arXiv subjects

Xiaohan Wang

Publications and source records attributed to Xiaohan Wang.

At least 91 records · Page 5Linked to original sources

arXiv2025

Temporal Preference Optimization for Long-Form Video Understanding

Despite significant advancements in video large multimodal models (video-LMMs), achieving effective temporal grounding in long-form videos remains a challenge for existing models. To address this limitation, we propose Temporal Preference Optimization (TPO), a novel post-training framework designed to enhance the temporal grounding capabilities of video-LMMs through preference learning. TPO adopts a self-training approach that enables models to differentiate between well-grounded and less accurate temporal responses by leveraging curated preference datasets at two granularities: localized temporal grounding, which focuses on specific video segments, and comprehensive temporal grounding, which captures extended temporal dependencies across entire video sequences. By optimizing on these preference datasets, TPO significantly enhances temporal understanding while reducing reliance on manually annotated data. Extensive experiments on three long-form video understanding benchmarks--LongVideoBench, MLVU, and Video-MME--demonstrate the effectiveness of TPO across two state-of-the-art video-LMMs. Notably, LLaVA-Video-TPO establishes itself as the leading 7B model on the Video-MME benchmark, underscoring the potential of TPO as a scalable and efficient solution for advancing temporal reasoning in long-form video understanding. Project page: https://ruili33.github.io/tpo_website.

cs.CV↗

arXiv2025

RLFactory: A Plug-and-Play Reinforcement Learning Post-Training Framework for LLM Multi-Turn Tool-Use

Large language models excel at basic reasoning but struggle with tasks that require interaction with external tools. We present RLFactory, a plug-and-play reinforcement learning post-training framework for multi-round tool use. RLFactory tackles (i) tool-call stability and adaptability amid tool heterogeneity and interface issues via an asyncio-based asynchronous caller and a decoupled tool/training architecture, and (ii) diverse evaluation needs via a reward layer supporting rule-based, model-judgment, and tool-verification signals. It reconstructs the MDP by introducing observation markers from tool feedback, closing the loop among model, tools, and environment, and implements a generate-parse-invoke-update workflow for dynamic policy optimization. On Search-R1 with Qwen3-4B, RLFactory achieves a 0.486 test score on the Natural Questions (NQ) dataset, surpassing larger models trained with similar techniques (e.g., Qwen2.5-7B-Instruct-GRPO at 0.473), and increases training throughput by 6.8x. RLFactory provides a low-barrier, highly adaptable framework for strengthening multi-round tool use of LLMs in real-world scenarios. Code: https://github.com/Simple-Efficient/RL-Factory.

cs.LG↗

arXiv2025

Transfer Learning for Classification under Decision Rule Drift with Application to Optimal Individualized Treatment Rule Estimation

In this paper, we extend the transfer learning classification framework from regression function-based methods to decision rules. We propose a novel methodology for modeling posterior drift through Bayes decision rules. By exploiting the geometric transformation of the Bayes decision boundary, our method reformulates the problem as a low-dimensional empirical risk minimization problem. Under mild regularity conditions, we establish the consistency of our estimators and derive the risk bounds. Moreover, we illustrate the broad applicability of our method by adapting it to the estimation of optimal individualized treatment rules. Extensive simulation studies and analyses of real-world data further demonstrate both superior performance and robustness of our approach.

stat.ML↗

arXiv2025

Seeing the Many: Exploring Parameter Distributions Conditioned on Features in Surrogates

Recently, neural surrogate models have emerged as a compelling alternative to traditional simulation workflows. This is accomplished by modeling the underlying function of scientific simulations, removing the need to run expensive simulations. Beyond just mapping from input parameter to output, surrogates have also been shown useful for inverse problems: output to input parameters. Inverse problems can be understood as search, where we aim to find parameters whose surrogate outputs contain a specified feature. Yet finding these parameters can be costly, especially for high-dimensional parameter spaces. Thus, existing surrogate-based solutions primarily focus on finding a small set of matching parameters, in the process overlooking the broader picture of plausible parameters. Our work aims to model and visualize the distribution of possible input parameters that produce a given output feature. To achieve this goal, we aim to address two challenges: (1) the approximation error inherent in the surrogate model and (2) forming the parameter distribution in an interactive manner. We model error via density estimation, reporting high density only if a given parameter configuration is close to training parameters, measured both over the input and output space. Our density estimate is used to form a prior belief on parameters, and when combined with a likelihood on features, gives us an efficient way to sample plausible parameter configurations that generate a target output feature. We demonstrate the usability of our solution through a visualization interface by performing feature-driven parameter analysis over the input parameter space of three simulation datasets. Source code is available at https://github.com/matthewberger/seeing-the-many

cs.LG↗

arXiv2025

Feather the Throttle: Revisiting Visual Token Pruning for Vision-Language Model Acceleration

Recent works on accelerating Vision-Language Models achieve strong performance across a variety of vision-language tasks despite highly compressing visual information. In this work, we examine the popular acceleration approach of early pruning of visual tokens inside the language model. Surprisingly, we find that while strong performance is maintained across many tasks, it exhibits drastically different behavior for a subset of vision-centric tasks such as localization. Upon further investigation, we uncover a core issue with the acceleration approach where most tokens towards the top of the image are pruned away. Yet, on many benchmarks aiming to evaluate vision-centric capabilities, strong performance persists with the flawed pruning strategy, highlighting these benchmarks' limited ability to assess fine-grained visual capabilities. Based on these findings, we propose FEATHER (Fast and Effective Acceleration wiTH Ensemble cRiteria), a straightforward approach that resolves the discovered early-layer pruning issue and further enhances the preservation of relevant tokens via multistage pruning with early uniform sampling to ensure broad image coverage. With comparable computational savings, we find that FEATHER achieves more than 5x performance improvement on the vision-centric localization benchmarks compared to the original acceleration approach.

cs.CV↗

arXiv2025

Closing the Modality Gap for Mixed Modality Search

Mixed modality search -- retrieving information across a heterogeneous corpus composed of images, texts, and multimodal documents -- is an important yet underexplored real-world application. In this work, we investigate how contrastive vision-language models, such as CLIP, perform on the mixed modality search task. Our analysis reveals a critical limitation: these models exhibit a pronounced modality gap in the embedding space, where image and text embeddings form distinct clusters, leading to intra-modal ranking bias and inter-modal fusion failure. To address this issue, we propose GR-CLIP, a lightweight post-hoc calibration method that removes the modality gap in CLIP's embedding space. Evaluated on MixBench -- the first benchmark specifically designed for mixed modality search -- GR-CLIP improves NDCG@10 by up to 26 percentage points over CLIP, surpasses recent vision-language generative embedding models by 4 percentage points, while using 75x less compute.

cs.CV↗

arXiv2025

Fabrication of airbridges with gradient exposure

In superconducting quantum circuits, airbridges are critical for eliminating parasitic slotline modes of coplanar waveguide circuits and reducing crosstalks between direct current magnetic flux biases. Here, we present a technique for fabricating superconducting airbridges. With this technique, a single layer of photoresist is employed, and the gradient exposure process is used to define the profile of airbridges. In order to properly obtain the bridge profile, we design exposure dosage based on residual photoresist thickness and laser power calibrations. Compared with other airbridge fabrication techniques, the gradient exposure fabrication technique provides the ability to produce lossless superconducting airbridges with flexible size and, thus, is more suitable for large-scale superconducting quantum circuits. Furthermore, this method reduces the complexity of the fabrication process and provides a high fabrication yield.

quant-ph↗

arXiv2025

Hidden Representation Clustering with Multi-Task Representation Learning towards Robust Online Budget Allocation

Marketing optimization, commonly formulated as an online budget allocation problem, has emerged as a pivotal factor in driving user growth. Most existing research addresses this problem by following the principle of 'first predict then optimize' for each individual, which presents challenges related to large-scale counterfactual prediction and solving complexity trade-offs. Note that the practical data quality is uncontrollable, and the solving scale tends to be tens of millions. Therefore, the existing approaches make the robust budget allocation non-trivial, especially in industrial scenarios with considerable data noise. To this end, this paper proposes a novel approach that solves the problem from the cluster perspective. Specifically, we propose a multi-task representation network to learn the inherent attributes of individuals and project the original features into high-dimension hidden representations through the first two layers of the trained network. Then, we divide these hidden representations into $K$ groups through partitioning-based clustering, thus reformulating the problem as an integer stochastic programming problem under different total budgets. Finally, we distill the representation module and clustering model into a multi-category model to facilitate online deployment. Offline experiments validate the effectiveness and superiority of our approach compared to six state-of-the-art marketing optimization algorithms. Online A/B tests on the Meituan platform indicate that the approach outperforms the online algorithm by 0.53% and 0.65%, considering order volume (OV) and gross merchandise volume (GMV), respectively.

cs.LG↗

arXiv2025

Topology Guidance: Controlling the Outputs of Generative Models via Vector Field Topology

For domains that involve numerical simulation, it can be computationally expensive to run an ensemble of simulations spanning a parameter space of interest to a user. To this end, an attractive surrogate for simulation is the generative modeling of fields produced by an ensemble, allowing one to synthesize fields in a computationally cheap, yet accurate, manner. However, for the purposes of visual analysis, a limitation of generative models is their lack of control, as it is unclear what one should expect when sampling a field from a model. In this paper we study how to make generative models of fields more controllable, so that users can specify features of interest, in particular topological features, that they wish to see in the output. We propose topology guidance, a method for guiding the sampling process of a generative model, specifically a diffusion model, such that a topological description specified as input is satisfied in the generated output. Central to our method, we couple a coordinate-based neural network used to represent fields, with a diffusion model used for generation. We show how to use topologically-relevant signals provided by the coordinate-based network to help guide the denoising process of a diffusion model. This enables us to faithfully represent a user's specified topology, while ensuring that the output field remains within the generative data distribution. Specifically, we study 2D vector field topology, evaluating our method over an ensemble of fluid flows, where we show that generated vector fields faithfully adhere to the location, and type, of critical points over the spatial domain. We further show the benefits of our method in aiding the comparison of ensembles, allowing one to explore commonalities and differences in distributions along prescribed topological features.

cs.LG↗

arXiv2025

Event-aware analysis of cross-city visitor flows using large language models and social media data

Public events, such as music concerts and fireworks displays, can cause irregular surges in cross-city travel demand, leading to potential overcrowding, travel delays, and public safety concerns. To better anticipate and accommodate such demand surges, it is essential to estimate cross-city visitor flows with awareness of public events. Although prior studies typically focused on the effects of a single mega event or disruptions around a single venue, this study introduces a generalizable framework to analyze visitor flows under diverse and concurrent events. We propose to leverage large language models (LLMs) to extract event features from multi-source online information and massive user-generated content on social media platforms. Specifically, social media popularity metrics are designed to capture the effects of online promotion and word-of-mouth in attracting visitors. An event-aware machine learning model is then adopted to uncover the specific impacts of different event features and ultimately predict visitor flows for upcoming events. Using Hong Kong as a case study, the framework is applied to predict daily flows of mainland Chinese visitors arriving at the city, achieving a testing R-squared of over 85%. We further investigate the heterogeneous event impacts on visitor numbers across different event types and major travel modes. Both promotional popularity and word-of-mouth popularity are found to be associated with increased visitor flows, but the specific effects vary by the event type. This association is more pronounced among visitors arriving by metro and high-speed rail, while it has less effect on air travelers. The findings can facilitate coordinated measures across government agencies and guide specialized transport policies, such as shuttle transit services to event venues, and comprehensive on-site traffic management strategies.

cs.SI↗

arXiv2025

Innovative Thinking, Infinite Humor: Humor Research of Large Language Models through Structured Thought Leaps

Humor is previously regarded as a gift exclusive to humans for the following reasons. Humor is a culturally nuanced aspect of human language, presenting challenges for its understanding and generation. Humor generation necessitates a multi-hop reasoning process, with each hop founded on proper rationales. Although many studies, such as those related to GPT-o1, focus on logical reasoning with reflection and correction, they still fall short in humor generation. Due to the sparsity of the knowledge graph in creative thinking, it is arduous to achieve multi-hop reasoning. Consequently, in this paper, we propose a more robust framework for addressing the humor reasoning task, named LoL. LoL aims to inject external information to mitigate the sparsity of the knowledge graph, thereby enabling multi-hop reasoning. In the first stage of LoL, we put forward an automatic instruction-evolution method to incorporate the deeper and broader thinking processes underlying humor. Judgment-oriented instructions are devised to enhance the model's judgment capability, dynamically supplementing and updating the sparse knowledge graph. Subsequently, through reinforcement learning, the reasoning logic for each online-generated response is extracted using GPT-4o. In this process, external knowledge is re-introduced to aid the model in logical reasoning and the learning of human preferences. Finally, experimental results indicate that the combination of these two processes can enhance both the model's judgment ability and its generative capacity. These findings deepen our comprehension of the creative capabilities of large language models (LLMs) and offer approaches to boost LLMs' creative abilities for cross-domain innovative applications.

cs.AI↗

arXiv2025

Automated Generation of Challenging Multiple-Choice Questions for Vision Language Model Evaluation

The rapid development of vision language models (VLMs) demands rigorous and reliable evaluation. However, current visual question answering (VQA) benchmarks often depend on open-ended questions, making accurate evaluation difficult due to the variability in natural language responses. To address this, we introduce AutoConverter, an agentic framework that automatically converts these open-ended questions into multiple-choice format, enabling objective evaluation while reducing the costly multiple-choice question creation process. Our experiments demonstrate that AutoConverter can generate correct and challenging multiple-choice questions, with VLMs demonstrating consistently similar or lower accuracy on these questions compared to human-created ones. Using AutoConverter, we construct VMCBench, a benchmark created by transforming 20 existing VQA datasets into a unified multiple-choice format, totaling 9,018 questions. We comprehensively evaluate 33 state-of-the-art VLMs on VMCBench, setting a new standard for scalable, consistent, and reproducible VLM evaluation.

cs.CV↗

arXiv2025

BIOMEDICA: An Open Biomedical Image-Caption Archive, Dataset, and Vision-Language Models Derived from Scientific Literature

The development of vision-language models (VLMs) is driven by large-scale and diverse multimodal datasets. However, progress toward generalist biomedical VLMs is limited by the lack of annotated, publicly accessible datasets across biology and medicine. Existing efforts are restricted to narrow domains, missing the full diversity of biomedical knowledge encoded in scientific literature. To address this gap, we introduce BIOMEDICA, a scalable, open-source framework to extract, annotate, and serialize the entirety of the PubMed Central Open Access subset into an easy-to-use, publicly accessible dataset. Our framework produces a comprehensive archive with over 24 million unique image-text pairs from over 6 million articles. Metadata and expert-guided annotations are also provided. We demonstrate the utility and accessibility of our resource by releasing BMCA-CLIP, a suite of CLIP-style models continuously pre-trained on the BIOMEDICA dataset via streaming, eliminating the need to download 27 TB of data locally. On average, our models achieve state-of-the-art performance across 40 tasks - spanning pathology, radiology, ophthalmology, dermatology, surgery, molecular biology, parasitology, and cell biology - excelling in zero-shot classification with a 6.56% average improvement (as high as 29.8% and 17.5% in dermatology and ophthalmology, respectively), and stronger image-text retrieval, all while using 10x less compute. To foster reproducibility and collaboration, we release our codebase and dataset for the broader research community.

cs.CV↗

arXiv2025

Zero-shot Action Localization via the Confidence of Large Vision-Language Models

Precise action localization in untrimmed video is vital for fields such as professional sports and minimally invasive surgery, where the delineation of particular motions in recordings can dramatically enhance analysis. But in many cases, large scale datasets with video-label pairs for localization are unavailable, limiting the opportunity to fine-tune video-understanding models. Recent developments in large vision-language models (LVLM) address this need with impressive zero-shot capabilities in a variety of video understanding tasks. However, the adaptation of LVLMs, with their powerful visual question answering capabilities, to zero-shot localization in long-form video is still relatively unexplored. To this end, we introduce a true Zero-shot Action Localization method (ZEAL). Specifically, we leverage the built-in action knowledge of a large language model (LLM) to inflate actions into detailed descriptions of the archetypal start and end of the action. These descriptions serve as queries to LVLM for generating frame-level confidence scores which can be aggregated to produce localization outputs. The simplicity and flexibility of our method lends it amenable to more capable LVLMs as they are developed, and we demonstrate remarkable results in zero-shot action localization on a challenging benchmark, without any training. Our code is publicly available at $\href{https://github.com/josaklil-ai/zeal}{github.com/josaklil-ai/zeal}$.

cs.CV↗

arXiv2025

Video Action Differencing

How do two individuals differ when performing the same action? In this work, we introduce Video Action Differencing (VidDiff), the novel task of identifying subtle differences between videos of the same action, which has many applications, such as coaching and skill learning. To enable development on this new task, we first create VidDiffBench, a benchmark dataset containing 549 video pairs, with human annotations of 4,469 fine-grained action differences and 2,075 localization timestamps indicating where these differences occur. Our experiments demonstrate that VidDiffBench poses a significant challenge for state-of-the-art large multimodal models (LMMs), such as GPT-4o and Qwen2-VL. By analyzing failure cases of LMMs on VidDiffBench, we highlight two key challenges for this task: localizing relevant sub-actions over two videos and fine-grained frame comparison. To overcome these, we propose the VidDiff method, an agentic workflow that breaks the task into three stages: action difference proposal, keyframe localization, and frame differencing, each stage utilizing specialized foundation models. To encourage future research in this new task, we release the benchmark at https://huggingface.co/datasets/jmhb/VidDiffBench and code at http://jmhb0.github.io/viddiff.

cs.CV↗

arXiv2025

SurgiSAM2: Fine-tuning a foundational model for surgical video anatomy segmentation and detection

Background: We evaluate SAM 2 for surgical scene understanding by examining its semantic segmentation capabilities for organs/tissues both in zero-shot scenarios and after fine-tuning. Methods: We utilized five public datasets to evaluate and fine-tune SAM 2 for segmenting anatomical tissues in surgical videos/images. Fine-tuning was applied to the image encoder and mask decoder. We limited training subsets from 50 to 400 samples per class to better model real-world constraints with data acquisition. The impact of dataset size on fine-tuning performance was evaluated with weighted mean Dice coefficient (WMDC), and the results were also compared against previously reported state-of-the-art (SOTA) results. Results: SurgiSAM 2, a fine-tuned SAM 2 model, demonstrated significant improvements in segmentation performance, achieving a 17.9% relative WMDC gain compared to the baseline SAM 2. Increasing prompt points from 1 to 10 and training data scale from 50/class to 400/class enhanced performance; the best WMDC of 0.92 on the validation subset was achieved with 10 prompt points and 400 samples per class. On the test subset, this model outperformed prior SOTA methods in 24/30 (80%) of the classes with a WMDC of 0.91 using 10-point prompts. Notably, SurgiSAM 2 generalized effectively to unseen organ classes, achieving SOTA on 7/9 (77.8%) of them. Conclusion: SAM 2 achieves remarkable zero-shot and fine-tuned performance for surgical scene segmentation, surpassing prior SOTA models across several organ classes of diverse datasets. This suggests immense potential for enabling automated/semi-automated annotation pipelines, thereby decreasing the burden of annotations facilitating several surgical applications.

cs.CV↗

arXiv2025

On-Demand Pulse Shaping with Partially Coherent Pulses in Nonlinear Dispersive Media

In this Letter, we employ the complex screen method to investigate the dynamic evolution of partially coherent pulses with specified properties as they propagate through a nonlinear Kerr medium. Our results reveal that partially coherent pulses can retain stable pulse characteristics and exhibit enhanced robustness when the source coherence is reduced. Importantly, by adjusting the source pulse properties, the far-zone pulse properties can be customized on demand, even in highly nonlinear environments. These findings are of significant importance for applications such as pulse shaping, free-space optical communication, information encryption etc. in nonlinear media. Notably, the results offer valuable insights for mitigating nonlinear effects in light beams within the spatial domain.

physics.optics↗

arXiv2024

LLMs for Knowledge Graph Construction and Reasoning: Recent Capabilities and Future Opportunities

This paper presents an exhaustive quantitative and qualitative evaluation of Large Language Models (LLMs) for Knowledge Graph (KG) construction and reasoning. We engage in experiments across eight diverse datasets, focusing on four representative tasks encompassing entity and relation extraction, event extraction, link prediction, and question-answering, thereby thoroughly exploring LLMs' performance in the domain of construction and inference. Empirically, our findings suggest that LLMs, represented by GPT-4, are more suited as inference assistants rather than few-shot information extractors. Specifically, while GPT-4 exhibits good performance in tasks related to KG construction, it excels further in reasoning tasks, surpassing fine-tuned models in certain cases. Moreover, our investigation extends to the potential generalization ability of LLMs for information extraction, leading to the proposition of a Virtual Knowledge Extraction task and the development of the corresponding VINE dataset. Based on these empirical findings, we further propose AutoKG, a multi-agent-based approach employing LLMs and external sources for KG construction and reasoning. We anticipate that this research can provide invaluable insights for future undertakings in the field of knowledge graphs. The code and datasets are in https://github.com/zjunlp/AutoKG.

cs.CL↗