• seminar

seminar

SEMINAR

Reasoning and Language Generation in the SUMO Ontology

Abstract: All popular ontology description languages like OWL, RDFS
and KIF don't have type systems. This means that it is hard to
guarantee that the information in the ontology is consistent. One of
the contributions of the thesis is that we explore how to model
ontologies in a strictly typed language like GF. The test case of the
work is SUMO - the biggest open source ontology available today. This
has several advantages: first inconsistencies are quickly discovered,
and second, since GF is both a logical framework and a powerful natural
language generation system, this ontology has a direct verbalization.
The original ontology has some NLG rules that are quite primitive, but
are used as the starting point for building the concrete syntax in GF for
English. Furthermore experiments were done in the generation of natural
language in Romanian and French. Since we don't have enough data for
complete verbalization, the experiments are just intended to show that
our approach is substantially better for these languages. From the
abstract syntax in GF it is possible also to generate axioms in TPTP
(the language used in the state of the art automated theorem provers)
which means that we can connect GF to theorem provers in order to check
the ontology for logical consistency.

 

Date: 2010-02-11 10:30 - 11:30

Location: EDIT Room, 3364, Hörsalsvägen, Chalmers

Permalink

SEMINAR

Michael McTear (Computer Science Research Institute, University of Ulster)

Spoken dialogue technology has developed considerably over the past thirty years both in terms of research activity as well as in terms of commercially deployed applications. In this talk I will begin with an overview of trends in dialogue research based on an analysis of papers that were presented at Eurospeech-Interspeech conferences in 1989, 1999, and 2009. Following this some challenges for dialogue researchers will be discussed, in particular, the issues faced by individual researchers and those in small groups who wish to build viable end-to-end dialogue systems and the problems involved in accessing the many resources and toolkits that are required for dialogue research. The talk will conclude with a brief overview of two recent PhD projects in dialogue at the University of Ulster and a discussion of some opportunities for future research involving the integration of dialogue technology into voice search applications and the application of spoken dialogue technology in ambient intelligence environments.

Date: 2010-05-20 10:15 - 12:00

Location: Department of Swedish Language, Lennart Torstenssonsgatan 8, room L307

Permalink

SEMINAR

Date: 2010-04-22 10:15 - 12:00

Location:

Permalink

SEMINAR

In this seminar, we will talk about our work on an integrated lexical resource for Swedish language technology. We call this resource the Swedish FrameNet++, the "++" part of the name reflecting the fact that it will include much more information than is normally found in a framenet.

Lars Borin
Dana Dannélls
Markus Forsberg
Maria Toporowska Gronostaj
Dimitrios Kokkinakis

Date: 2010-02-25 10:15

Location: L307

Permalink

SEMINAR

Meaning accommodation occurs when a dialogue participant learns new meanings by observing utterances in context and  adapting to the way linguistic expressions are used. This paper sketches a formal account of meaning accommodation, focusing on the detection of subjectively innovative uses of an expression. It also provides an example of adaptation of lexical meaning to accommodate innovation.

Date: 2010-06-03 10:15 - 12:00

Location: L307, Lennart Torstenssonsgatan 8

Permalink

SEMINAR

Josef Ruppenhofer (Saarland University).

In my presentation, I will highlight the differences in the work of
FrameNet and Salsa, the most important of which is whether one
proceeds through the lexicon frame-wise or lemma-wise. I will report
on an experiment we carried out in Saarbrücken to study the usefulness
of automatic pre-annotations for speeding up labeling.  I will also
talk a bit about the recent Semeval shared Task that my colleague
Caroline Sporleder and I organized on the topic of "Linking Events and
their Participants in Discourse". This involved full-text annotation
of extracts from Arthur Conan Doyle. These Sherlock Holmes-texts were
annotated with frames, coreference, and information about the
antecedents of null instantiated frame elements. Other recent work
that I've been involved with is a tool for reconfiguring FrameNet
semi-automatically, in particular in the context of textual entailment
recognition.

Date: 2010-05-06 10:15 - 12:00

Location: room L307, Lennart Torstenssonsgatan 8

Permalink

SEMINAR

Rising Frames

In this paper I will review arguments I have made before for using frame-like objects in a formal semantic account of Montague's puzzles about the rising of prices and temperatures.  I shall then look in a little detail at the view of semantic coordination that Staffan Larsson and I have been working on and suggest that the questions we should be asking are not "What is the meaning of rise?" but "Given certain resources how will an agent exploit her semantic resources for rise when exposed to a new utterance of the word?" and "How will this modify the agent's resources?".  I will suggest that such a view might give a simple explanation for Zipf's law concerning the relationship of meanings and frequency.

Slides.

Date: 2010-03-04 10:15

Location: room L307, Lennart Torstenssonsgatan 8

Permalink

SEMINAR

GRASP: Grammatikbaserad Språkinlärning (Grammar-based language learning)

I am developing a pedagogical tool for children with communication disabilities. The tool will consist of a computer system which gives support for language learning and language training. The system is graphical, where the child can point to words or phrases to get explanations in the form of symbols, written text, or spoken words. The child can also reformulate the sentences, by moving words to new places or change the inflection of words. The system will then automatically reformulate the sentence, so that it will always be grammatically correct. The tool is intended to stimulate - in a playful way - the child's will to explore the possibilities of language.

GRASP is a small one-year project which started in January 2010. It is run by DART (Kommunikations- och dataresurscenter för personer med funktionshinder), and is financed by Sunnerdahls Handikappfond.

In this talk I will report on the current status of the tool (less of this), and the underlying ideas (more of this). If you are lucky, I might even show some pseudo-mathematical theories about tree similarity in type-theoretical grammars...

Date: 2010-05-27 10:15 - 12:00

Location: L307, Lennart Torstenssonsgatan 8

Permalink

SEMINAR

 

I will present our ongoing work in the project CONPLISIT, where we investigate how semantic search methods can be used to detect changes in consumption patterns and life-style in 19th century Swedish literature. To reach this goal we are currently linking SALDO, a Swedish semantic lexicon, with Dalin, a Swedish 19th century dictionary, and moreover, adapting the morphological description of SALDO to Dalin. An interesting aspect of this work is that it adds a diachronic facet to the lexical resource SweFN++.The current state of the work is updated daily at: http://spraakbanken.gu.se/eng/research/swefn/dalin

 

Date: 2010-04-08 10:15 - 12:00

Location: Department of Swedish Language, Lennart Torstenssonsgatan 8, room L307

Permalink

SEMINAR

This seminar is about people with communication disabilities, and how language technology can be of help. More specifically, Mats Lundälv will present the EU project ÆGIS – Open Accessibility Everywhere: Groundwork, Infrastructure, Standards:

"The ÆGIS project seeks to determine whether 3rd generation access techniques will provide a more accessible, more exploitable and deeply embeddable approach in mainstream ICT (desktop, rich Internet and mobile applications)."

Mats is working at DART – Centre for Augmentative and Alternative Communication (AAC) and Assistive Technology (AT).

For more information:

 

 

Date: 2010-02-18 10:15 - 12:00

Location: L307

Permalink

X
Loading