arXiv subjects

Wenyuan Xue

Publications and source records attributed to Wenyuan Xue.

6 recordsLinked to original sources

arXiv2026

Generalized Forcing Method: Generation of Diverse Data for Training Linear Transport PDE Closure Models

Data-driven closure modeling for transport partial differential equations requires training data that are accurate, affordable, diverse, and directly tailored to the target closure fields. We develop the Generalized Forcing Method (GFM), a data-generation framework for training linear transport closure models. GFM generates such data by running simulations with a zero initial condition and an extra body force that is constructed compatibly with the reduced dynamics. This framework leads to implicit GFM (iGFM), which prescribes resolved trajectories, and explicit GFM (eGFM), which constructs a basis of admissible forcings. We apply eGFM to three linear transport closure problems: homogeneous shear flows, spatially inhomogeneous flows, and homogeneous shear flows with random coefficients. The results show that eGFM can identify accurate and stable reduced models when the reduced variables and model form are consistent with the underlying closure relation.

physics.flu-dyn↗

arXiv2026

A variable-coefficient model for decay of isotropic turbulence capturing effects of finite cascade time and Reynolds number

We study isotropic turbulence decay in the context of the k-epsilon model, which solves the dissipation and kinetic energy equations. In modeling the dissipation equation, the coefficient C_epsilon2, suggested by Hanjalic and Launder [Journal of Fluid Mechanics, 1972] [1], is related to the temporal decay power-law by n = 1/(C_epsilon2 -1 )) and is assumed to be a constant value. In this work, we perform high-fidelity numerical simulations to examine the mathematical terms responsible for the decay of isotropic turbulence, considering both scenarios of forced and decaying turbulence. Our data suggest that the instantaneous C_epsilon2 not only depends on the instantaneous Reynolds number but is also sensitive to the history of energy injection in turbulence. We attribute these observations to the finite time required for the cascade from energetic to dissipative scales. Considering data from both decaying and growing forced turbulence, we develop an evolution equation for C_epsilon2 with Reynolds-dependent coefficients. We demonstrate that this model accurately captures the time evolution of dissipation and kinetic energy over a wide range of Reynolds numbers under a wide range of forced and decay scenarios.

physics.flu-dyn↗

arXiv2024

A Survey on Hallucination in Large Vision-Language Models

Recent development of Large Vision-Language Models (LVLMs) has attracted growing attention within the AI landscape for its practical implementation potential. However, ``hallucination'', or more specifically, the misalignment between factual visual content and corresponding textual generation, poses a significant challenge of utilizing LVLMs. In this comprehensive survey, we dissect LVLM-related hallucinations in an attempt to establish an overview and facilitate future mitigation. Our scrutiny starts with a clarification of the concept of hallucinations in LVLMs, presenting a variety of hallucination symptoms and highlighting the unique challenges inherent in LVLM hallucinations. Subsequently, we outline the benchmarks and methodologies tailored specifically for evaluating hallucinations unique to LVLMs. Additionally, we delve into an investigation of the root causes of these hallucinations, encompassing insights from the training data and model components. We also critically review existing methods for mitigating hallucinations. The open questions and future directions pertaining to hallucinations within LVLMs are discussed to conclude this survey.

cs.CV↗

arXiv2024

Align before Adapt: Leveraging Entity-to-Region Alignments for Generalizable Video Action Recognition

Large-scale visual-language pre-trained models have achieved significant success in various video tasks. However, most existing methods follow an "adapt then align" paradigm, which adapts pre-trained image encoders to model video-level representations and utilizes one-hot or text embedding of the action labels for supervision. This paradigm overlooks the challenge of mapping from static images to complicated activity concepts. In this paper, we propose a novel "Align before Adapt" (ALT) paradigm. Prior to adapting to video representation learning, we exploit the entity-to-region alignments for each frame. The alignments are fulfilled by matching the region-aware image embeddings to an offline-constructed text corpus. With the aligned entities, we feed their text embeddings to a transformer-based video adapter as the queries, which can help extract the semantics of the most important entities from a video to a vector. This paradigm reuses the visual-language alignment of VLP during adaptation and tries to explain an action by the underlying entities. This helps understand actions by bridging the gap with complex activity semantics, particularly when facing unfamiliar or unseen categories. ALT demonstrates competitive performance while maintaining remarkably low computational costs. In fully supervised experiments, it achieves 88.1% top-1 accuracy on Kinetics-400 with only 4947 GFLOPs. Moreover, ALT outperforms the previous state-of-the-art methods in both zero-shot and few-shot experiments, emphasizing its superior generalizability across various learning scenarios.

cs.CV↗

arXiv2023

ChartDETR: A Multi-shape Detection Network for Visual Chart Recognition

Visual chart recognition systems are gaining increasing attention due to the growing demand for automatically identifying table headers and values from chart images. Current methods rely on keypoint detection to estimate data element shapes in charts but suffer from grouping errors in post-processing. To address this issue, we propose ChartDETR, a transformer-based multi-shape detector that localizes keypoints at the corners of regular shapes to reconstruct multiple data elements in a single chart image. Our method predicts all data element shapes at once by introducing query groups in set prediction, eliminating the need for further postprocessing. This property allows ChartDETR to serve as a unified framework capable of representing various chart types without altering the network architecture, effectively detecting data elements of diverse shapes. We evaluated ChartDETR on three datasets, achieving competitive results across all chart types without any additional enhancements. For example, ChartDETR achieved an F1 score of 0.98 on Adobe Synthetic, significantly outperforming the previous best model with a 0.71 F1 score. Additionally, we obtained a new state-of-the-art result of 0.97 on ExcelChart400k. The code will be made publicly available.

cs.CV↗

arXiv2021

TGRNet: A Table Graph Reconstruction Network for Table Structure Recognition

A table arranging data in rows and columns is a very effective data structure, which has been widely used in business and scientific research. Considering large-scale tabular data in online and offline documents, automatic table recognition has attracted increasing attention from the document analysis community. Though human can easily understand the structure of tables, it remains a challenge for machines to understand that, especially due to a variety of different table layouts and styles. Existing methods usually model a table as either the markup sequence or the adjacency matrix between different table cells, failing to address the importance of the logical location of table cells, e.g., a cell is located in the first row and the second column of the table. In this paper, we reformulate the problem of table structure recognition as the table graph reconstruction, and propose an end-to-end trainable table graph reconstruction network (TGRNet) for table structure recognition. Specifically, the proposed method has two main branches, a cell detection branch and a cell logical location branch, to jointly predict the spatial location and the logical location of different cells. Experimental results on three popular table recognition datasets and a new dataset with table graph annotations (TableGraph-350K) demonstrate the effectiveness of the proposed TGRNet for table structure recognition. Code and annotations will be made publicly available.

cs.CV↗