[ Skip to the content ]

Institute of Formal and Applied Linguistics Wiki


[ Back to the navigation ]

11. 11. 2010 (MFF UK, Malostranske nam. 25, 1. patro, mistnost S9)
IDO DAGAN (Bar-Ilan University, Israel)
PROSPECTS FOR APPLIED SEMANTICS: A TEXTUAL ENTAILMENT PERSPECTIVE

Abstract:
How should an applied model for natural language semantics look like?
While the desired functionality and output of models for other levels of
language, like morphology and syntax, are quite consensual, this is not
the case for semantics. Classical formal approaches suggest that semantic
models should produce logic representations of text. Yet, the formal
logic-based approach remained a rather small niche in state of the art
NLP. Common practice, on the other hand, is rather chaotic, with a
plethora of scattered semantic processing tasks whose relationships are
largely unclear.
This talk will argue that the textual entailment paradigm may provide a
suitable encompassing framework for much of the applied semantics space.
Under this approach the underlying semantic modeling task should be
mapping between natural language units whose meanings entail one another,
rather than mapping language units onto an extrinsic logical language.
This approach is motivated by the observation that many semantic inference
needs across NLP can be reduced to the entailment framework, which, in
turn, may encompass much of the common-practice techniques in applied
semantics.
The first part of the talk will motivate and present the entailment
framework. We will then review BIUTEE, the Bar-Ilan University Textual
Entailment Engine, illustrating practical entailment modeling and
interesting research directions, including a proof system over parse-trees
and global learning of entailment graphs. Finally, I will suggest some
vision along two lines: (a) creating generic entailment engines that may
power semantic processing across NLP tasks and applications; (b) teasing
ideas, triggered by the textual entailment paradigm, for potential leaps
in long-awaited application areas, including text exploration, intelligent
tutoring and natural language interfaces.


[ Back to the navigation ] [ Back to the content ]