Accueil > Stages/Emploi > Post-doc


publié le

A post-doctoral position is openin in October 2019, as part of the ANR funded project Chunked, with Benoit Lemaire and Sophie Portrat.

starting date : Octobre 2019 (adaptable) - duration : 18 months

Contact :

Post-doc : Computational cognitive modeling of chunking processes in working memory

Working memory is crucial for everyday life because it maintains information for the purpose of ongoing cognitive activities. It has been widely studied, in particular by means of computational models that are useful to investigate complex models whose outcomes is difficult to predict. The verbal Time-Based Resource Sharing model of working memory (Barrouillet, Portrat, & Camos, 2011) is such a complex system. It describes how specific operations such as encoding, forgetting, attentional refreshing and retrieval interact to allow memorization. Several computational models have implemented these processes (Oberauer & Lewandowsky, 2011 ; Gauvrit & Mathy, 2018 ; Glavan & Houpt, 2018). In our lab, computational models were designed to specify some characteristics of attentional refreshing (Portrat & Lemaire, 2015) and the way time-based decay and interference may jointly operate as sources of forgetting (Lemaire & Portrat, 2018).

For a more efficient memory storage, people are used to group information into meaningful units called chunks (the sequence X V P D F T is easy to memorize as soon as one recognizes the chunk PDF). A first proposal implementing chunking processes within a working memory model has been proposed in our lab (Portrat et al., 2016) with chunks being groups of 3 letters in a sequence of 7 letters. The goal of this post-doctoral project is to extend this framework because many factors, that are not considered yet, are known to influence chunking (e.g., prior knowledge, perceptual characteristics). The candidate should be able to understand the existing computational model, extend it considering the literature and other models and compare it to experimental data. Experiments should also be designed to collect data that would be necessary to test the model.

Candidates should have a PhD in one of the domains of cognitive sciences (cognitive psychology, artificial intelligence, neurosciences, computer science) and a background in both computer programming and cognitive theories.

read more on the Chunked project :

Voir en ligne : Chunked Project