arXiv subjects

Han Wang

Publications and source records attributed to Han Wang.

At least 199 records · Page 11Linked to original sources

arXiv2025

Dynamic Double Space Tower

The Visual Question Answering (VQA) task requires the simultaneous understanding of image content and question semantics. However, existing methods often have difficulty handling complex reasoning scenarios due to insufficient cross-modal interaction and capturing the entity spatial relationships in the image.\cite{huang2023adaptive}\cite{liu2021comparing}\cite{guibas2021adaptive}\cite{zhang2022vsa}We studied a brand-new approach to replace the attention mechanism in order to enhance the reasoning ability of the model and its understanding of spatial relationships.Specifically, we propose a dynamic bidirectional spatial tower, which is divided into four layers to observe the image according to the principle of human gestalt vision. This naturally provides a powerful structural prior for the spatial organization between entities, enabling the model to no longer blindly search for relationships between pixels but make judgments based on more meaningful perceptual units. Change from "seeing images" to "perceiving and organizing image content".A large number of experiments have shown that our module can be used in any other multimodal model and achieve advanced results, demonstrating its potential in spatial relationship processing.Meanwhile, the multimodal visual question-answering model July trained by our method has achieved state-of-the-art results with only 3B parameters, especially on the question-answering dataset of spatial relations.

cs.CV↗

arXiv2025

Learning-Based Stable Optimal Control for Infinite-Time Nonlinear Regulation Problems

Infinite-time nonlinear optimal regulation control is widely utilized in aerospace engineering as a systematic method for synthesizing stable controllers. However, conventional methods often rely on linearization hypothesis, while recent learning-based approaches rarely consider stability guarantees. This paper proposes a learning-based framework to learn a stable optimal controller for nonlinear optimal regulation problems. First, leveraging the equivalence between Pontryagin Maximum Principle (PMP) and Hamilton-Jacobi-Bellman (HJB) equation, we improve the backward generation of optimal examples (BGOE) method for infinite-time optimal regulation problems. A state-transition-matrix-guided data generation method is then proposed to efficiently generate a complete dataset that covers the desired state space. Finally, we incorporate the Lyapunov stability condition into the learning framework, ensuring the stability of the learned optimal policy by jointly learning the optimal value function and control policy. Simulations on three nonlinear optimal regulation problems show that the learned optimal policy achieves near-optimal regulation control and the code is provided at https://github.com/wong-han/PaperNORC

eess.SY↗

arXiv2025

Predicting ICU In-Hospital Mortality Using Adaptive Transformer Layer Fusion

Early identification of high-risk ICU patients is crucial for directing limited medical resources. We introduce ALFIA (Adaptive Layer Fusion with Intelligent Attention), a modular, attention-based architecture that jointly trains LoRA (Low-Rank Adaptation) adapters and an adaptive layer-weighting mechanism to fuse multi-layer semantic features from a BERT backbone. Trained on our rigorous cw-24 (CriticalWindow-24) benchmark, ALFIA surpasses state-of-the-art tabular classifiers in AUPRC while preserving a balanced precision-recall profile. The embeddings produced by ALFIA's fusion module, capturing both fine-grained clinical cues and high-level concepts, enable seamless pairing with GBDTs (CatBoost/LightGBM) as ALFIA-boost, and deep neuro networks as ALFIA-nn, yielding additional performance gains. Our experiments confirm ALFIA's superior early-warning performance, by operating directly on routine clinical text, it furnishes clinicians with a convenient yet robust tool for risk stratification and timely intervention in critical-care settings.

cs.LG↗

arXiv2025

CLaMR: Contextualized Late-Interaction for Multimodal Content Retrieval

Online video web content is richly multimodal: a single video blends vision, speech, ambient audio, and on-screen text. Retrieval systems typically treat these modalities as independent retrieval sources, which can lead to noisy and subpar retrieval. We explore multimodal video content retrieval, where relevance can be scored from one particular modality or jointly across multiple modalities simultaneously. Consequently, an effective retriever must dynamically choose which modality (or set of modalities) best addresses the query. We introduce CLaMR, a multimodal, late-interaction retriever that jointly indexes 4 modalities: video frames, transcribed speech, on-screen text, and metadata. CLaMR jointly encodes all modalities with a unified multimodal backbone for improved contextualization and is trained to enhance dynamic modality selection via two key innovations. First, given the lack of training data for multimodal retrieval, we introduce MultiVENT 2.0++, a large-scale synthetic training dataset built on MultiVENT 2.0 (event-centric videos in various languages paired with queries) with modality-targeted queries. Next, we propose a modality-aware loss that jointly trains according to a standard contrastive objective alongside an objective for learning correct modality usage. On the test sets of MultiVENT 2.0++ and MSRVTT, conventional aggregation strategies, such as averaging similarities for baseline retrievers, degrade performance by introducing noise from irrelevant modalities. In contrast, CLaMR consistently outperforms existing retrievers: on MultiVENT 2.0++, CLaMR improves nDCG@10 by 25.6 over the best single-modality retriever and by 35.4 over the best multi-modality retriever. We illustrate CLaMR's downstream utility on long-video QA, retrieving relevant frames and obtaining a 3.50% boost over LanguageBind on Video-MME and 1.42% over dense sampling on LongVideoBench.

cs.CV↗

arXiv2025

Learning Neural Controllers with Optimality and Stability Guarantees Using Input-Output Dissipativity

Deep learning methods have demonstrated significant potential for addressing complex nonlinear control problems. For real-world safety-critical tasks, however, it is crucial to provide formal stability guarantees for the designed controllers. In this paper, we propose a new framework for designing neural controllers that achieve both stability and optimality with respect to certain functions. Our key idea is to exploit the concept of input-output dissipativity of nonlinear systems by learning neural storage functions and supply rate functions. As a generalization of Lyapunov theory, dissipativity theory provides a natural connection to optimal control theory, offering both stability guarantees and meaningful optimality certificates. The neural controllers can be directly derived from the learned supply rate functions and guarantee closed-loop stability while inheriting optimality properties that can be shaped towards user-defined control objectives. Extensive numerical experiments demonstrate the effectiveness of our approach.

eess.SY↗

arXiv2025

AlphaOne: Reasoning Models Thinking Slow and Fast at Test Time

This paper presents AlphaOne ($α$1), a universal framework for modulating reasoning progress in large reasoning models (LRMs) at test time. $α$1 first introduces $α$ moment, which represents the scaled thinking phase with a universal parameter $α$. Within this scaled pre-$α$ moment phase, it dynamically schedules slow thinking transitions by modeling the insertion of reasoning transition tokens as a Bernoulli stochastic process. After the $α$ moment, $α$1 deterministically terminates slow thinking with the end-of-thinking token, thereby fostering fast reasoning and efficient answer generation. This approach unifies and generalizes existing monotonic scaling methods by enabling flexible and dense slow-to-fast reasoning modulation. Extensive empirical studies on various challenging benchmarks across mathematical, coding, and scientific domains demonstrate $α$1's superior reasoning capability and efficiency. Project page: https://alphaone-project.github.io/

cs.CL↗

arXiv2025

WildDoc: How Far Are We from Achieving Comprehensive and Robust Document Understanding in the Wild?

The rapid advancements in Multimodal Large Language Models (MLLMs) have significantly enhanced capabilities in Document Understanding. However, prevailing benchmarks like DocVQA and ChartQA predominantly comprise \textit{scanned or digital} documents, inadequately reflecting the intricate challenges posed by diverse real-world scenarios, such as variable illumination and physical distortions. This paper introduces WildDoc, the inaugural benchmark designed specifically for assessing document understanding in natural environments. WildDoc incorporates a diverse set of manually captured document images reflecting real-world conditions and leverages document sources from established benchmarks to facilitate comprehensive comparisons with digital or scanned documents. Further, to rigorously evaluate model robustness, each document is captured four times under different conditions. Evaluations of state-of-the-art MLLMs on WildDoc expose substantial performance declines and underscore the models' inadequate robustness compared to traditional benchmarks, highlighting the unique challenges posed by real-world document understanding. Our project homepage is available at https://bytedance.github.io/WildDoc.

cs.CV↗

arXiv2025

Neural Canonical Transformations for Quantum Anharmonic Solids of Lithium

Lithium is a typical quantum solid, characterized by cubic structures at ambient pressure. As the pressure increases, it forms more complex structures and undergoes a metal-to-semiconductor transformation, complicating theoretical and experimental analyses. We employ the neural canonical transformation approach, an \textit{ab initio} variational method based on probabilistic generative models, to investigate the quantum anharmonic effects in lithium solids at finite temperatures. This approach combines a normalizing flow for phonon excited-state wave functions with a probabilistic model for the occupation of energy levels, optimized jointly to minimize the free energy. Our results indicate that quantum anharmonicity lowers the \textit{bcc}-\textit{fcc} transition temperature compared to classical molecular dynamics predictions. At high pressures, the predicted fractional coordinates of lithium atoms in the \textit{cI16} structure show good quantitative agreement with experimental observations. Finally, contrary to previous beliefs, we find that the poor metallic \textit{oC88} structure is stabilized by the potential energy surface obtained via high-accuracy electronic structure calculations, rather than thermal or quantum nuclear effects.

cond-mat.mtrl-sci↗

arXiv2025

An Empirical Study of LLM Reasoning Ability Under Strict Output Length Constraint

Recent work has demonstrated the remarkable potential of Large Language Models (LLMs) in test-time scaling. By making models think before answering, they are able to achieve much higher accuracy with extra inference computation. However, in many real-world scenarios, models are used under time constraints, where an answer should be given within a certain output length. It is unclear whether and how the reasoning ability of different LLMs remain effective under strict constraints. We take a first look at this problem by conducting an in-depth empirical study. Specifically, we test 30 LLMs on common reasoning datasets under a wide range of output length budgets, and we analyze the correlation between the inference accuracy and various properties including model type, model size, prompt style, etc. We also consider the mappings between token budgets and actual on-device latency budgets. The results have demonstrated several interesting findings regarding the budget-aware LLM reasoning ability that differ from the unconstrained situation, e.g. the optimal choices of either model size or prompt style change under different budgets. These findings offer timely evaluation to this area and practical guidance for users to deploy LLMs under real-world latency constraints.

cs.AI↗

arXiv2025

Prolonged Reasoning Is Not All You Need: Certainty-Based Adaptive Routing for Efficient LLM/MLLM Reasoning

Recent advancements in reasoning have significantly enhanced the capabilities of Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs) across diverse tasks. However, excessive reliance on chain-of-thought (CoT) reasoning can impair model performance and brings unnecessarily lengthened outputs, reducing efficiency. Our work reveals that prolonged reasoning does not universally improve accuracy and even degrade performance on simpler tasks. To address this, we propose Certainty-based Adaptive Reasoning (CAR), a novel framework that dynamically switches between short answers and long-form reasoning based on the model perplexity. CAR first generates a short answer and evaluates its perplexity, triggering reasoning only when the model exhibits low confidence (i.e., high perplexity). Experiments across diverse multimodal VQA/KIE benchmarks and text reasoning datasets show that CAR outperforms both short-answer and long-form reasoning approaches, striking an optimal balance between accuracy and efficiency.

cs.CL↗

arXiv2025

RT-APNN for Solving Gray Radiative Transfer Equations

The Gray Radiative Transfer Equations (GRTEs) are high-dimensional, multiscale problems that pose significant computational challenges for traditional numerical methods. Current deep learning approaches, including Physics-Informed Neural Networks (PINNs) and Asymptotically Preserving Neural Networks (APNNs), are largely restricted to low-dimensional or linear GRTEs. To address these challenges, we propose the Radiative Transfer Asymptotically Preserving Neural Network (RT-APNN), an innovative framework extending APNNs. RT-APNN integrates multiple neural networks into a cohesive architecture, reducing training time while ensuring high solution accuracy. Advanced techniques such as pre-training and Markov Chain Monte Carlo (MCMC) adaptive sampling are employed to tackle the complexities of long-term simulations and intricate boundary conditions. RT-APNN is the first deep learning method to successfully simulate the Marshak wave problem. Numerical experiments demonstrate its superiority over existing methods, including APNNs and MD-APNNs, in both accuracy and computational efficiency. Furthermore, RT-APNN excels at solving high-dimensional, nonlinear problems, underscoring its potential for diverse applications in science and engineering.

physics.comp-ph↗

arXiv2025

Detection for Intermediate-mass Binary Black Holes in Population III Star Clusters with TianQin

Context. Population III star clusters are predicted to form in unenriched dark matter halos. Direct N-body simulation of Pop III clusters implies the possible formation and merger of intermediate-mass binary black holes (IMBBHs). The gravitational wave signals could be detected by space-borne gravitational wave detectors like TianQin. Aims. This study evaluates the potential of TianQin in detecting IMBBHs from Pop III star clusters, focusing on key factors such as the detection horizon, detection number, and Fisher uncertainty. Methods. A Monte Carlo simulation is employed to derive IMBBH sources, utilizing the IMRPhenomD waveform model to simulate catalogs of observable IMBBH mergers. The mass and redshift distributions are derived from direct N-body simulations of IMBBHs in Population III star clusters. Detection numbers are determined by calculating the signal-to-noise ratios (SNR) of the simulated signals and applying thresholds for detection. Fisher uncertainty is obtained through Fisher information matrix analysis. Results. The findings suggest that TianQin could achieve detection numbers within 5 years ranging from 1 in the most pessimistic scenario to 253 in the most optimistic scenario. Furthermore, TianQin can precisely constrain the IMBBH mass with a relative uncertainty of $10^{-6}$, coalescence time $t_c$ within 1 second, and sky location $\barΩ_S$ within 1 $\rm{deg}^2$. However, the luminosity distance $D_L$ and inclination angle $ι$ exhibit substantial degeneracies, limiting their precise estimation.

astro-ph.GA↗

arXiv2025

A Bounding Box is Worth One Token: Interleaving Layout and Text in a Large Language Model for Document Understanding

Recently, many studies have demonstrated that exclusively incorporating OCR-derived text and spatial layouts with large language models (LLMs) can be highly effective for document understanding tasks. However, existing methods that integrate spatial layouts with text have limitations, such as producing overly long text sequences or failing to fully leverage the autoregressive traits of LLMs. In this work, we introduce Interleaving Layout and Text in a Large Language Model (LayTextLLM)} for document understanding. LayTextLLM projects each bounding box to a single embedding and interleaves it with text, efficiently avoiding long sequence issues while leveraging autoregressive traits of LLMs. LayTextLLM not only streamlines the interaction of layout and textual data but also shows enhanced performance in KIE and VQA. Comprehensive benchmark evaluations reveal significant improvements of LayTextLLM, with a 15.2% increase on KIE tasks and 10.7% on VQA tasks compared to previous SOTA OCR-based LLMs. All resources are available at https://github.com/LayTextLLM/LayTextLLM.

cs.CL↗

arXiv2025

A Deep Learning Potential for Accurate Shock Response Simulations in Tin

Tin (Sn) plays a crucial role in studying the dynamic mechanical responses of ductile metals under shock loading. Atomistic simulations serves to unveil the nano-scale mechanisms for critical behaviors of dynamic responses. However, existing empirical potentials for Sn often lack sufficient accuracy when applied in such simulation. Particularly, the solid-solid phase transition behavior of Sn poses significant challenges to the accuracy of interatomic potentials. To address these challenges, this study introduces a machine-learning potential model for Sn, specifically optimized for shock-response simulations. The model is trained using a dataset constructed through a concurrent learning framework and is designed for molecular simulations across thermodynamic conditions ranging from 0 to 100 GPa and 0 to 5000 K, encompassing both solid and liquid phases as well as structures with free surfaces. It accurately reproduces density functional theory (DFT)-derived basic properties, experimental melting curves, solid-solid phase boundaries, and shock Hugoniot results. This demonstrates the model's potential to bridge ab initio precision with large-scale dynamic simulations of Sn.

cond-mat.mtrl-sci↗

arXiv2025

Demonstration of low-overhead quantum error correction codes

Quantum computers hold the potential to surpass classical computers in solving complex computational problems. However, the fragility of quantum information and the error-prone nature of quantum operations make building large-scale, fault-tolerant quantum computers a prominent challenge. To combat errors, pioneering experiments have demonstrated a variety of quantum error correction codes. Yet, most of these codes suffer from low encoding efficiency, and their scalability is hindered by prohibitively high resource overheads. Here, we report the demonstration of two low-overhead quantum low-density parity-check (qLDPC) codes, a distance-4 bivariate bicycle code and a distance-3 qLDPC code, on our latest superconducting processor, Kunlun, featuring 32 long-range-coupled transmon qubits. Utilizing a two-dimensional architecture with overlapping long-range couplers, we demonstrate simultaneous measurements of all nonlocal weight-6 stabilizers via the periodic execution of an efficient syndrome extraction circuit. We achieve a logical error rate per logical qubit per cycle of $(8.91 \pm 0.17)\%$ for the distance-4 bivariate bicycle code with four logical qubits and $(7.77 \pm 0.12)\%$ for the distance-3 qLDPC code with six logical qubits. Our results establish the feasibility of implementing various qLDPC codes with long-range coupled superconducting processors, marking a crucial step towards large-scale low-overhead quantum error correction.

quant-ph↗

arXiv2025

Synthesis of safety certificates for discrete-time uncertain systems via convex optimization

We study the problem of co-designing control barrier functions and linear state feedback controllers for discrete-time linear systems affected by additive disturbances. For disturbances of bounded magnitude, we provide a semi-definite program whose feasibility implies the existence of a control law and a certificate ensuring safety in the infinite horizon with respect to the worst-case disturbance realization in the uncertainty set. For disturbances with unbounded support, we rely on martingale theory to derive a second semi-definite program whose feasibility provides probabilistic safety guarantees holding joint-in-time over a finite time horizon. We examine several extensions, including (i) encoding of different types of input constraints, (ii) robustification against distributional ambiguity around the true distribution, (iii) design of safety filters, and (iv) extension to general safety specifications such as obstacle avoidance.

math.OC↗

arXiv2025

Clickbait Detection via Large Language Models

Clickbait, which aims to induce users with some surprising and even thrilling headlines for increasing click-through rates, permeates almost all online content publishers, such as news portals and social media. Recently, Large Language Models (LLMs) have emerged as a powerful instrument and achieved tremendous success in a series of NLP downstream tasks. However, it is not yet known whether LLMs can be served as a high-quality clickbait detection system. In this paper, we analyze the performance of LLMs in the few-shot and zero-shot scenarios on several English and Chinese benchmark datasets. Experimental results show that LLMs cannot achieve the best results compared to the state-of-the-art deep and fine-tuning PLMs methods. Different from human intuition, the experiments demonstrated that LLMs cannot make satisfied clickbait detection just by the headlines.

cs.CL↗

arXiv2025

PSSketch: Finding Persistent and Sparse Flow with High Accuracy and Efficiency

Finding persistent sparse (PS) flow is critical to early warning of many threats. Previous works have predominantly focused on either heavy or persistent flows, with limited attention given to PS flows. Although some recent studies pay attention to PS flows, they struggle to establish an objective criterion due to insufficient data-driven observations, resulting in reduced accuracy. In this paper, we define a new criterion "anomaly boundary" to distinguish PS flows from regular flows. Specifically, a flow whose persistence exceeds a threshold will be protected, while a protected flow with a density lower than a threshold is reported as a PS flow. We then introduce PSSketch, a high-precision layered sketch to find PS flows. PSSketch employs variable-length bitwise counters, where the first layer tracks the frequency and persistence of all flows, and the second layer protects potential PS flows and records overflow counts from the first layer. Some optimizations have also been implemented to reduce memory consumption further and improve accuracy. The experiments show that PSSketch reduces memory consumption by an order of magnitude compared to the strawman solution combined with existing work. Compared with SOTA solutions for finding PS flows, it outperforms up to 2.94x in F1 score and reduces ARE by 1-2 orders of magnitude. Meanwhile, PSSketch achieves a higher throughput than these solutions.

cs.DS↗