You are here
DH Seminar: "Algorithms for Temporal Information Processing"
Speaker: Oleksandr Kolomiyets, Katholieke Universiteit Leuven
Title: "Algorithms for Temporal Information Processing of Text and their Applications"
Temporal information processing of text is a complex information extraction task in which temporally relevant information in text has to be extracted and properly represented in order to be used by a machine. In general the temporal information processing task regards the major concepts of temporal cognition such as time, events, and relations between events and times when they are encoded in language.
In this talk I will highlight the algorithms for temporal information processing of text and focus on the automated extraction of temporal information. Three major temporal concepts in language are identified: time expressions - expressions in text that denote time, temporal events - events that happen or last in time, and temporal relations between events and times. With respect to this distinction temporal information processing of text can be divided into a number of corresponding sub-tasks, such as recognition and normalization of time expressions, recognition of events, and recognition of temporal relations between events and times.
In detail, the talk focuses on (i) the supervised learning algorithms for temporal expression recognition in text with sparse training data and a bootstrapping approach that addresses the sparsity problem, (ii) the novel paradigm for modularized normalization of temporal expressions based on a deep semantic analysis of temporal expression constituents, (iii) the novel annotation paradigm of temporal information that aims at a full and coherent set of annotated temporal relations, but does not require annotating an exhaustive set of temporal relations, and (iv) the novel algorithms for the temporal document structure recognition composed of temporal events and temporal relations which is an important step towards automated story understanding.