Tahiti Meeting
Tahiti Meeting is the annual “conference” for DREAM team in IRISA. The name comes from the fact that they always choose an island in Bretagne to hold the meeting. This year the meeting was held on L’ile d’Yeu which is in Vendee, the province south to Bretagne. We were 17 people, including all the profs in DREAM team, ph.d. students and 4 invited attendants from France Telcom, and other universities.
Day One’s presentations were on model-based diagnosis using discrete-event system. The current challenges are incremental diagnosis, diagnosability analysis. Alban Gretain who was just from AI05 presented his framework of incremental diagnosis and its usage in other analysis, such as system reconfiguration. Yannick Penncole, who is a French Australian now, presented his work of using diagnosability analysis to construct diagnoser. This kind of method works for discrete-event system only. Because DIAMOND project is launching in this team, Mao presented a new paper from Luca’s team on using MBD in Web Services diagnosis. This paper is a first step in this topic. I presented a short tutorial in Web Services and our work for OISEE project . The paper was accepted in ICWS05.
Day Two’s presentations were a mixture of data mining and diagnosis. I got to know they have a project on cardiology diagnosis. They collaborate with a prof who is specialized in cardiology single. DREAM team uses symbolic reasoning approach to help choosing signal processing algorithms and diagnosing the diseases. I will use their knowledge to try our algorithms developed for monotonicity analysis. This would be interesting work.
Day Two’s rest presentations and one presentation in Day Three were about logic research. One was on disjunctive logic programming by Yves Moinard . Though this French presentation was difficulty to fully understand, I saw how the research is done. Since disjunctive logic cannot be processed using …, it is a hard topic. Another presentation by Philippe Besnard was about causal logic reasoning. Since causality is important for diagnosis, I once also thought about the logic behind it. The presentation of course was a systematic analysis on this topic. I felt happy after I saw this work.
Day One’s presentations were on model-based diagnosis using discrete-event system. The current challenges are incremental diagnosis, diagnosability analysis. Alban Gretain who was just from AI05 presented his framework of incremental diagnosis and its usage in other analysis, such as system reconfiguration. Yannick Penncole, who is a French Australian now, presented his work of using diagnosability analysis to construct diagnoser. This kind of method works for discrete-event system only. Because DIAMOND project is launching in this team, Mao presented a new paper from Luca’s team on using MBD in Web Services diagnosis. This paper is a first step in this topic. I presented a short tutorial in Web Services and our work for OISEE project . The paper was accepted in ICWS05.
Day Two’s presentations were a mixture of data mining and diagnosis. I got to know they have a project on cardiology diagnosis. They collaborate with a prof who is specialized in cardiology single. DREAM team uses symbolic reasoning approach to help choosing signal processing algorithms and diagnosing the diseases. I will use their knowledge to try our algorithms developed for monotonicity analysis. This would be interesting work.
Day Two’s rest presentations and one presentation in Day Three were about logic research. One was on disjunctive logic programming by Yves Moinard . Though this French presentation was difficulty to fully understand, I saw how the research is done. Since disjunctive logic cannot be processed using …, it is a hard topic. Another presentation by Philippe Besnard was about causal logic reasoning. Since causality is important for diagnosis, I once also thought about the logic behind it. The presentation of course was a systematic analysis on this topic. I felt happy after I saw this work.
0 Comments:
Post a Comment
<< Home