Speakers
Dr. Xia Hu
Associate Professor at Rice University
ChatGPT in Action: An Experimental Investigation of Its Effectiveness in NLP Tasks
Dr. Yixin Cao
Assistant Professor at Singapore Management University
Trustworthy Natural Language Processing with Knowledge Guidance
Accepted Papers
Mind the Gap: An Investigation of Link Prediction Metrics on the Quality of Logic Rules
Harmanpreet Singh (LG Toronto AI Lab), Royal Sequiera (LG Toronto AI Lab), Sen Jia (LG Electronics), Maxime Gazeau (LG Toronto AI Lab), Homa Fashandi (LG Electronics)
Syntactic Complexity Measurement, and Reduction Through Controlled Simplification; An Approach to Extract Complete Knowledge from Unstructured Text for Knowledge Graphs
Muhammad Salman (The Australian National University)
College-Related Question Answering based on Knowledge Graph
Cheng Peng (The Hong Kong Polytechnic University), Hao Jiang (The Hong Kong Polytechnic University), Junnan Dong (The Hong Kong Polytechnic University), Xiao Huang (The Hong Kong Polytechnic University)
Error Detection on Knowledge Graphs with Triple Embedding
Yezi Liu (University of California Irvine), Qinggang Zhang (The Hong Kong Polytechnic University), Mengnan Du (New Jersey Institute of Technology), Xiao Huang (The Hong Kong Polytechnic University), Xia Hu (Rice University)
An Efficient Meta-path Fusion Network for Heterogeneous Recommendation
Shuai Chen (Beihang University), Hao Chen (Tencent), Liqun Yang (Beihang University), Zhoujun Li (Beihang University)
URL-BERT: Training Webpage Representation via Social Media Engagements
Ayesha Qamar (Texas A&M university), Ahmed El-Kishky (Twitter), Sumit Binnani (Twitter), Sneha Mehta (Independent), Taylor Berg-Kirkpatrick (UCSD), Chetan Verma (VerSe Innovation Labs)
About This Workshop
A knowledge graph (KG) consists of numerous triples, in which each triple, i.e., (head entity, relation, tail entity), denotes a real-world assertion. Many large-scale KGs have been developed, e.g., general-purpose KGs Freebase and YAGO. Also, lots of domainspecific KGs are emerging, e.g., COVID-19 KGs, biomedical KGs, and agricultural KGs. By embedding KGs into low-dimensional vectors, i.e., representations of entities and relations, we could integrate KGs into machine learning models and enhance the performance of many prediction tasks, including search, recommendations, and question answering. During the construction, refinement, embedding, and application of KGs, a variety of KG learning algorithms have been developed to handle challenges in various real-world scenarios. Moreover, graph neural networks have also brought new opportunities to KG learning.
This workshop aims to engage with active researchers from KG communities, recommendation communities, natural language processing communities, and other communities, and deliver state-of-the-art research insights into the core challenges in KG learning. Researchers, students, and practitioners who are working on or interested in knowledge graphs, recommender systems, question answering, and network analysis are welcome to attend the workshop. Researchers who are not familiar with KGs are also welcome to attend the workshop, in which they could learn how KGs can potentially be applied to and benefit their research.
Relevance to WSDM
WSDM is one of the premier conferences on web-inspired research involving search and data mining. Researchers, students, and practitioners in the area of recommendations, web mining, social network analysis, etc, will gather and communicate the latest research progress and research findings. Among them, many studies are related to KGs. This workshop matches the topic and scope of WSDM well.
Web mining and content analysis
Many KGs are constructed based on web content, especially encyclopedia KGs and commonsense KGs. E.g., Freebase, YAGO, and NELL.
KG-enhanced web search
To improve the quality of search results, many corporations have integrated KGs into their information retrieval processes, e.g., Google, Baidu, Microsoft Bing, and Yahoo!.
KG-based recommender systems
KGs provide auxiliary information of items, which could facilitate the recommendation models in learning more informative embedding representations. Also, KGs have been employed to alleviate the cold start issue in recommendations. Many e-commerce companies have established their own KGs to enhance their recommender systems. Taobao and Amazon have constructed KGs connecting items, scenes, users, and user profiles. These KGs could help the mining of actual needs of eCommerce customers.
KG-based virtual assistant systems
Internal KGs have been built to power many real-world virtual assistant services since triples in KGs could be employed to answer questions in natural languages.