UMR CNRS 7253

Site Tools

research

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Next revision		Previous revision
research [2015/10/19 13:48] – external edit 127.0.0.1research [2021/01/05 10:31] (current) – [Learning problems] sdesterc
Line 5: Line 5:
 ====== Practical uncertainty representations ====== ====== Practical uncertainty representations ======
  
-Work (mainly) benefiting from collaborations and discussions with D. Dubois, M. Troffaes, E. Miranda, L. Utkin, E. Chojnacki, E. Quaeghebeur and I. Sanchez+Work (mainly) benefiting from collaborations and discussions with D. Dubois, M. Troffaes, E. Miranda, L. Utkin, E. Chojnacki, E. Quaeghebeur and I. Montes
  
 There exist many practical representations in imprecise probability theories, including possibility distributions, belief functions, imprecise probability assignments, pari-mutuel models, imprecise cumulative distributions (p-boxes), clouds, ... There exist many practical representations in imprecise probability theories, including possibility distributions, belief functions, imprecise probability assignments, pari-mutuel models, imprecise cumulative distributions (p-boxes), clouds, ...
Line 27: Line 27:
 ====== Uncertainty propagation and (in)dependence modelling ====== ====== Uncertainty propagation and (in)dependence modelling ======
  
-Work (mainly) benefiting from collaborations and discussions with D. Dubois, G. De Cooman, E. Chojnacki, J. Baccou, T. Burger, M. Sallak, M.C.M. Troffaes, F. Coolen, S. Ferson, F. Aguirre and I. Sanchez+Work (mainly) benefiting from collaborations and discussions with D. Dubois, G. De Cooman, E. Chojnacki, J. Baccou, T. Burger, M. Sallak, M.C.M. Troffaes, F. Coolen, S. Ferson, F. Aguirre and I. Montes
  
 How to propagate uncertainty analysis in various models is an important issue that may face several difficulties. Most of my research in this domain has concerned the propagation of uncertainty model through deterministic functions with methods combining Monte-Carlo simulation and interval analysis, with an industrial risk-assessment purpose.  How to propagate uncertainty analysis in various models is an important issue that may face several difficulties. Most of my research in this domain has concerned the propagation of uncertainty model through deterministic functions with methods combining Monte-Carlo simulation and interval analysis, with an industrial risk-assessment purpose. 
Line 39: Line 39:
 ====== Learning problems ====== ====== Learning problems ======
  
-Work (mainly) benefiting from collaborations and discussions with B. Quost, T. Denoeux, B. Ben Yaghlane, N. Sutton-Charani, G. Yang, M. Masson, E. Hüllermeier, A. Antonucci, G. Corani, M. Poss and N. Ben Abdallah+Work (mainly) benefiting from collaborations and discussions with B. Quost, T. Denoeux, B. Ben Yaghlane, N. Sutton-Charani, G. Yang, M. Masson, E. Hüllermeier, A. Antonucci, G. Corani, M. Poss, V-L. Nguyen, Y. Alarcon, S. Messoudi and N. Ben Abdallah
  
 Outside of extending some classical classifiers (k-NN methods, Naïve networks) to imprecise probabilistic settings, our work currently focuses on the combination of classifiers, to address both the usual multi-classification problem, as well as more complex problems such as label ranking and multilabel classification. Outside of extending some classical classifiers (k-NN methods, Naïve networks) to imprecise probabilistic settings, our work currently focuses on the combination of classifiers, to address both the usual multi-classification problem, as well as more complex problems such as label ranking and multilabel classification.
  
-One of our current favorite field of investigation is the so-called binary decomposition, where complex problems are decomposed in several binary ones (facilitating the learning but increasing the number of models to learn). In the future, we plan to focus more on active learning topics, where imprecise probabilistic methods can have an important role to play, due to their ability to identify cases where information is missing. +Our research currently focus on three issues: 
 +  * Learning from imprecise/soft data, where we propose methods to learn from uncertain data in a general way; 
 +  * Robust/skeptical inference for complex problems, where we try to produce cautious models, that is models delivering set-valued rather than point-valued predictions, for complex problems typically presenting a combinatorial structures. This is in particular multi-task learning problems such as multi-target regression, multi-label issues or learning-to-rank problems; 
 +  * Instrumentalizing imprecision in learning, where we try to identify those learning scenarios where using imprecision can actually improve the results when compared to more traditional (e.g., probabilistic) methods.
  
  
Line 53: Line 56:
  
   * Flexible querying in data bases (P. Buche, V. Guillard)   * Flexible querying in data bases (P. Buche, V. Guillard)
-  * Signal filtering with kernels (O. Strauss, F. Comby)+  * Signal filtering with kernels (O. Strauss, F. Comby, A. Rico)
   * Knowledge Engineering (B. Charnomordic, R. Thomopoulos)   * Knowledge Engineering (B. Charnomordic, R. Thomopoulos)
   * Risk analysis and robust design (E. Chojancki, V. Guillard, M. Sallak)   * Risk analysis and robust design (E. Chojancki, V. Guillard, M. Sallak)
   * Process modelling (C. Baudrit)   * Process modelling (C. Baudrit)
   * Virtual training (I. Thouvenin)   * Virtual training (I. Thouvenin)

Page Tools

User Tools

research.1445255293.txt.gz · Last modified: 2015/10/19 13:48 by 127.0.0.1
Mentions légales