Models

Models of working memory

Computational models allow us to simulate working memory and better understand some of its features.

  • an extension of the TBRS* model (Oberauer & Lewandowsky, 2011) to deal with chunks (Portrat, Guida, Phénix, & Lemaire, 2016)
  • a study on the way computational models have to be adapted to fit older adult working memory data (Hoareau, Lemaire, Portrat, & Plancher, 2015)
  • a computational investigation of the refreshing mechanism within the TBRS model (Portrat & Lemaire, 2015)
  • a computational investigation of working memory capacity, suggesting to consider it as a quantity of information rather than a fixed number of items (Lemaire, Robinet, & Portrat, 2012).

Models of eye movements in information search

Searching information on a web page involves different kinds of cognitive processes and especially visual processes, semantic processes and obviously processes of memory management. Our goal is to design a computational cognitive model of the way humans combine visual, semantic and memory information in a web search  task.

  • a model of information search in textual material (Chanceaux, M., Guérin-Dugué, A., Lemaire, B., Baccino, T., 2014)
  • a model of the decision to stop or continue reading in information search (Lopez Orozco F.,Guérin-Dugué, A., Lemaire, B., 2012 ; Frey, A., Ionescu, G., Lemaire, B., Lopez-Orozco, F., Baccino, T., Guérin-Dugué, A., 2013).
  • a model for generating different visual strategies (Lemaire, B., Guérin-Dugué, A., Baccino, T., Chanceaux, M., Pasqualotti, L., 2011).
  • a model integrating visual, semantic and memory processes (Chanceaux, M., Guérin-Dugué, A., Lemaire, B., Baccino, T., 2009)

Models of inductive learning

The goal is to model the way humans inductively identify and aggregate concepts from the low-level stimuli they are exposed to. We use the idea that humans tend to select the simplest structures and implement it based on an information-theoretic criterion, the Minimum Description Length (MDL) principle.

  • MDLChunker, a model of chunking (Robinet, V., Lemaire, B., Gordon, M., 2011) that has been applied to working memory (Lemaire, B., Robinet, V., Portrat S., 2012) and word segmentation (Robinet, V., Lemaire, B., 2009)
  • a MDL-based Model of gender knowledge acquisition (Marchal, H., Lemaire, B., Bianco, M., Dessus, P, 2008)
  • a model of the acquisition of lexical gender in French (Marchal, H., Bianco, M., Dessus, P., Lemaire, B., 2007)

Models of semantic association

We mainly use Latent Semantic Analysis (LSA) to model the human jdgements of semantic association.

  • a model of children’s semantic memory (Denhière, G., Lemaire, B., Bellissens, C., Jhean-Larose,  S., 2007)
  • ICAN, a model of the incremental construction of an associative network from a corpus (Lemaire, B., Denhière, G., 2004)

Models of text comprehension and summarization

We use LSA as a model of semantic memory to model text comprehension and summarization.

  • a computational model for simulating text comprehension based on Kintsch’s Construction-Integration model (Lemaire, B., Denhière, G., Bellissens, C., Jhean-Larose, S., 2006)
  • a computational cognitive model of summarization assessment skills (Lemaire, B., Mandin, S., Dessus, P. & Denhière, G., 2005)
  • a model of metaphor comprehension (Lemaire B., Bianco M., 2003)

Models of text assessment

LSA can also be used to assess texts in an educational perspective

  • a model of students for educational environments (Zampa V., Lemaire B., 2002)
  • a model to assess the semantic content of student essays (Lemaire B., Dessus P., 2001 ; Dessus P, Lemaire B., 2002 ; Dessus, P., Lemaire, B., Loiseau, M., Mandin, S., Villiot-Leclercq, E. & Zampa, V., 2011)