检索结果-内蒙古大学图书馆

Causal structure learning of nonlinear additive noise model based on streaming feature

学校读者我要写书评

暂无评论

Causal structure learning of nonlinear additive noise model ...

IEEE International Conference on data Mining Workshops (ICDM Workshops)

作者： Jing Yang Gaojin Fan Anbo Shen Aiguo Wang Ministry of Education Key Laboratory of Knowledge Engineering With Big Data(Hefei University of Technology) School of Computer Science and Information Engineering Hefei University of Technology Hefei P. R. China School of Electronic and Information Engineering Foshan University China

ISBN: (纸本)9781665424288

The current streaming feature structure learning needs to be improved in the processing of nonlinear continuous data and the dynamic acquisition of causal structures. In this paper, we propose a causal structure learning algorithm, CANSF, based on the streaming feature of additive noise models. We have made three contributions. First, by using the information carried by the noise of nonlinear continuous data, we propose a real correlation identification method based on logarithmic likelihood, which can identify the real correlation and redundant features of target features, and dynamically select parent and child nodes for each feature. Second, based on regression analysis, a method to determine the causal direction is proposed, which can be used for dynamic orientation. Third, a learning method of causal structure based on streaming features is proposed, which can obtain the Causal structure diagram directly and dynamically.

关键词： Additive noise Learning systems Correlation Heuristic algorithms Conferences data models Regression analysis

Analyzing the Spillover Effects of the Binance Incident on Financial Markets: A Machine Learning Approach

学校读者我要写书评

暂无评论

SSRN

SSRN 2024年

作者： Feng, Lingbing Qi, Jiajun Liu, Ye Wang, Wei School of Statistics and Data Science Jiangxi University of Finance and Economics Jiangxi Nanchang330013 China Key Laboratory of Data Science in Finance and Economics Jiangxi University of Finance and Economics Jiangxi Nanchang330013 China Research Centre for Financial Development and Risk Prevention Jiangxi University of Finance and Economics Jiangxi Nanchang330013 China School of Applied Economics Jiangxi University of Finance and Economics Jiangxi Nanchang330013 China School of Management and Engineering Nanjing University Jiangsu Nanjing210093 China School of Economics and Business Administration Chongqing University Chongqing400044 China

This paper integrates event study with machine learning to analyze the spillover effects of the "Binance Incident" in the cryptocurrency market on financial markets. Utilizing the Lasso model, this paper predicts normal returns during the event window, outperforming traditional market models. Findings reveal positive effects on precious metal markets due to their safe-haven nature, while energy markets exhibit negative effects due to reduced investor confidence and weaker safe-haven attributes. Fiat currencies show negligible reactions. Summarizing across assets, this paper finds a negative overall impact of this adverse event on the financial markets. © 2024, The Authors. All rights reserved.

关键词： Machine learning

NEW BOOLEAN SATISFIABILITY PROBLEM HEURISTIC STRATEGY: MINIMAL POSITIVE NEGATIVE PRODUCT STRATEGY

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Qun, Zhao Xintao, Wang Menghui, Yang School of Information Resource Management Renmin University of China Beijing100872 China Institute of Economics Chinese Academy of Social Sciences Beijing100872 China Key Laboratory of Data Engineering and Knowledge Engineering Ministry of Education Beijing100872 China

This study presents a novel heuristic algorithm called the "Minimal Positive Negative Product Strategy" to guide the CDCL algorithm in solving the Boolean satisfiability problem. It provides a mathematical explanation for the superiority of this algorithm over widely used heuristics such as the Dynamic Largest Individual Sum (DLIS) and the Variable State Independent Decaying Sum (VSIDS). Experimental results further confirm the effectiveness of this heuristic strategy in *** Codes 03 Copyright © 2023, The Authors. All rights reserved.

关键词： Heuristic algorithms

An improved elastic net with heuristic cloning strategy for traveling salesman problem

学校读者我要写书评

暂无评论

ICIC Express Letters 2011年第11期5卷 3995-4000页

作者： Yang, Gang Yang, Nan Yi, Junyan Key Laboratory of Data Engineering and Knowledge Engineering MOE School of Information Renmin University of China No. 59 Zhongguancun Street Haidian District Beijing 100872 China Department of Computer Science and Technology Zhejiang University of Technology No. 18 Chaowang Road Hangzhou 310014 China

Based on analysis of relationships between solution qualities and number of initial dynamic points in elastic net algorithm, we propose an improved elastic network algorithm (IENA) introduced in a heuristic cloning strategy for traveling salesman problem (TSP). The improved elastic net consists of four evolution processes with less dynamic points and gradually reducing cities, and it accelerates the convergent process without compromising the equality of solutions. IENA is conducted on some random TSP instances to verify its performance by comparison with other algorithms. © 2011 ICIC International.

关键词： Traveling salesman problem

Unifying Large Language Models and knowledge Graphs: A Roadmap

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Pan, Shirui Luo, Linhao Wang, Yufei Chen, Chen Wang, Jiapu Wu, Xindong Griffith University QLD Australia The Department of Data Science and AI Monash University Melbourne Australia The Nanyang Technological University Singapore The Faculty of Information Technology Beijing University of Technology Beijing China The Key Laboratory of Knowledge Engineering with Big Data The Ministry of Education of China Hefei University of Technology Hefei China The Research Center for Knowledge Engineering Zhejiang Lab Hangzhou China

Large language models (LLMs), such as ChatGPT and GPT4, are making new waves in the field of natural language processing and artificial intelligence, due to their emergent ability and generalizability. However, LLMs are black-box models, which often fall short of capturing and accessing factual knowledge. In contrast, knowledge Graphs (KGs), Wikipedia and Huapu for example, are structured knowledge models that explicitly store rich factual knowledge. KGs can enhance LLMs by providing external knowledge for inference and interpretability. Meanwhile, KGs are difficult to construct and evolve by nature, which challenges the existing methods in KGs to generate new facts and represent unseen knowledge. Therefore, it is complementary to unify LLMs and KGs together and simultaneously leverage their advantages. In this article, we present a forward-looking roadmap for the unification of LLMs and KGs. Our roadmap consists of three general frameworks, namely, 1) KG-enhanced LLMs, which incorporate KGs during the pre-training and inference phases of LLMs, or for the purpose of enhancing understanding of the knowledge learned by LLMs;2) LLM-augmented KGs, that leverage LLMs for different KG tasks such as embedding, completion, construction, graph-to-text generation, and question answering;and 3) Synergized LLMs + KGs, in which LLMs and KGs play equal roles and work in a mutually beneficial way to enhance both LLMs and KGs for bidirectional reasoning driven by both data and knowledge. We review and summarize existing efforts within these three frameworks in our roadmap and pinpoint their future research directions. Copyright © 2023, The Authors. All rights reserved.

关键词： knowledge graph

Exploring Heterogeneity and Uncertainty for Graph-based Cognitive Diagnosis Models in Intelligent Education

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Shao, Pengyang Yang, Yonghui Gao, Chen Chen, Lei Zhang, Kun Zhuang, Chenyi Wu, Le Li, Yong Wang, Meng Key Laboratory of Knowledge Engineering with Big Data Hefei University of Technology Anhui Hefei China BNRist Tsinghua University Beijing China Department of Electronic Engineering BNRist Tsinghua University Beijing China Ant Group Zhejiang Hangzhou China

Graph-based Cognitive Diagnosis (CD) has attracted much research interest due to its strong ability on inferring students’ proficiency levels on knowledge concepts. While graph-based CD models have demonstrated remarkable performance, we contend that they still cannot achieve optimal performance due to the neglect of edge heterogeneity and uncertainty. Edges involve both correct and incorrect response logs, indicating heterogeneity. Meanwhile, a response log can have uncertain semantic meanings, e.g., a correct log can indicate true mastery or fortunate guessing, and a wrong log can indicate a lack of understanding or a careless mistake. In this paper, we propose an Informative Semantic-aware Graph-based Cognitive Diagnosis model (ISG-CD), which focuses on how to utilize the heterogeneous graph in CD and minimize effects of uncertain edges. Specifically, to explore heterogeneity, we propose a semantic-aware graph neural networks based CD model. To minimize effects of edge uncertainty, we propose an Informative Edge Differentiation layer from an information bottleneck perspective, which suggests keeping a minimal yet sufficient reliable graph for CD in an unsupervised way. We formulate this process as maximizing mutual information between the reliable graph and response logs, while minimizing mutual information between the reliable graph and the original graph. After that, we prove that mutual information maximization can be theoretically converted to the classic binary cross entropy loss function, while minimizing mutual information can be realized by the Hilbert-Schmidt Independence Criterion. Finally, we adopt an alternating training strategy for optimizing learnable parameters of both the semantic-aware graph neural networks based CD model and the edge differentiation layer. Extensive experiments on three real-world datasets have demonstrated the effectiveness of ISG-CD. © 2024, CC BY.

关键词： knowledge graph

Visible Thermal Person Re-Identification Via Multi-Branch Modality Residual Complementary Learning

学校读者我要写书评

暂无评论

SSRN

SSRN 2024年

作者： Chen, Long Sun, Rui Yu, Yiheng Du, Yun Zhang, Xudong Key Laboratory of Knowledge Engineering with Big Data Ministry of Education School of Computer and Information Anhui Province Key Laboratory of Industry Safety and Emergency Technology Hefei University of Technology Hefei230009 China The School of Computer and Information Anhui Province Key Laboratory of Industry Safety and Emergency Technology Hefei University of Technology Hefei230009 China

Visible-thermal Person Re-identification (VT Re-ID) is a challenging task in all-weather surveillance system. Existing methods focus on extracting the modality-shared features, ignoring the discriminative inter-modality complementary features. To tackle this issue, we propose the multi-branch modality residual complementary learning method consisting of modality residual complementary learning (MRCL) module and multi-branch feature learning (MBFL) module. The MRCL module can be easily integrated into existing CNN and Transformer baselines and drive the network to focus on both intra-modality and inter-modality information simultaneously. On one hand, we adopt the basic two-stream network to obtain the intra-modality features, on the other hand, we feed the residual image acquired by cross-modality correlation saliency erasing operation to another two-stream network to capture the inter-modality complementary features. Moreover, we employ the MBFL module to capture local spatial features and local channel features and integrate them with global features to increase the discriminability and robustness of the ultimate representations for handling intra-modality variations. The extensive experiments on RegDB and SYSU-MM01 datasets demonstrate the superiority of our proposed method compared with state-of-the-art methods. © 2024, The Authors. All rights reserved.

关键词： Learning systems

Adaptive unimodal cost volume filtering for deep stereo matching

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Zhang, Youmin Chen, Yimin Bai, Xiao Yu, Suihanjin Yu, Kun Li, Zhiwei Yang, Kuiyuan State Key Laboratory of Software Development Environment School of Computer Science and Engineering Beijing Advanced Innovation Center for Big Data and Brain Computing Jiangxi Research Institute Beihang University Beijing China DeepMotion

State-of-the-art deep learning based stereo matching approaches treat disparity estimation as a regression problem, where loss function is directly defined on true disparities and their estimated ones. However, disparity is just a byproduct of a matching process modeled by cost volume, while indirectly learning cost volume driven by disparity regression is prone to overfitting since the cost volume is under constrained. In this paper, we propose to directly add constraints to the cost volume by filtering cost volume with unimodal distribution peaked at true disparities. In addition, variances of the unimodal distributions for each pixel are estimated to explicitly model matching uncertainty under different contexts. The proposed architecture achieves state-of-the-art performance on Scene Flow and two KITTI stereo benchmarks. In particular, our method ranked the 1st place of KITTI 2012 evaluation and the 4th place of KITTI 2015 evaluation (recorded on 2019.8.20). The codes of AcfNet are available at: https://***/DeepMotionAIResearch/ DenseMatchingBenchmark. Copyright © 2019, The Authors. All rights reserved.

关键词： Uncertainty analysis

Hierarchical Multi-Granular Decision Networks for Hierarchical Classification

学校读者我要写书评

暂无评论

SSRN

SSRN 2024年

作者： Liu, Haoyang Hu, Xue gang Bai, Shengxing Lin, Yaojin Key Laboratory of Knowledge Engineering with Big Data Hefei University of Technology Ministry of Education China School of Computer Science and Information Engineering Hefei University of Technology Anhui Hefei230009 China Anhui Province Key Laboratory of Industry Safety and Emergency Technology Anhui Hefei230009 China School of Computer Science Minnan Normal University Fujian Zhangzhou363000 China Key Laboratory of Data Science and Intelligence Application Minnan Normal University Fujian Zhangzhou363000 China

Hierarchical classification learning is an effective means of solving large-scale classification problems, which models classification problems at different levels of granularity according to a hierarchical structure. In hierarchical classification tasks, classification results at different levels of granularity are mutually beneficial, where coarse-grained results can reduce the number of fine-grained candidate categories, and fine-grained results can provide a basis for coarse-grained classification. However, for the final decision, the contribution of the granularity levels in the different hierarchies is different and unknown. Feedback from different granularity levels should be flexibly integrated. In addition, errors in hierarchical classification are not equivalent, and some errors can have serious results for practical applications. In order to address the above challenges, we propose a new hierarchical network to integrate multi-granularity knowledge for decision making. On the one hand, the network computes the decision scores of instances at each granularity level and learns a set of weights to combine all scores. Based on the learnable weights, the model can effectively utilize the multi-granularity classification results to obtain the final decision. On the other hand, we design local and global classification losses to guide model optimization, which allows the model to gain a hierarchical perspective and the ability to understand the degree of hierarchical misclassification. As a result, the extent to which the method results in misclassification consequences is limited. Experimental results on five benchmark datasets show that our method arises as a state-of-the-art hierarchical classification method. © 2024, The Authors. All rights reserved.

关键词： Deep neural networks