|
Keynote Talks
-
University of Massachusetts Amherst, USA
The Relevance of Answers [slides]
Abstract:
There has been an increasing interest in information retrieval tasks that involve retrieving answers rather than documents in response to users' questions. This has been driven primarily by the demands of limited bandwidth environments such as voice-based search or mobile search, but also by progress in the development of retrieval models that have improved performance on these tasks. Research in this area has led to a growing understanding of the differences between answers and relevant documents, but the implications of these differences are not, in my opinion, fully appreciated. In this talk, I will review the work that has been done on ranking and understanding answers with the aim of identifying some of the key aspects of answers that should be studied to support the ongoing development of more effective search systems.
Biography:
W. Bruce Croft is a Distinguished Professor Emeritus at the University of Massachusetts Amherst, Director of the Center for Intelligent Information Retrieval (CIIR), and Senior Research Fellow at RMIT University, Melbourne. He is an ACM Fellow and has received four lifetime achievement awards for his research, including the Gerard Salton Award from ACM SIGIR. Five of his papers have received SIGIR Test of Time Awards and another three received honorable mentions. His Google h-index is currently 103.
-
Ben-Gurion University, Israel
Automatic Analysis of Personality Dimensions through Digital Signatures:
Vision, Achievements and Challenges [slides]
Abstract:
The term “personality” is used to describe relatively stable patterns of thought, emotion and behavior. While the analysis of personality has been traditionally conducted through the expert’s “manual” analysis and personality questionnaires, it is currently possible to identify personality dimensions and disorders through the analysis of digital signatures produced by individuals in the social media. The emerging technologies for computational personality analysis have clear practical implications from the screening of depression, to the design of targeted advertisements suited for the individual’s personality and up to the screening for solo perpetrators such as school shooters. In this talk, I will present the vision of computational personality analysis and its relevance for current challenges in various fields. I will explain why the computational analysis of personality must rely on more sophisticated theories of personality than those used so far and will illustrate some notable achievements and directions from the automatic analysis of depression to the screening for solo terrorists. The talk will also include a critical examination of some ventures, such as those developed by IBM Personality Insights and the late Cambridge Analytica, and will point to the challenges facing those who are interested in advancing the field.
Biography:
YAIR NEUMAN (b. 1968) received his PhD.in cognition from the Hebrew University and he is
currently a full Prof. at the Department of Cognitive and Brain Sciences, BGU, Israel. He is interested in developing novel computational models for various interdisciplinary problems of data analytics and has developed cutting edge algorithms for various contexts and agencies ranging from IARPA to CitiBank. Has made intensive contributions to the automatic analysis of personality dimensions in academic and non-academic comtexts. The author of numerous academic papers and six books, the most recent is "Mathematical structures of natural intelligence" (Springer).
-
Signal AI
Analysing the world's news: Learnings and challenges from industry [slides]
Abstract:
The talk will focus on the learnings from the last 6 years of applying research at Signal AI. It will showcase the efforts to transform the best research in IR/NLP to build a large-scale text analytics pipeline capable of processing millions of documents daily to power the Signal AI media monitoring and intelligence platform and a review of the open challenges that could be interesting for the CLEF community. The talk will touch not only technical details of the pipeline and its components but also the challenges and differences of working and evaluating different solutions in industry.
Biography:
Dr. Miguel Martinez is the Co-Founder and Chief Data Scientist of Signal AI, a growing UK company that analyses millions of news articles every day in real-time in order to improve business intelligence and decision making. During the last 6 years, Miguel has led the efforts to create and maintain the Signal Research team with strong university collaborations, with the main goal of transforming the best research principles, models and algorithms from academia into real, scalable products. Research interests include news processing, information retrieval, text analysis, natural language processing and evaluation, among others. He completed his PhD from Queen Mary University in 2014. During his time at Signal AI, the company has grown from 3 to more than 150 people, and raised more than $30M. He has been awarded the Business Leader of Tomorrow award 2014 by Innovate UK and was included in the list of UK Business Innovators in 2016 by Bloomberg. Signal AI has won numerous awards, including the Fujitsu AI Innovator in the Lloyds National Business awards 2018 and the Hottest Enterprise SaaS or B2B in the Europas 2018.
|
|
