arXiv subjects

Wei Lin

Publications and source records attributed to Wei Lin.

At least 145 records · Page 8Linked to original sources

arXiv2025

Exploring Modality Guidance to Enhance VFM-based Feature Fusion for UDA in 3D Semantic Segmentation

Vision Foundation Models (VFMs) have become a de facto choice for many downstream vision tasks, like image classification, image segmentation, and object localization. However, they can also provide significant utility for downstream 3D tasks that can leverage the cross-modal information (e.g., from paired image data). In our work, we further explore the utility of VFMs for adapting from a labeled source to unlabeled target data for the task of LiDAR-based 3D semantic segmentation. Our method consumes paired 2D-3D (image and point cloud) data and relies on the robust (cross-domain) features from a VFM to train a 3D backbone on a mix of labeled source and unlabeled target data. At the heart of our method lies a fusion network that is guided by both the image and point cloud streams, with their relative contributions adjusted based on the target domain. We extensively compare our proposed methodology with different state-of-the-art methods in several settings and achieve strong performance gains. For example, achieving an average improvement of 6.5 mIoU (over all tasks), when compared with the previous state-of-the-art.

cs.CV↗

arXiv2025

ViG3D-UNet: Volumetric Vascular Connectivity-Aware Segmentation via 3D Vision Graph Representation

Accurate vascular segmentation is essential for coronary visualization and the diagnosis of coronary heart disease. This task involves the extraction of sparse tree-like vascular branches from the volumetric space. However, existing methods have faced significant challenges due to discontinuous vascular segmentation and missing endpoints. To address this issue, a 3D vision graph neural network framework, named ViG3D-UNet, was introduced. This method integrates 3D graph representation and aggregation within a U-shaped architecture to facilitate continuous vascular segmentation. The ViG3D module captures volumetric vascular connectivity and topology, while the convolutional module extracts fine vascular details. These two branches are combined through channel attention to form the encoder feature. Subsequently, a paperclip-shaped offset decoder minimizes redundant computations in the sparse feature space and restores the feature map size to match the original input dimensions. To evaluate the effectiveness of the proposed approach for continuous vascular segmentation, evaluations were performed on two public datasets, ASOCA and ImageCAS. The segmentation results show that the ViG3D-UNet surpassed competing methods in maintaining vascular segmentation connectivity while achieving high segmentation accuracy. Our code will be available soon.

eess.IV↗

arXiv2025

Semiparametric Causal Discovery and Inference with Invalid Instruments

Learning causal relationships among a set of variables, as encoded by a directed acyclic graph, from observational data is complicated by the presence of unobserved confounders. Instrumental variables (IVs) are a popular remedy for this issue, but most existing methods either assume the validity of all IVs or postulate a specific form of relationship, such as a linear model, between the primary variables and the IVs. To overcome these limitations, we introduce a partially linear structural equation model for causal discovery and inference that accommodates potentially invalid IVs and allows for general dependence of the primary variables on the IVs. We establish identification under this semiparametric model by constructing surrogate valid IVs, and develop a finite-sample procedure for estimating the causal structures and effects. Theoretically, we show that our procedure consistently learns the causal structures, yields asymptotically normal estimates, and effectively controls the false discovery rate in edge recovery. Simulation studies demonstrate the superiority of our method over existing competitors, and an application to inferring gene regulatory networks in Alzheimer's disease illustrates its usefulness.

stat.ME↗

arXiv2025

Multi-Agent Deep Reinforcement Learning for Multiple Anesthetics Collaborative Control

Automated control of personalized multiple anesthetics in clinical Total Intravenous Anesthesia (TIVA) is crucial yet challenging. Current systems, including target-controlled infusion (TCI) and closed-loop systems, either rely on relatively static pharmacokinetic/pharmacodynamic (PK/PD) models or focus on single anesthetic control, limiting personalization and collaborative control. To address these issues, we propose a novel framework, Value Decomposition Multi-Agent Deep Reinforcement Learning (VD-MADRL). VD-MADRL optimizes the collaboration between two anesthetics propofol (Agent I) and remifentanil (Agent II). And It uses a Markov Game (MG) to identify optimal actions among heterogeneous agents. We employ various value function decomposition methods to resolve the credit allocation problem and enhance collaborative control. We also introduce a multivariate environment model based on random forest (RF) for anesthesia state simulation. Additionally, a data resampling and alignment technique ensures synchronized trajectory data. Our experiments on general and thoracic surgery datasets show that VD-MADRL performs better than human experience. It improves dose precision and keeps anesthesia states stable, providing great clinical value.

eess.SY↗

arXiv2025

DropMAE: Learning Representations via Masked Autoencoders with Spatial-Attention Dropout for Temporal Matching Tasks

This paper studies masked autoencoder (MAE) video pre-training for various temporal matching-based downstream tasks, i.e., object-level tracking tasks including video object tracking (VOT) and video object segmentation (VOS), self-supervised visual correspondence learning, dense tracking tasks including optical flow estimation and long-term point tracking, and 3D point cloud tracking. Specifically, our work explores to provide a general representation to boost the temporal matching ability in various downstream tracking tasks. To achieve this, we firstly find that a simple extension of MAE, which randomly masks out frame patches in videos and reconstruct the frame pixels, heavily relies on spatial cues while ignoring temporal relations for frame reconstruction, thus leading to sub-optimal temporal matching representations. To alleviate this, we propose DropMAE, which adaptively performs spatial-attention dropout in the frame reconstruction to facilitate temporal correspondence learning in videos. We obtain several important findings with DropMAE: 1) DropMAE is a strong and efficient temporal matching learner, which achieves better fine-tuning results on matching-based tasks than the ImageNet-based MAE with 2x faster pre-training speed. 2) DropMAE is effective for different tracking tasks, i.e., object-level matching tasks including VOT and VOS, dense tracking tasks including optical flow estimation and tracking any point (TAP), and even 3D tracking in the different modality of point cloud data. Since none exists, we build ViT-based trackers for different downstream tracking tasks, and our pre-trained DropMAE model can be directly loaded in these ViT-based trackers for fine-tuning without further modifications. Experiments on 6 downstream tracking tasks demonstrate the effectiveness of DropMAE as a general pre-trained representation for diverse tracking tasks.

cs.CV↗

arXiv2025

PerLA: Perceptive 3D Language Assistant

Enabling Large Language Models (LLMs) to understand the 3D physical world is an emerging yet challenging research direction. Current strategies for processing point clouds typically downsample the scene or divide it into smaller parts for separate analysis. However, both approaches risk losing key local details or global contextual information. In this paper, we introduce PerLA, a 3D language assistant designed to be more perceptive to both details and context, making visual representations more informative for the LLM. PerLA captures high-resolution (local) details in parallel from different point cloud areas and integrates them with (global) context obtained from a lower-resolution whole point cloud. We present a novel algorithm that preserves point cloud locality through the Hilbert curve and effectively aggregates local-to-global information via cross-attention and a graph neural network. Lastly, we introduce a novel loss for local representation consensus to promote training stability. PerLA outperforms state-of-the-art 3D language assistants, with gains of up to +1.34 CiDEr on ScanQA for question answering, and +4.22 on ScanRefer and +3.88 on Nr3D for dense captioning. https://gfmei.github.io/PerLA/

cs.CV↗

arXiv2025

Behavior Modeling Space Reconstruction for E-Commerce Search

Delivering superior search services is crucial for enhancing customer experience and driving revenue growth. Conventionally, search systems model user behaviors by combining user preference and query item relevance statically, often through a fixed logical 'and' relationship. This paper reexamines existing approaches through a unified lens using both causal graphs and Venn diagrams, uncovering two prevalent yet significant issues: entangled preference and relevance effects, and a collapsed modeling space. To surmount these challenges, our research introduces a novel framework, DRP, which enhances search accuracy through two components to reconstruct the behavior modeling space. Specifically, we implement preference editing to proactively remove the relevance effect from preference predictions, yielding untainted user preferences. Additionally, we employ adaptive fusion, which dynamically adjusts fusion criteria to align with the varying patterns of relevance and preference, facilitating more nuanced and tailored behavior predictions within the reconstructed modeling space. Empirical validation on two public datasets and a proprietary search dataset underscores the superiority of our proposed methodology, demonstrating marked improvements in performance over existing approaches.

cs.IR↗

arXiv2025

Teaching VLMs to Localize Specific Objects from In-context Examples

Vision-Language Models (VLMs) have shown remarkable capabilities across diverse visual tasks, including image recognition, video understanding, and Visual Question Answering (VQA) when explicitly trained for these tasks. Despite these advances, we find that present-day VLMs (including the proprietary GPT-4o) lack a fundamental cognitive ability: learning to localize specific objects in a scene by taking into account the context. In this work, we focus on the task of few-shot personalized localization, where a model is given a small set of annotated images (in-context examples) -- each with a category label and bounding box -- and is tasked with localizing the same object type in a query image. Personalized localization can be particularly important in cases of ambiguity of several related objects that can respond to a text or an object that is hard to describe with words. To provoke personalized localization abilities in models, we present a data-centric solution that fine-tunes them using carefully curated data from video object tracking datasets. By leveraging sequences of frames tracking the same object across multiple shots, we simulate instruction-tuning dialogues that promote context awareness. To reinforce this, we introduce a novel regularization technique that replaces object labels with pseudo-names, ensuring the model relies on visual context rather than prior knowledge. Our method significantly enhances the few-shot localization performance of recent VLMs ranging from 7B to 72B in size, without sacrificing generalization, as demonstrated on several benchmarks tailored towards evaluating personalized localization abilities. This work is the first to explore and benchmark personalized few-shot localization for VLMs -- exposing critical weaknesses in present-day VLMs, and laying a foundation for future research in context-driven vision-language applications.

cs.CV↗

arXiv2025

Exact first passage time distribution for nonlinear chemical reaction networks II: monomolecular reactions and a A + B - C type of second-order reaction with arbitrary initial conditions

In biochemical reaction networks, the first passage time (FPT) of a reaction quantifies the time it takes for the reaction to first occur, from the initial state. While the mean FPT historically served as a summary metric, a far more comprehensive characterization of the dynamics of the network is contained within the complete FPT distribution. The relatively uncommon theoretical treatments of the FPT distribution that have been given in the past have been confined to linear systems, with zero and first-order processes. Recently, we presented theoretically exact solutions for the FPT distribution, within nonlinear systems involving two-particle collisions, such as A+B - C. Although this research yielded invaluable results, it was based upon the assumption of initial conditions in the form of a Poisson distribution. This somewhat restricts its relevance to real-world biochemical systems, which frequently display intricate behaviour and initial conditions that are non-Poisson in nature. Our current study extends prior analyses to accommodate arbitrary initial conditions, thereby expanding the applicability of our theoretical framework and providing a more adaptable tool for capturing the dynamics of biochemical reaction networks.

q-bio.MN↗

arXiv2025

IterQR: An Iterative Framework for LLM-based Query Rewrite in e-Commercial Search System

The essence of modern e-Commercial search system lies in matching user's intent and available candidates depending on user's query, providing personalized and precise service. However, user's query may be incorrect due to ambiguous input and typo, leading to inaccurate search. These cases may be released by query rewrite: modify query to other representation or expansion. However, traditional query rewrite replies on static rewrite vocabulary, which is manually established meanwhile lacks interaction with both domain knowledge in e-Commercial system and common knowledge in the real world. In this paper, with the ability to generate text content of Large Language Models (LLMs), we provide an iterative framework to generate query rewrite. The framework incorporates a 3-stage procedure in each iteration: Rewrite Generation with domain knowledge by Retrieval-Augmented Generation (RAG) and query understanding by Chain-of-Thoughts (CoT); Online Signal Collection with automatic positive rewrite update; Post-training of LLM with multi task objective to generate new rewrites. Our work (named as IterQR) provides a comprehensive framework to generate \textbf{Q}uery \textbf{R}ewrite with both domain / real-world knowledge. It automatically update and self-correct the rewrites during \textbf{iter}ations. \method{} has been deployed in Meituan Delivery's search system (China's leading food delivery platform), providing service for users with significant improvement.

cs.IR↗

arXiv2025

Adversarial Transform Particle Filters

The particle filter (PF) and the ensemble Kalman filter (EnKF) are widely used for approximate inference in state-space models. From a Bayesian perspective, these algorithms represent the prior by an ensemble of particles and update it to the posterior with new observations over time. However, the PF often suffers from weight degeneracy in high-dimensional settings, whereas the EnKF relies on linear Gaussian assumptions that can introduce significant approximation errors. In this paper, we propose the Adversarial Transform Particle Filter (ATPF), a novel filtering framework that combines the strengths of the PF and the EnKF through adversarial learning. Specifically, importance sampling is used to ensure statistical consistency as in the PF, while adversarially learned transformations, such as neural networks, allow accurate posterior matching for nonlinear and non-Gaussian systems. In addition, we incorporate kernel methods to ease optimization and leverage regularization techniques based on optimal transport for better statistical properties and numerical stability. We provide theoretical guarantees, including generalization bounds for both the analysis and forecast steps of ATPF. Extensive experiments across various nonlinear and non-Gaussian scenarios demonstrate the effectiveness and practical advantages of our method.

stat.ME↗

arXiv2025

A 3D Multimodal Feature for Infrastructure Anomaly Detection

Ageing structures require periodic inspections to identify structural defects. Previous work has used geometric distortions to locate cracks in synthetic masonry bridge point clouds but has struggled to detect small cracks. To address this limitation, this study proposes a novel 3D multimodal feature, 3DMulti-FPFHI, that combines a customized Fast Point Feature Histogram (FPFH) with an intensity feature. This feature is integrated into the PatchCore anomaly detection algorithm and evaluated through statistical and parametric analyses. The method is further evaluated using point clouds of a real masonry arch bridge and a full-scale experimental model of a concrete tunnel. Results show that the 3D intensity feature enhances inspection quality by improving crack detection; it also enables the identification of water ingress which introduces intensity anomalies. The 3DMulti-FPFHI outperforms FPFH and a state-of-the-art multimodal anomaly detection method. The potential of the method to address diverse infrastructure anomaly detection scenarios is highlighted by the minimal requirements for data compared to learning-based methods. The code and related point cloud dataset are available at https://github.com/Jingyixiong/3D-Multi-FPFHI.

cs.CV↗

arXiv2025

Training an LLM-as-a-Judge Model: Pipeline, Insights, and Practical Lessons

The rapid advancement of large language models (LLMs) has opened new possibilities for their adoption as evaluative judges. This paper introduces Themis, a fine-tuned LLM judge that delivers sophisticated context-aware evaluations. We provide a comprehensive overview of the development pipeline for Themis, highlighting its scenario-dependent evaluation prompts and two novel methods for controlled instruction generation. These designs enable Themis to effectively distill evaluative skills from teacher models, while retaining flexibility for continuous development. We introduce two human-labeled benchmarks for meta-evaluation, demonstrating that Themis can achieve high alignment with human preferences in an economical manner. Additionally, we explore insights into the LLM-as-a-judge paradigm, revealing nuances in performance and the varied effects of reference answers. Notably, we observe that pure knowledge distillation from strong LLMs, though common, does not guarantee performance improvement through scaling. We propose a mitigation strategy based on instruction-following difficulty. Furthermore, we provide practical guidelines covering data balancing, prompt customization, multi-objective training, and metric aggregation. We aim for our method and findings, along with the fine-tuning data, benchmarks, and model checkpoints, to support future research and development in this area.

cs.CL↗

arXiv2025

Quantifying energy landscape of high-dimensional oscillatory systems by diffusion decomposition

High-dimensional networks producing oscillatory dynamics are ubiquitous in biological systems. Unravelling the mechanism of oscillatory dynamics in biological networks with stochastic perturbations becomes paramountly significant. Although the classical energy landscape theory provides a tool to study this problem in multistable systems and explain cellular functions, it remains challenging to quantify the landscape for high-dimensional oscillatory systems accurately. Here we propose an approach called the diffusion decomposition of Gaussian approximation (DDGA). We demonstrate the efficacy of the DDGA in quantifying the energy landscape of oscillatory systems and corresponding stochastic dynamics, in comparison with existing approaches. By further applying the DDGA to high-dimensional biological networks, we are able to uncover more intricate biological mechanisms efficiently, which deepens our understanding of cellular functions.

q-bio.QM↗

arXiv2025

Flexible delivery of broadband, 100-fs mid-infrared pulses in the water-absorption band using hollow-core photonic crystal fibre

High quality free-space and over-fibre transmission of mid-IR light is limited by factors such as material-related absorption, diffraction, light leakage and nonlinearity. Conventional vacuum apparatus can be utilized for high-quality laser-beam delivery to address these issues, the deployment of such apparatus would, however, increase the system complexity, being detrimental to their practical applications. Here we report the successful use of evacuated hollow-core photonic crystal fibre (PCF) to flexibly transmit ultrafast mid-IR pulses over several meters, while preserving exceptional spatial, spectral and temporal fidelity. The PCF was engineered to feature a low-loss transmission band within the water absorption range, and an evacuated 5-m length was used to transmit Watt-level, 100 fs pulses centred at around 2.8 microns. A comparison between free-space transmission and air-filled PCF highlights the superior performance of the evacuated hollow-core PCF, indicating its strong suitability for the flexible delivery of sub-ps laser pulses in the mid-IR.

physics.optics↗

arXiv2025

Three-stage dynamics of nonlinear pulse amplification in ultrafast mid-infrared fiber amplifier with anomalous dispersion

Nonlinear pulse amplification in optical fiber, with capability of breaking the gain-bandwidth limitation, is a key technique for high-energy, ultrafast pulse generation. In the longer wavelength region (including 1.55 μm, 2 μm and 2.8 μm) where the gain fiber has normally strong anomalous dispersion, the nonlinear amplification process over fiber exhibits more complicated dynamics than that of its 1-μm counterpart, and the underlying mechanism of the nonlinear pulse propagation process in high-gain anomalous fiber is still elusive so far. Here, we demonstrate an in-depth study on the nonlinear amplification process in high-gain ultrafast mid-infrared fiber, providing clear physical understanding on the debate of adiabatic soliton compression. We unveil that under the high-gain condition, the ultrafast pulse launched into the anomalous gain fiber experiences successively three distinct stages, named as the balance between linear and nonlinear chirp, high-order-soliton-like pulse compression and pulse splitting due to high-order effects. While a relatively-clean ultrafast pulse can be obtained immediately after the high-order-soliton-like compression stage, excessive gain fiber length could hardly enhance further the pulse peak power due to soliton splitting. Our findings can provide several critical guidelines for designing high-power ultrafast fiber amplifiers at near- and mid-infrared wavelengths.

physics.optics↗

arXiv2025

High-throughput calculations of two-dimensional auxetic $M_4X_8$ with magnetism, electrocatalysis, and alkali metal battery applications

Two-dimensional (2D) materials with multifunctional properties, such as negative Poisson's ratio (NPR), magnetism, catalysis, and energy storage capabilities, are of significant interest for advanced applications in flexible electronics, spintronics, catalysis, and lithium-ion batteries. However, the discovery of such materials, particularly in low-dimensional forms, remains a challenge. In this study, we perform high-throughput density-functional theory (DFT) calculations to explore a new class of 2D V-shaped monolayers with remarkable physicochemical properties. Among 18 stable $M_4X_8$ (M = transition metal; X = halogen) compounds, we identify 9 auxetic monolayers, with \ce{Pd4I8} standing out for its exceptionally high NPR of -0.798. Notably, 4 of these materials exhibit half semiconductor properties, while 5 others are bipolar magnetic semiconductors, offering a unique combination of electronic and magnetic behavior. Additionally, these materials demonstrate promising catalytic activity for hydrogen and oxygen evolution reactions (HER/OER) and show potential as anodes for rechargeable metal-ion batteries, particularly in alkali-ion systems. This work not only expands the family of 2D NPR materials but also introduces new candidates with multifunctional capabilities for a wide range of applications in nanoelectronics, catalysis, and energy storage.

cond-mat.mes-hall↗

arXiv2025

Physics-Informed Machine Learning for Efficient Reconfigurable Intelligent Surface Design

Reconfigurable intelligent surface (RIS) is a two-dimensional periodic structure integrated with a large number of reflective elements, which can manipulate electromagnetic waves in a digital way, offering great potentials for wireless communication and radar detection applications. However, conventional RIS designs highly rely on extensive full-wave EM simulations that are extremely time-consuming. To address this challenge, we propose a machine-learning-assisted approach for efficient RIS design. An accurate and fast model to predict the reflection coefficient of RIS element is developed by combining a multi-layer perceptron neural network (MLP) and a dual-port network, which can significantly reduce tedious EM simulations in the network training. A RIS has been practically designed based on the proposed method. To verify the proposed method, the RIS has also been fabricated and measured. The experimental results are in good agreement with the simulation results, which validates the efficacy of the proposed method in RIS design.

cs.LG↗