A5100408983- Topic Modeling
- Natural Language Processing Techniques
- Advanced Text Analysis Techniques
- Text and Document Classification Technologies
- Speech and dialogue systems
- Complex Network Analysis Techniques
- Semantic Web and Ontologies
- Recommender Systems and Techniques
- Multimodal Machine Learning Applications
- Advanced Graph Neural Networks
- Sentiment Analysis and Opinion Mining
- Web Data Mining and Analysis
- Advanced Computational Techniques and Applications
- Biomedical Text Mining and Ontologies
- Speech Recognition and Synthesis
- Opinion Dynamics and Social Influence
- Optimization and Search Problems
- Advanced Image and Video Retrieval Techniques
- Scheduling and Optimization Algorithms
- Data Management and Algorithms
- Machine Learning and Data Classification
- Music and Audio Processing
- Generative Adversarial Networks and Image Synthesis
- Advanced Bandit Algorithms Research
- Anomaly Detection Techniques and Applications
Hong Kong Polytechnic University
2016-2025
Central South University
2019-2025
University Town of Shenzhen
2025
Tsinghua University
2003-2025
Beijing Shijitan Hospital
2024-2025
Capital Medical University
2024-2025
Changzhou University
2016-2025
Guiyang Medical University
2024-2025
Wuhan Polytechnic University
2021-2025
New York University
2023-2025
To address the sparsity and cold start problem of collaborative filtering, researchers usually make use side information, such as social networks or item attributes, to improve recommendation performance. This paper considers knowledge graph source information. limitations existing embedding-based path-based methods for knowledge-graph-aware recommendation, we propose RippleNet, an end-to-end framework that naturally incorporates into recommender systems. Similar actual ripples propagating...
To alleviate sparsity and cold start problem of collaborative filtering based recommender systems, researchers engineers usually collect attributes users items, design delicate algorithms to exploit these additional information. In general, the are not isolated but connected with each other, which forms a knowledge graph (KG). this paper, we propose Knowledge Graph Convolutional Networks (KGCN), an end-to-end framework that captures inter-item relatedness effectively by mining their...
We develop a high-quality multi-turn dialog dataset, DailyDialog, which is intriguing in several aspects. The language human-written and less noisy. dialogues the dataset reflect our daily communication way cover various topics about life. also manually label developed with intention emotion information. Then, we evaluate existing approaches on DailyDialog hope it benefit research field of systems.
Knowledge graphs capture structured information and relations between a set of entities or items. As such knowledge represent an attractive source that could help improve recommender systems. However, existing approaches in this domain rely on manual feature engineering do not allow for end-to-end training. Here we propose Knowledge-aware Graph Neural Networks with Label Smoothness regularization (KGNN-LS) to provide better recommendations. Conceptually, our approach computes user-specific...
Recommender systems are used to accurately and actively provide users with potentially interesting information or services. Collaborative filtering is a widely adopted approach recommendation, but sparse data cold-start often barriers providing high quality recommendations. To address such issues, we propose novel method that works improve the performance of collaborative recommendations by integrating rating given social trust network among these same users. This model-based adopts matrix...
Collaborative filtering often suffers from sparsity and cold start problems in real recommendation scenarios, therefore, researchers engineers usually use side information to address the issues improve performance of recommender systems. In this paper, we consider knowledge graphs as source information. We propose MKR, a Multi-task feature learning approach for Knowledge graph enhanced Recommendation. MKR is deep end-to-end framework that utilizes embedding task assist task. The two tasks...
Unlike extractive summarization, abstractive summarization has to fuse different parts of the source text, which inclines create fake facts. Our preliminary study reveals nearly 30% outputs from a state-of-the-art neural system suffer this problem. While previous approaches usually focus on improvement informativeness, we argue that faithfulness is also vital prerequisite for practical system. To avoid generating facts in summary, leverage open information extraction and dependency parse...
Although Generative Adversarial Networks achieve state-of-the-art results on a variety of generative tasks, they are regarded as highly unstable and prone to miss modes. We argue that these bad behaviors GANs due the very particular functional shape trained discriminators in high dimensional spaces, which can easily make training stuck or push probability mass wrong direction, towards higher concentration than data generating distribution. introduce several ways regularizing objective,...
Most previous seq2seq summarization systems purely depend on the source text to generate summaries, which tends work unstably. Inspired by traditional template-based approaches, this paper proposes use existing summaries as soft templates guide model. To end, we a popular IR platform Retrieve proper candidate templates. Then, extend framework jointly conduct template Reranking and template-aware summary generation (Rewriting). Experiments show that, in terms of informativeness, our model...
Jingjing Xu, Xu Sun, Qi Zeng, Xiaodong Zhang, Xuancheng Ren, Houfeng Wang, Wenjie Li. Proceedings of the 56th Annual Meeting Association for Computational Linguistics (Volume 1: Long Papers). 2018.
Despite the successes in capturing continuous distributions, application of generative adversarial networks (GANs) to discrete settings, like natural language tasks, is rather restricted. The fundamental reason difficulty back-propagation through random variables combined with inherent instability GAN training objective. To address these problems, we propose Maximum-Likelihood Augmented Discrete Generative Adversarial Networks. Instead directly optimizing objective, derive a novel and...
Unlike extractive summarization, abstractive summarization has to fuse different parts of the source text, which inclines create fake facts. Our preliminary study reveals nearly 30% outputs from a state-of-the-art neural system suffer this problem. While previous approaches usually focus on improvement informativeness, we argue that faithfulness is also vital prerequisite for practical system. To avoid generating facts in summary, leverage open information extraction and dependency parse...
Manufacturing organizations have a pivotal role in reducing the adverse impact of global warming by adopting sustainable practices and producing environmentally-friendly products. Organizations are engaged environmental corporate social responsibility (ECSR) emphasize green intellectual capital (GIC), innovative products support for business sustainability (BUS). The current study aims to analyze organizational ECSR GIC on innovation (GIN) BUS. data 237 participants from manufacturing firms...
It is difficult to identify sentence importance from a single point of view.In this paper, we propose learning-based approach combine various features.They are categorized as surface, content, relevance and event features.Surface features related extrinsic aspects sentence.Content measure based on contentconveying words.Event represent sentences by events they contained.Relevance evaluate its relatedness with other sentences.Experiments show that the combined improved summarization...
Topic modeling techniques have the benefits of words and documents uniformly under a probabilistic framework. However, they also suffer from limitations sensitivity to initialization unigram topic distribution, which can be remedied by deep learning techniques. To explore combination techniques, we first explain standard modelfrom perspective neural network. Based on this, propose novel model (NTM) where representation are efficiently naturally combined into uniform Extending NTM, easily add...
Ziqiang Cao, Furu Wei, Sujian Li, Wenjie Ming Zhou, Houfeng Wang. Proceedings of the 53rd Annual Meeting Association for Computational Linguistics and 7th International Joint Conference on Natural Language Processing (Volume 2: Short Papers). 2015.
To address the sparsity and cold-start problem of collaborative filtering, researchers usually make use side information, such as social networks or item attributes, to improve performance recommendation. In this article, we consider knowledge graph (KG) source information. limitations existing embedding-based path-based methods for KG-aware recommendation, propose RippleNet , an end-to-end framework that naturally incorporates KG into recommender systems. has two versions: (1) The outward...
Xiaoyu Shen, Hui Su, Yanran Li, Wenjie Shuzi Niu, Yang Zhao, Akiko Aizawa, Guoping Long. Proceedings of the 55th Annual Meeting Association for Computational Linguistics (Volume 2: Short Papers). 2017.
Previous researches on Text-level discourse parsing mainly made use of constituency structure to parse the whole document into one tree.In this paper, we present limitations based and first propose dependency directly represent relations between elementary units (EDUs).The state-of-the-art techniques, Eisner algorithm maximum spanning tree (MST) algorithm, are adopted an optimal arcfactored model large-margin learning techniques.Experiments show that our parsers achieve a competitive...
Graph representation learning aims to embed each vertex in a graph into low-dimensional vector space. Existing methods can be classified two categories: generative models that learn the underlying connectivity distribution graph, and discriminative predict probability of edge between pair vertices. In this paper, we propose GraphGAN, an innovative framework unifying above classes methods, which model play game-theoretical minimax game. Specifically, for given vertex, tries fit its true over...