Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Xiaowei XuSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works91 from public data
- Federated Class-Incremental Learning327
A novel Global-Local Forgetting Compensation (GLFC) model is developed, to learn a global class-incremental model for alleviating the catastrophic forgetting from both local and global perspectives, and a prototype gradient-based communication mechanism is developed to protect the privacy.
- From System 1 to System 2: A Survey of Reasoning Large Language Models314
This survey provides an overview of the progress in foundational LLMs and the early development of System 2 technologies, exploring how their combination has paved the way for reasoning LLMs and explores promising directions for advancing reasoning LLMs.
- Data Poisoning Attacks on Federated Machine Learning310
This work attempts to attacking a popular federated multitask learning framework, which uses a general multitask learning framework to handle statistical challenges in the federated learning setting, and proposes a novel systems-aware optimization method, called as attack on federated learning (AT2FL), to efficiently derive the implicit gradients for poisoned data, and further attain optimal attack strategies in the federated machine learning.
- 248
- 154
- What Can Be Transferred: Unsupervised Domain Adaptation for Endoscopic Lesions Segmentation150
A new unsupervised semantic transfer model including two complementary modules for endoscopic lesions segmentation, which can alternatively determine where and how to explore transferable domain-invariant knowledge between labeled source lesions dataset and unlabeled target diseases dataset is developed.
- BoW3D: Bag of Words for Real-Time Loop Closing in 3D LiDAR SLAM121
The proposed BoW3D builds the bag of words based on the 3D LiDAR feature LinK3D, which is efficient, pose-invariant and can be used for accurate point-to-point matching and shows better performance in terms of max and extended precision scores on most scenes.
- 117
- 106
- No One Left Behind: Real-World Federated Class-Incremental Learning102
A proxy server is designed to tackle global forgetting caused by Non-IID class imbalance between different clients, and a category-balanced gradient-adaptive compensation loss and a category gradient-induced semantic distillation loss are developed.
- Exploring the Equivalence of Siamese Self-Supervised Learning via A Unified Gradient Framework74
UniGrad is proposed, a simple but effective gradient form for self-supervised learning that does not require a memory bank or a predictor network, but can still achieve state-of-the-art performance and easily adopt other training strategies.
- Xformer: Hybrid X-Shaped Transformer for Image Denoising73
A hybrid X-shaped vision Transformer, named Xformer, which performs notably on image denoising tasks, and proposes the Bidirectional Connection Unit (BCU) to couple the learned representations from these two branches while providing enhanced information fusion.
- Federated Incremental Semantic Segmentation71
A Forgetting-Balanced Learning (FBL) model is proposed to address heterogeneous forgetting on old classes from both intra-client and interclient aspects to address catastrophic forgetting in Federated learning-based semantic segmentation.
- LinK3D: Linear Keypoints Representation for 3D LiDAR Point Cloud67
A novel 3D feature representation method that fully considers the characteristics (such as the sparsity and complexity) of LiDAR point clouds and represents the keypoint with its robust neighbor keypoints, which provides strong constraints in the description of the keypoint.
- Create Your World: Lifelong Text-to-Image Diffusion66
A model which intends to overcome knowledge “catastrophic forgetting” for the past encountered concepts, and semantic “catastrophic neglecting” for one or more concepts in the text prompt is proposed.
- 51
- Semantic-Transferable Weakly-Supervised Endoscopic Lesions Segmentation50
A pseudo label generator is proposed to leverage seed information to generate highly-confident pseudo pixel labels by incorporating class balance and super-pixel spatial prior in a new semantic lesions representation transfer model for weakly-supervised endoscopic lesions segmentation.
- CSCL: Critical Semantic-Consistent Learning for Unsupervised Domain Adaptation47
A critical transfer based adversarial framework is designed to highlight transferable domain-wise knowledge while neglecting untransferable knowledge, and a symmetric soft divergence loss is presented to explore inter-class relationships and facilitate category-wise distribution alignment.
- Weakly-Supervised Cross-Domain Adaptation for Endoscopic Lesions Segmentation45
A new weakly-supervised lesions transfer framework is proposed, which can not only explore transferable domain-invariant knowledge across different datasets, but also prevent the negative transfer of untransferable representations.
- I3DOL: Incremental 3D Object Learning without Catastrophic Forgetting42
A new Incremental 3D Object Learning (i.e., I3DOL) model is proposed, which is the first exploration to learn new classes of 3D object continually, and an adaptive-geometric centroid module is designed to construct discriminative local geometric structures, which can better characterize the irregular point cloud representation for3D object.
- 40
- 40
- Lifelong robotic visual-tactile perception learning38
A new L ifelong V isual- T actile L earning (LVTL) model is proposed for continuous robotic visual-tactile perception tasks, which fully explores the latent correlations in both intra- modality and cross-modality aspects.
- 36
- Gradient-Semantic Compensation for Incremental Semantic Segmentation35
A step-aware gradient compensation that can balance forgetting paces of previously seen classes by re-weighting gradient back-propagation is developed and a soft-sharp semantic relation distillation is proposed to distill consistent inter-class semantic relations via soft labels for alleviating catastrophic forgetting from the semantic aspect.
- 35
- 33
- Self-Paced Weight Consolidation for Continual Learning30
A self-paced Weight Consolidation framework to attain robust continual learning via evaluating the discriminative contributions of previous tasks is proposed and an alternative convex search to iteratively update the model parameters and priority weights in the bi-convex formulation is adopted.
- Heterogeneous Forgetting Compensation for Class-Incremental Learning30
A novel Heterogeneous Forgetting Compensation (HFC) model is developed, which can resolve heterogeneous forgetting of easy-to-forget and hard- to- forget old categories from both representation and gradient aspects.
- 27
- 25
- Continual LLaVA: Continual Instruction Tuning in Large Vision-Language Models24
This work proposes a new benchmark for COntinuAl inStruction Tuning on LVLMs (COAST), which encompasses the aforementioned domain-incremental, capability-incremental, and dataset-incremental configurations and proposes Continual LLaVA, a rehearsal-free method tailored for continual instruction tuning in LVLMs.
- Task Relation Distillation and Prototypical Pseudo Label for Incremental Named Entity Recognition22
To tackle catastrophic forgetting, a task relation distillation scheme is introduced that serves two purposes: ensuring inter-task semantic consistency across different incremental learning tasks by minimizing inter- task relation Distillation loss, and enhancing the model's prediction confidence by minimizing intra-task self-entropy loss.
- PixelVLA: Advancing Pixel-level Understanding in Vision-Language-Action Model21
The first VLA model designed to support both pixel-level reasoning and multimodal prompting with text and visual inputs is introduced, built on a new visuomotor instruction tuning framework that integrates a multiscale pixel-aware encoder with a visual promptaware encoder.
- 20
- InOR-Net: Incremental 3-D Object Recognition Network for Point Cloud Representation19
A novel Incremental 3-D Object Recognition Network (i.e., InOR-Net) is developed, which could recognize new classes of 2-D objects continuously by overcoming the catastrophic forgetting on old classes and a dual adaptive fairness compensations strategy is designed to overcome the forgetting brought by class imbalance.
- MuseumMaker: Continual Style Customization Without Catastrophic Forgetting17
This paper proposes MuseumMaker, a method that enables the synthesis of images by following a set of customized styles in a never-end manner, and gradually accumulates these creative artistic works as a Museum, and designs a task-wise token learning module where a unique token embedding is learned to denote a new style.
- Generative Partial Visual-Tactile Fused Object Clustering16
A conditional cross-modal clustering generative adversarial network is developed to synthesize one modality conditioning on the other modality, which can compensate missing samples and align the visual and tactile modalities naturally by adversarial learning.
- Evolving Metric Learning for Incremental and Decremental Features15
A new online Evolving Metric Learning (EML) model for incremental and decremental features, which can handle the instance and feature evolutions simultaneously by incorporating with a smoothed Wasserstein metric distance is developed.
- 14
- Never-Ending Behavior-Cloning Agent for Robotic Manipulation12
This work investigates the above challenges with a never-ending behavior-cloning agent (NBAgent) in embodied robots, a pioneering language-conditioned NBAgent, which can continually learn observation knowledge of novel 3-D scene semantics and robot manipulation skills from skill-shared and skill-specific attributes.
- 10
- 10
- Dual Hyperspectral Mamba for Efficient Spectral Compressive Imaging9
This work proposes a novel Dual Hyperspectral Mamba (DHM) to explore both global long-range dependencies and local contexts for efficient HSI reconstruction, and demonstrates the benefits of the DHM for HSI reconstruction.
- 9
- 9
- PECTP: Parameter-Efficient Cross-Task Prompts for Incremental Vision Transformer8
A Parameter-Efficient Cross-Task Prompt (PECTP) framework with Prompt Retention Module (PRM) and classifier Head Retention Module (HRM) and HRM to inherit old knowledge in the previously learned classifier heads to facilitate the cross-task prompts’ generalization ability.
- MedKGent: a large language model agent framework for constructing temporally evolving medical knowledge graph7
In downstream evaluations, MedKGent-KG significantly improved retrieval-augmented generation for five LLMs across seven medical question-answering benchmarks, and position MedKGent as a scalable and temporally aware infrastructure for medical knowledge representation and literature-grounded AI research.
- 7
- Federated Continual Novel Class Learning7
A Global Alignment Learning (GAL) framework is proposed that can accurately estimate the global novel class number and provide effective guidance for local training from a global perspective and provide effective guidance for local training from a global perspective, all while maintaining privacy protection.
- Federated Incremental Named Entity Recognition6
A structural knowledge distillation loss to retain the latent space's feature structure and a pseudo-label-guided inter-type contrastive loss to enhance discriminative capability over different entity types are developed, effectively preserving previously learned knowledge within local clients.
- Federated Learning with New Knowledge: Fundamentals, Advances, and Futures6
This paper systematically defines the main sources of new knowledge in FL, including new features, tasks, models, and algorithms, and comprehensively discusses the potential future directions for FL with new knowledge, considering a variety of factors such as scenario setups, efficiency, and security.
- Accurate Feature Extraction from Historical Geologic Maps Using Open-Set Segmentation and Detection6
This study presents a novel AI method for extracting polygon and point features from historical geologic maps, representing a pivotal step for assessing the mineral resources needed for energy transition, using map units in the legends as prompts for one-shot segmentation and detection in geological feature extraction.
- I3DOD: Towards Incremental 3D Object Detection via Prompting6
This work presents a novel Incremental 3D Object Detection framework with the guidance of prompting and proposes a task-shared prompts mechanism to learn the matching relationships between the object localization information and category semantic information.
- Rehearsal-free Federated Domain-incremental Learning5
A rehearsal-free federated domain incremental learning framework, RefFiL, based on a global prompt-sharing paradigm to alleviate catastrophic forgetting challenges in federated domain-incremental learning, where unseen domains are continually learned.
- Deep self‐cleansing for medical image segmentation with noisy labels5
This research presents a novel approach to segmentation that combines supervised deep learning with reinforcement learning to solve the challenge of integrating image classification and annotation into a discrete-time model.
- Unsupervised Dense Deformation Embedding Network for Template-Free Shape Correspondence5
A new Unsupervised Dense Deformation Embedding Network, which learns to predict deformations between non-rigid shapes from dense local features and introduces a coarse parameterized deformation graph, for which a novel trace and propagation algorithm is proposed to improve both the quality and efficiency of the deformation.
- BioXArena: Benchmarking LLM Agents on Multi-Modal Biomedical Machine Learning Tasks4
BioXArena is a biomedical machine learning benchmark designed to evaluate whether agents can generate task-specific model training pipelines for heterogeneous and multi-modal biomedical datasets, and performs extensive ablation studies, robustness evaluations, scaling analyses, cost analyses, and failure-mode investigations.
- 4
- CRISP: Contrastive Residual Injection and Semantic Prompting for Continual Video Instance Segmentation3
This work introduces Contrastive Residual Injection and Semantic Prompting (CRISP), a framework tailored to address instance-wise, category-wise, and task-wise confusion in CVIS, and introduces a semantic consistency loss based on the contrastive learning.
- Thinking with Drafts- Speculative Temporal Reasoning for Efficient Long Video Understanding3
Experiments across multiple video understanding benchmarks demonstrate that SpecTemp not only maintains competitive accuracy but also significantly accelerates inference compared with existing thinking-with-frames methods.
- Hierarchical Visual Prompt Learning for Continual Video Instance Segmentation3
A novel Hierarchical Visual Prompt Learning (HVPL) model is developed that overcomes catastrophic forgetting of previous categories from both frame-level and video-level perspectives and proves to be more effective than baseline approaches.
- 3
- 3
- Bring Your Dreams to Life: Continual Text-to-Video Customization2
A novel Continual Customized Video Diffusion model is developed that can continuously learn new concepts to generate videos across various text-to-video generation tasks by tackling forgetting and concept neglect and outperforms existing CTVG models.
- SCas4D: Structural Cascaded Optimization for Boosting Persistent 4D Novel View Synthesis2
SCas4D, a cascaded optimization framework that leverages structural patterns in 3D Gaussian Splatting for dynamic scenes, achieves convergence within 100 iterations per time frame and produces results comparable to existing methods with only one-twentieth of the training iterations.
- 2
- Towards AI-Architecture Liberty: A Comprehensive Survey on Design and Generation of Virtual Architecture by Deep Learning2
This survey focuses on investigating and comparing the current latest approaches to 3D object generation with deep generative models (DGMs), including Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), 3D-aware images, and diffusion models.
- Crafting Your Evolving Dreams: Concept-Incremental Versatile Customization1
This work develops a novel Continually Customizable Diffusion Model (CCDM), enabling users to perform concept-incremental versatile customization, and proposes a controllable regional context synthesis strategy that performs multi-concept composition in alignment with user-provided conditions.
- Complementary Information Guided Occupancy Prediction via Multi-Level Representation Fusion1
CIGOcc extracts segmentation, graphics, and depth features from an input image and introduces a deformable multi-level fusion mechanism to fuse these three multi-level features, which achieves state-of-the-art performance on the SemanticKITTI benchmark.
- Exploring Stability-Plasticity Trade-Offs for Continual Named Entity Recognition1
The SPT method surpasses previous CNER approaches in achieving a suitable stability-plasticity trade-off, and develops a confidence-based pseudo-labeling approach for the current non-entity type, which predicts entity types using the old model to handle the semantic shift of the non-entity type.
- 1
- 1
- 1
- SimC3D: A Simple Contrastive 3D Pretraining Framework Using RGB Images1
It is empirically demonstrated that the primary benefit of the 2D modality stems from the incorporation of locality information, and SimC3D directly employs 2D positional embeddings as a stronger contrastive objective, eliminating the necessity for 2D backbones and leading to considerable performance improvements.
- 1
- 1
- 1
- –
- ReasonWalker: Reasoning Iterative Vision-and-Language Navigation With Implicit Instructions–
This article proposes ReasonWalker, a novel navigation model designed to enable reasoning-based navigation using implicit instructions over time, and proposes a new hierarchical learning paradigm, where the model first learns navigation actions and then acquires scene associations for implicit instruction reasoning.
- Patch-Based Active Source-Free Domain Adaptation for Annotation-Efficient Medical Image Segmentation–
- –
- –
- –
- Accurate Tracking of Arabidopsis Root Cortex Cell Nuclei in 3D Time-Lapse Microscopy Images Based on Genetic Algorithm–
This work proposes an accurate tracking method based on a Genetic Algorithm that incorporates knowledge of Arabidopsis root cellular patterns and spatial relationships among volumes that achieves accurate nuclear tracking in time-lapse microscopy of the root meristem.
- Optimizing Vision-Language Consistency via Cross-Layer Regional Attention Alignment–
Consistent Cross-layer Regional Alignment (CCRA) is proposed, which introduces Layer-Patch-wise Cross Attention (LPWCA) to capture fine-grained regional-semantic correlations by jointly weighting patch and layer-wise embedding, and Progressive Attention Integration (PAI) that systematically coordinates LPWCA, layer-wise, and patch-wise attention mechanisms in sequence.
- –
- –
- –
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-14. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.