• seminar

seminar

SEMINAR


Jeffrey Parrott (LANCHART, Copenhagen)

http://lanchart.hum.ku.dk/staff_scan/medarbejderdetaljer/?id=340729&f=1

 

Date: 2011-03-14 15:15 - 17:00

Location: L308, Lennart Torstenssonsgatan 8

Permalink

SEMINAR

Grounding the meaning of spatial words in a mobile robot setting

Simon Dobnik

Oxford University Computing Laboratory

 

We describe a system where the semantics of spatial referential
expressions have been automatically learned by finding mappings between
symbolic natural language descriptions of the environment and non-symbolic
representations from the sensory data of a mobile robot used for
localisation and map building (SLAM). Although the success of learning can
be measured by examining classifier performance on held-out data, this
does not in itself guarantee that the descriptions or actions generated
will be natural and informative for a human observer. To demonstrate that
the robot has captured human-like spatial knowledge the classifiers were
integrated with a system that drives a mobile robot and its performance in
generating descriptions and answering questions was evaluated by human
observers.

Date: 2011-03-18 14:15 - 15:15

Location: T116, Gamla Hovrätten (FLoV)

Permalink

SEMINAR

 


http://www.informatik.uni-konstanz.de/en/arbeitsgruppen/infovis/members/...

 

Date: 2011-02-23 13:30 - 15:30

Location: L307, Lennart Torstenssonsgatan 8

Permalink

SEMINAR


The existence of the GF resource grammar for Dutch and the well-known linguistic similarity between Dutch and Afrikaans provide a basis for the accelerated development of a GF resource grammar for Afrikaans. The main purpose of this talk is to provide a status report on recent progress made in this regard, with specific reference to the lexicon and certain word categories.

--

Professor Laurette Pretorius (University of South Africa) is at a the moment visiting the Language Technology Group at Computer Science and Engineering at GU/Chalmers.

Date: 2011-01-25 16:15 - 17:00

Location: EDIT room, EDIT building, Chalmers

Permalink

SEMINAR

Nancy Chang will give a guest talk at the Human reasoning seminar series

Abstract:

In this talk I present an overview of how certain relational concepts can be captured using neurally plausible representations. The approach exploits structured connectionist models that are, like traditional neural networks, inspired by the general properties of human neural computation. But these models also recognize the highly structured nature of neural circuitry, and the possibility of using local clusters of neurons to represent traditionally symbolic relational concepts. I will describe how several such concepts – including feature-value association, variable binding and dynamic action control – can be modeled by neurally plausible systems.

 

Date: 2011-06-09 10:15 - 12:00

Location: L308, Lennart Torstenssonsgatan 8

Permalink

SEMINAR

Date: 2011-05-26 10:15 - 12:00

Location: L308, Lennart Torstenssonsgatan 8

Permalink

SEMINAR

There are lots of reasons why setting up a CLT Cloud in the form of web-API:s to our language resources and processors is a good idea. In my talk, I will begin by reviewing some such reasons. The question is, how do we design and implement a cloud once we know we want one? A web-API request usually results in a chunk of XML or JSON being returned to the client, and this can of course easily be hacked. But XML and JSON is just syntax, and it seems we want to take a more pricipled and perhaps semantical approach to the API design problem. Markus Forsberg and myself have a project where we intend to explore such an approach. In most of my talk – or hey, let's call it a working seminar! – I intend to take a first stab at this, based on the concept of a graph.

Date: 2011-05-19 10:15 - 12:00

Location: L308, Lennart Torstenssonsgatan 8

Permalink

SEMINAR


Preparing for NODALIDA:

Markus Forsberg - Green resources in plain sight: opening up the SweFN++ project

SweFN++ is a project conducted at Språkbanken with two main objectives: the creation of a Swedish framenet covering at least 50,000 lexical units built on the same principles as the English Berkeley FrameNet; an integration of a number of existing free lexical resources, constructed by harmonizing, standardizing and merging these resources, and thereby reusing the valuable grammatical and semantic information painstakingly collected in these resources.

An important theme of the project is openness, which manifests itself in many different ways: openness from day one; open content; downloadable development versions updated daily; the use of open standards and open licenses; the use and production of open source tools; and open web service APIs for both resources and tools.

This is a philosophical stance – we believe that research should be carried out in the open to enable scrutinization and increased collaboration. It is, from our point of view, more valuable that anyone is allowed to download and inspect unfinished work today, and, at the same time, run the risk that it is confused with something more mature, rather than taking the safer, but less productive, road of publishing the "finished" product at the end of the project.

The project is open for inspection at: http://spraakbanken.gu.se/eng/swefn/

Peter Ljunglöf - Editing Syntax Trees on the Surface

We describe a system for interactive modification of syntax trees by intuitive editing operations on the surface string. The system has a graphical interface, where the user can move, replace, add, and in other ways modify, words or phrases. During editing, the sentence is kept grammatical, by automatically rearranging words and changing inflection, if necessary. This is accomplished by combining constraints on syntax trees with a distance measure between trees.

Dimitrios Kokkinakis - Evaluating the Coverage of three Controlled Health Vocabularies with Focus on Findings, Signs & Symptoms

The medical domain is blessed with a magnitude of terminological resources of various characteristics, sizes, structure, depth and breadth of descriptive power, granularity etc. In this domain a particularly interesting and difficult entity type are signs, symptoms and findings which to a large extend are expressed in a periphrastic manner, sometimes by the use of figurative or metaphorical language, or contextualized using a wealth of vague variant expressions. We hypothesize therefore that no major official terminology source alone can accommodate for the variation and complexity present in real text data, such as electronic medical records, notes or health related documents. In this paper we evaluate the content of the three largest medical control vocabularies available for Swedish on extracted reference symptom lists and initiate a discussion on how we should proceed in order to accommodate for increased coverage on similar genres.

 

Date: 2011-05-05 10:15 - 12:00

Location: L308, Lennart Torstenssonsgatan 8

Permalink

SEMINAR

Date: 2011-04-14 10:15 - 12:00

Location: L308, Lennart Torstenssonsgatan 8

Permalink

SEMINAR


The notion of "the content of a dialogue" is shown to be problematic in light of the phenomena of semantic coordination in dialogue, and the associated notion of semantic plasticity - the ability of meanings to change as a result of language use. Specifically, it appears that any notion of content in dialogue based on classical model-theoretical semantics will be insufficient for capturing semantic plasticity. An alternative formal semantics, type theory with records (TTR) is briefly introduced and is shown to be better equipped to deal with semantic coordination and plasticity. However, it is also argued that any account of content in dialogue which takes semantic coordination seriously will also need to consider the problems it raises for some concepts central to traditional notions of meaning, namely inference and truth.

 

Date: 2011-04-07 10:15 - 12:00

Location: L308, Lennart Torstenssonsgatan 8

Permalink

X
Loading