检索结果-内蒙古大学图书馆

Multi-Task Self-Supervised Learning for Medical Image Segmentation

学校读者我要写书评

暂无评论

Multi-Task Self-Supervised Learning for Medical Image Segmen...

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Bo Wang Hang Zhao Xiongfei Li Mingjie Tian Bo Huang Feiyang Yang Intelligent Information Processing Laboratory Yanbian University Yanji China College of Computer Science and Technology Jilin University Changchun China Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education Jilin University Changchun China School of Artificial Intelligence Jilin University Changchun China

Although medical image segmentation has achieved remarkable results with supervised learning, obtaining labeled data remains challenging and costly. To counteract this, we present the MTSPSeg, a multi-task self-supervised learning framework. We establish the dynamic gradient learning rate (DGLR) strategy and the dynamic multi-task loss weight adjustment method (DWL) to relieve potential early convergence and diverse learning difficulties of self-supervised learning. Our method deeply integrates the prior knowledge inherent in medical images, facilitates collaborative learning among tasks, and delves further into the latent information within the images. Experimental results demonstrate that regardless of the proportion of labeled data, our method outperforms baseline models on the KiTS19 and LiTS datasets, achieving state-of-the-art performance on the KiTS19 dataset.

关键词：

END-TO-END ENTITY DETECTION WITH PROPOSER AND REGRESSOR

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Wen, Xueru Zhou, Changjiang Tang, Haotian Liang, Luguang Jiang, Yu Qi, Hong College of Computer Science and Technology Jilin University Changchun China Key Laboratory of Symbolic Computation and Knowledge Engineering Ministry of Education Jilin University China

Named entity recognition is a traditional task in natural language processing. In particular, nested entity recognition receives extensive attention for the widespread existence of the nesting scenario. The latest research migrates the well-established paradigm of set prediction in object detection to cope with entity nesting. However, the manual creation of query vectors, which fail to adapt to the rich semantic information in the context, limits these approaches. An end-to-end entity detection approach with proposer and regressor is presented in this paper to tackle the issues. First, the proposer utilizes the feature pyramid network to generate high-quality entity proposals. Then, the regressor refines the proposals for generating the final prediction. The model adopts encoder-only architecture and thus obtains the advantages of the richness of query semantics, high precision of entity localization, and easiness of model training. Moreover, we introduce the novel spatially modulated attention and progressive refinement for further improvement. Extensive experiments demonstrate that our model achieves advanced performance in flat and nested NER, achieving a new state-of-the-art F1 score of 80.74 on the GENIA dataset and 72.38 on the WeiboNER dataset. Copyright © 2022, The Authors. All rights reserved.

关键词： Forecasting

Multimodal Hypothetical Summary for Retrieval-based Multi-image Question Answering

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Li, Peize Si, Qingyi Fu, Peng Lin, Zheng Wang, Yan School of Artificial Intelligence Jilin University Changchun China Institute of Information Engineering Chinese Academy of Sciences Beijing China Huawei Technologies Co. Ltd. Beijing China School of Cyber Security University of Chinese Academy of Sciences Beijing China Key Laboratory of Symbol Computation and Knowledge Engineering Ministry of Education College of Computer Science and Technology Jilin University Changchun China

Retrieval-based multi-image question answering (QA) task involves retrieving multiple question-related images and synthesizing these images to generate an answer. Conventional "retrieve-then-answer" pipelines often suffer from cascading errors because the training objective of QA fails to optimize the retrieval stage. To address this issue, we propose a novel method to effectively introduce and reference retrieved information into the QA. Given the image set to be retrieved, we employ a multimodal large language model (visual perspective) and a large language model (textual perspective) to obtain multimodal hypothetical summary in question-form and description-form. By combining visual and textual perspectives, MHyS captures image content more specifically and replaces real images in retrieval, which eliminates the modality gap by transforming into text-to-text retrieval and helps improve retrieval. To more advantageously introduce retrieval with QA, we employ contrastive learning to align queries (questions) with MHyS. Moreover, we propose a coarse-to-fine strategy for calculating both sentence-level and word-level similarity scores, to further enhance retrieval and filter out irrelevant details. Our approach achieves a 3.7% absolute improvement over state-of-the-art methods on RETVQA and a 14.5% improvement over CLIP. Comprehensive experiments and detailed ablation studies demonstrate the superiority of our method. Copyright © 2024, The Authors. All rights reserved.

关键词： Question answering

Sfmfold: A Multi-Modal Contrast Learning Model for Rna Secondary Structure Prediction

学校读者我要写书评

暂无评论

SSRN

SSRN 2025年

作者： Li, Ang Liu, Yuanning Zhang, Zhaoyang Li, Shuwang Dong, Hao Lv, Meng College of Computer Science and Technology Jilin University Changchun130012 China Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education Jilin University Changchun130012 China College of Software Jilin University Changchun130012 China College of Information Engineering Chang’an University Xi’an710064 China

The secondary structure of RNA can regulate the interaction between RNA and protein and perform various biological tasks through specific folding shapes, which plays an important role in revealing biomolecular mechanisms and developing new vaccines and therapeutic drugs. However, existing RNA secondary structure prediction methods based on thermodynamics and traditional neural network models usually lack the ability to model long-distance dependencies of RNA molecules and cannot handle complex non-nested structures in RNA molecules. In this paper, we propose a multi-modal contrastive learning method SFMFold based on state space model, which captures long-distance interactions in RNA sequences through RNA sequence representation learning part, and models the structure of RNA molecules through RNA structure representation learning part. Finally, the features of the two modalities are aligned and deeply fused to obtain the final RNA molecule prediction map. We test SFMFold on several datasets, and the experiments show that our proposed model shows excellent performance in cross-family prediction and long sequence prediction tasks. In addition, SFMFold can also effectively predict non-nested structures such as pseudoknot with better accuracy than the SOTA methods. © 2025, The Authors. All rights reserved.

关键词： Prediction models

Discrepancy-Guided Reconstruction Learning for Image Forgery Detection

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Shi, Zenan Chen, Haipeng Chen, Long Zhang, Dong College of Computer Science and Technology Jilin University China Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education Jilin University China Department of CSE The Hong Kong University of Science and Technology Hong Kong

In this paper, we propose a novel image forgery detection paradigm for boosting the model learning capacity on both forgery-sensitive and genuine compact visual patterns. Compared to the existing methods that only focus on the discrepant-specific patterns (e.g., noises, textures, and frequencies), our method has a greater generalization. Specifically, we first propose a Discrepancy-Guided Encoder (DisGE) to extract forgery-sensitive visual patterns. DisGE consists of two branches, where the mainstream backbone branch is used to extract general semantic features, and the accessorial discrepant external attention branch is used to extract explicit forgery cues. Besides, a Double-Head Reconstruction (DouHR) module is proposed to enhance genuine compact visual patterns in different granular spaces. Under DouHR, we further introduce a Discrepancy-Aggregation Detector (DisAD) to aggregate these genuine compact visual patterns, such that the forgery detection capability on unknown patterns can be improved. Extensive experimental results on four challenging datasets validate the effectiveness of our proposed method against state-of-the-art competitors. Copyright © 2023, The Authors. All rights reserved.

关键词： Textures

Efficient Heuristics for Learning Scalable Bayesian Network Classifier from Labeled and Unlabeled Data

学校读者我要写书评

暂无评论

SSRN

SSRN 2023年

作者： Wang, Limin Wang, Junjie Guo, Lu Li, Qilong Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education Jilin University Changchun130012 China College of Software Jilin University Changchun130012 China College of Instrumentation and Electrical Engineering Jilin University Changchun130012 China

Naive Bayes (NB) is one of the top ten machine learning algorithms whereas its attribute independence assumption rarely holds in practice. A feasible and efficient approach to improving NB is relaxing the assumption by adding augmented edges to the restricted topology of NB. In this paper we prove theoretically that the generalized topology may be a suboptimal solution to model multivariate probability distributions if its fitness to data cannot be measured. Thus we propose to apply log-likelihood function as the scoring function, then introduce an efficient heuristic search strategy to explore high-dependence relationships, and for each iteration the learned topology will be improved to fit data better. The proposed algorithm, called log-likelihood Bayesian classifier (LLBC), can respectively learn two submodels from labeled training set and individual unlabeled testing instance, and then make them work jointly for classification in the framework of ensemble learning. Our extensive experimental evaluations on 36 UCI benchmark datasets reveal that, LLBC demonstrates excellent classification performance and provides a competitive approach to learn from labeled and unlabeled data. © 2023, The Authors. All rights reserved.

关键词： Classification (of information)

Data-Efficient CLIP-Powered Dual-Branch Networks for Source-Free Unsupervised Domain Adaptation

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Li, Yongguang Cao, Yueqi Li, Jindong Wang, Qi Wang, Shengsheng College of Software Jilin University Changchun130012 China College of Computer Science and Technology Jilin University Changchun130012 China Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education Jilin University Changchun130012 China School of Artificial Intelligence Jilin University Changchun130012 China Engineering Research Center of Knowledge-Driven Human-Machine Intelligence Ministry of Education Jilin University Changchun130012 China

Source-free Unsupervised Domain Adaptation (SF-UDA) aims to transfer a model’s performance from a labeled source domain to an unlabeled target domain without direct access to source samples, addressing critical data privacy concerns. However, most existing SF-UDA approaches assume the availability of abundant source domain samples, which is often impractical due to the high cost of data annotation. To address the dual challenges of limited source data and privacy concerns, we introduce a data-efficient, CLIP-powered dual-branch network (CDBN). This architecture consists of a cross-domain feature transfer branch and a target-specific feature learning branch, leveraging high-confidence target domain samples to transfer text features of source domain categories while learning target-specific soft prompts. By fusing the outputs of both branches, our approach not only effectively transfers source domain category semantic information to the target domain but also reduces the negative impacts of noise and domain gaps during target training. Furthermore, we propose an unsupervised optimization strategy driven by accurate classification and diversity, preserving the classification capability learned from the source domain while generating more confident and diverse predictions in the target domain. CDBN achieves near state-of-the-art performance with far fewer source domain samples than existing methods across 31 transfer tasks on seven datasets. © 2024, CC BY.

关键词： Adversarial machine learning

ON A NOVEL UCP RESULT AND ITS APPLICATION TO INVERSE CONDUCTIVE SCATTERING

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Diao, Huaian Fei, Xiaoxu Liu, Hongyu School of Mathematics Key Laboratory of Symbolic Computation and Knowledge Engineering Ministry of Education Jilin University Changchun130012 China School of Mathematics and Statistics Central South University Changsha410083 China Department of Mathematics City University of Hong Kong Kowloon Hong Kong

In this paper, we derive a novel Unique Continuation Principle (UCP) for a system of second-order elliptic PDEs and apply it to investigate inverse problems in conductive scattering. The UCP relaxes the typical assumptions imposed on the domain or boundary with certain interior transmission conditions. This is motivated by the study of the associated inverse scattering problem and enables us to establish several novel unique identifiability results for the determination of generalized conductive scatterers using a single far-field pattern, significantly extending the results in [15, 21]. A key technical advancement in our work is the combination of Complex Geometric Optics (CGO) techniques from [15, 21] with the Fourier expansion method to microlocally analyze corner singularities and their implications for inverse problems. We believe that the methods developed can have broader applications in other contexts. © 2024, CC BY.

关键词： Inverse problems

Multifaceted User Modeling in Recommendation: A Federated Foundation Models Approach 39

学校读者我要写书评

暂无评论

Multifaceted User Modeling in Recommendation: A Federated Fo...

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Zhang, Chunxu Long, Guodong Guo, Hongkuan Liu, Zhaojie Zhou, Guorui Zhang, Zijian Liu, Yang Yang, Bo College of Computer Science and Technology Jilin University China Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education China Australian Artificial Intelligence Institute FEIT University of Technology Sydney Australia Kuaishou Technology China Institute for AI Industry Research Tsinghua University China

ISBN: (纸本)157735897X

Multifaceted user modeling aims to uncover fine-grained patterns and learn representations from user data, revealing their diverse interests and characteristics, such as profile, preference, and personality. Recent studies on foundation model-based recommendation have emphasized the Transformer architecture's remarkable ability to capture complex, non-linear user-item interaction relationships. This paper aims to advance foundation model-based recommendersystems by introducing enhancements to multifaceted user modeling capabilities. We propose a novel Transformer layer designed specifically for recommendation, using the self-attention mechanism to capture sequential user-item interaction patterns. Specifically, we design a group gating network to identify user groups, enabling hierarchical discovery across different layers, thereby capturing the multifaceted nature of user interests through multiple Transformer layers. Furthermore, to broaden the data scope and further enhance multifaceted user modeling, we extend the framework to a federated setting, enabling the use of private datasets while ensuring privacy. Experimental validations on benchmark datasets demonstrate the superior performance of our proposed method. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Federated learning