs), ultralow threshold energy (2.2 Wcm-2), and giant nonlinear refraction indexes (10-5 cm2/W) in the wide spectral range (400-1064 nm), yielding performance surpassing state-of-the-art nonlinear carbon materials (i.e. carbon nanotube) by orders of magnitude. Spectroscopic and bandgap analyses attribute these exotic performances to enhanced n-pi interaction enabled by nitrogen doping, which amplifies nonlinear polarization dynamics. Crucially, ultrafast fluorescence spectroscopy reveals a large two-photon absorption cross-section of the N-CQDs, challenging the conventional cognition that broadband SSPM necessitates single-photon excitation. This discovery unveils a multi-channel AOS rooted in synergistic single-photon and two-photon processes.. This work demonstrates a new paradigm for achieving ultrafast, broadband, and energy-efficient AOS by heteroatom doping engineering.

physics.optics↗

arXiv2025

LAMDAS: LLM as an Implicit Classifier for Domain-specific Data Selection

Adapting large language models (LLMs) to specific domains often faces a critical bottleneck: the scarcity of high-quality, human-curated data. While large volumes of unchecked data are readily available, indiscriminately using them for fine-tuning risks introducing noise and degrading performance. Strategic data selection is thus crucial, requiring a method that is both accurate and efficient. Existing approaches, categorized as similarity-based and direct optimization methods, struggle to simultaneously achieve these goals. In this paper, we introduce LAMDAS (LLM As an iMplicit classifier for domain-specific DAta Selection), a novel approach that leverages the pre-trained LLM itself as an implicit classifier, thereby bypassing explicit feature engineering and computationally intensive optimization process. LAMDAS reframes data selection as a one-class classification problem, identifying candidate data that "belongs" to the target domain defined by a small reference dataset. Extensive experimental results demonstrate that LAMDAS not only exceeds the performance of full-data training using a fraction of the data but also outperforms nine state-of-the-art (SOTA) baselines under various scenarios. Furthermore, LAMDAS achieves the most compelling balance between performance gains and computational efficiency compared to all evaluated baselines.

cs.CL↗

arXiv2025

Icon$^{2}$: Aligning Large Language Models Using Self-Synthetic Preference Data via Inherent Regulation

Large Language Models (LLMs) require high quality preference datasets to align with human preferences. However, conventional methods for constructing such datasets face significant challenges: reliance on pre-collected instructions often leads to distribution mismatches with target models, while the need for sampling multiple stochastic responses introduces substantial computational overhead. In this work, we explore a paradigm shift by leveraging inherent regulation of LLMs' representation space for efficient and tailored preference dataset construction, named Icon$^{2}$. Specifically, it first extracts layer-wise direction vectors to encode sophisticated human preferences and then uses these vectors to filter self-synthesized instructions based on their inherent consistency. During decoding, bidirectional inherent control is applied to steer token representations, enabling the precise generation of response pairs with clear alignment distinctions. Experimental results demonstrate significant improvements in both alignment and efficiency. Llama3-8B and Qwen2-7B achieve an average win rate improvement of 13.89% on AlpacaEval 2.0 and 13.45% on Arena-Hard, while reducing computational costs by up to 48.1%.

cs.CL↗

arXiv2025

Toward Robust URL Extraction for Open Science: A Study of arXiv File Formats and Temporal Trends

In this work, we study how URL extraction results depend on input format. We compiled a pilot dataset by extracting URLs from 10 arXiv papers and used the same heuristic method to extract URLs from four formats derived from the PDF files or the source LaTeX files. We found that accurate and complete URL extraction from any single format or a combination of multiple formats is challenging, with the best F1-score of 0.71. Using the pilot dataset, we evaluate extraction performance across formats and show that structured formats like HTML and XML produce more accurate results than PDFs or Text. Combining multiple formats improves coverage, especially when targeting research-critical resources. We further apply URL extraction on two tasks, namely classifying URLs into open-access datasets and software and the others, and analyzing the trend of URLs usage in arXiv papers from 1992 to 2024. These results suggest that using a combination of multiple formats achieves better performance on URL extraction than a single format, and the number of URLs in arXiv papers has been steadily increasing since 1992 to 2014 and has been drastically increasing from 2014 to 2024. The dataset and the Jupyter notebooks used for the preliminary analysis are publicly available at https://github.com/lamps-lab/arxiv-urls

cs.DL↗

arXiv2025

Over-Barrier Ionization Dynamics Studied by Backpropagation

Tunneling and over-barrier ionization are the primary processes of strong-field ionization of atoms and molecules. While the dynamics of tunneling ionization have been extensively studied, exploration of over-barrier ionization dynamics has remained a significant challenge. In this study, we investigate the dynamics of over-barrier ionization using the backpropagation method specifically adapted for this context. By analyzing the topology of the backpropagating trajectories, we differentiate the contributions of tunneling and over-barrier ionizations to the distributions of photoelectron momentum and ionization time. While the transition from tunneling to over-barrier ionization is known to depend on the field strength, our results reveal that it is also influenced by the initial transverse momentum of the outgoing electron. We clarify how ionization probabilities vary with intensity for each mechanism, highlighting a competitive relationship between them. We further find that accounting for the Stark shift is essential for accurately determining the threshold field strength for over-barrier ionization. Our work provides a detailed understanding of the dynamics of over-barrier ionization and lays the groundwork for exploring new mechanisms in intense laser-matter interactions.

physics.atom-ph↗