Content: Third Workshop

Third Workshop


The third Hybris Workshop will take place on Nov 20th-21st 2013 at the technical faculty of the University of Dresden, Germany.

Venue

For detailed information, please see this document. 

Program

Tuesday, November 19th

Travel to Dresden.

Wednesday, Nov 20th

08:45Introduction
09:00-09:45Invited talk by Sophia Ananiadou, University of Manchester, UK.
Pathway curation using text mining (Abstract)
09:45-10:30Coffee Break
10:30-12:30Talks by project partners
B1Yue Ma, TU Dresden.
Learning Snomed GCIs from Textual Data: the first report
Alina Petrova, TU Dresden
Learning formal definitions for Medical Subject Headings
B2Maria Kissa, TU Dresden
Towards the automated curation of the Comparative Toxicogenomics Database using ASP


Max Ostrowski, University of Potsdam
Constraint Answer Set Programming for Biological Applications
12:30-14:00Lunch
14:00-15:00Invited Talk by Mutsunori Banbara, Kobe University, Japan
University Course Timetabling with Answer Set Programming (abstract)
15:00-15:30Coffee Break
15:30-17:30Talks by project partners
A2Stefan Ellmauthaler, University of Leipzig
Generalizing Multi-Context Systems for Reactive Stream Reasoning Applications

Orkunt Sabuncu, University of Potsdam
ROSoClingo: A ROS package for ASP-based robot control
A1
Jens Claßen, RWTH Aachen University
On the Decidability of Verifying LTL Properties of Golog Programs
C1Nichola Abdo, University of Freiburg
Inferring What to Imitate in Manipulation Actions by Using a Recommender System

Andreas Hertle,  University of Freiburg
Task Planning for Robots: Efficiency vs. Expressiveness

Tim Niemueller, RWTH Aachen University
Towards Complex Sensing in the Situation Calculus
20:00Dinner

Thursday, Nov 21st

09:00-12:00Tutorial by Wolfram Burgard, University of Freiburg
09:00-13:00PI Meeting
12:00Lunch and Farewell

Abstracts

Pathway curation using text mining

Invited talk by Sophia Ananiadou, University of Manchester

Abstract: The construction of large-scale pathway models requires substantial human effort, and their maintenance requires continuous monitoring of recent publications, to ensure that models reflect the current state of knowledge. Model construction requires complex evidence from the literature, since individual findings must be considered within their biological context -- the relevant cell types, chemical environment, etc. -- to ensure that they are interpreted correctly.  Locating and analysing such complex evidence requires techniques that go beyond basic keyword searching or the calculation of simple co-occurrence statistics between bio-entities and their interactions. However,
such simple co-occurrence approaches cannot deal with the potential complexities of reactions that involve multiple bio-entities, each of which may have a different role in the reaction (e.g., initiator, location, etc.). It is precisely complex interactions such as these that are targeted by event extraction.  The aim of event extraction is to recognise and structure information regarding interactions (e.g. transcription, apoptosis) and reactions in text, and to characterise the ways in which bio-entities participate in them. We have recently demonstrated the ability of event extraction techniques to support pathway curation which was designed to alleviate a frequent problem that is found in pathways, i.e., the lack of references to supporting evidence from the literature. The resulting system, PathText (http://nactem.ac.uk/pathtext2/), allows curators to find snippets from the literature that are relevant to a curator-selected reaction in a pathway network. To enable the manual inspection and correction of automatically produced results, we  have developed a web-based, text mining platform Argo (http://argo.nactem.ac.uk/) which allows users to combine a range of text mining solutions with the manual annotation of the resources produced.

University Course Timetabling with Answer Set Programming

Invited talk by Mutsunori Banbara, Kobe University, Japan

Abstract: In recent years, timetabling has become an area of increasing interest in an international community involving both researchers and practitioners, such as the international series of PATAT conferences. The typical topics of this area include educational timetabling, transport timetabling, employee timetabling, sports timetabling, and so on.

The educational timetabling problem can be generally defined as the task of assigning a number of events, such as lectures and examinations, to a limited set of timeslots (and perhaps rooms), subject to a given set of hard and soft constraints.

The modeling language and problem modeling play a very important role in the real-world timetable generation.  The latter is particularly challenging because different institutions have their own needs and policies, and problem formulation (a specific set of soft constraints) may change from institution to institution and from time to time.

In this talk, we show that ASP (Answer Set Programming) is an ideal modeling language for educational timetabling, through an ASP encoding of the curriculum-based course timetabling problem proposed in the third track of the second international timetabling competition (ITC-2007, http://www.cs.qub.ac.uk/itc2007/).

(Joint work with Takehide Soh, Naoyuki Tamura, Katsumi Inoue, and Torsten Schaub)