To attack this problem, we propose a novel table cell search framework, in which we develop chain representations for both table cells and questions, and further employ deep neural networks to semantically match them. First, we tackle the challenge of modeling high-dimensional multi-modal correlations in the spatio-temporal data, as accurate modeling of correlations is the key to accurate predictive analysis.
Starting with the challenges and issues existing in knowledge graph query processing, I will discuss our efforts in addressing these issues, including schemaless graph querying, user feedback, factoid question benchmark, natural language questions, and query routing in collaborative networks.
Moreover, we further show that our model can assist optimizing collaborative networks via hypothesis testing. Towards realizing such great potential of question answering, we make two key observations on the status quo: The developed methodologies and frameworks in this dissertation pave the path for an array of exciting research directions including QA in various domains like healthcare and business intelligence, via using disparate and complementary data sources, and combining human intelligence and machine intelligence for problem solving and decision making through QA interactions.
As is known to all, the big data age contributes large-scale diversified information sources, such as structured knowledge bases KBsunstructured texts, and semi-structured tables.
Knowledge Graph and Question Answering Abstract: Human intelligence, contributed by crowdsourcing platforms and collaborative networks, should be exploited to complement automated question answering and problem solving. Monday, March 27, Date: Therefore, we investigate an important yet largely under-addressed problem: Second, what current automated QA systems can achieve is still limited in many situations, e.
In this thesis, we have studied the key challenges in large scale multivariate time series analysis and proposed novel and scalable solutions. This thesis contributes i scalable, principled algorithms that combine globality with locality to understand graphs, and ii applications in two areas: Analysis of large scale time series data collected from diverse applications has created new multi-faceted challenges and opportunities.
The paradigm of information search is undergoing a significant transformation due to the rise of mobile devices. Within this deluge of data, how can we find its most important structures?
We show how to interpretably summarize a graph with its important structures, and complement that with inference, which leverages little prior information and the network structure to efficiently learn information about all the entities.
We propose two solutions to address this challenge: Moreover, we empirically verify that tables supply rich knowledge that might not exist or is difficult to be identified in existing KBs. We cast the problem as a low-rank tensor learning problem with side information incorporated via a graph Laplacian regularization.Short Bio: Xifeng Yan is a professor at the University of California, Santa Barbara.
He holds the Venkatesh Narayanamurti Chair of Computer Science. He received his Ph.D. from the University of Illinois at Urbana-Champaign in and was a research staff member at the IBM T. J.
Watson Research Center between and This "Cited by" count includes citations to the following articles in Scholar. The ones marked * may be different from the article in the profile.
Xifeng Yan of University of California, Santa Barbara, CA UCSB with expertise in Information Science, Algorithms, Artificial Intelligence. Read. Outstanding Dissertation Award, Computer Science Department, UCSB Ph.D.
Progress Award, Computer Science Department, UCSB Alex Morales, Huan Sun, Xifeng Yan, \Synthetic Review Spamming and Immu-nization" The 22nd International Conference on World Wide Web (WWW)poster.
ACM SIGKDD dissertation awards recognize outstanding work done by graduate students in the areas of data science, machine learning and data mining. Review Criteria: Relevance of the Dissertation to KDD; Huan Sun (student) and Xifeng Yan (advisor) at University of California, Santa Barbara.
The primary goal of Professor Yan research is to develop fundamental concepts and new principles of data mining, design intelligent algorithms and build scalable systems.Download