CDD – Post doc (12-24 mois) – GRADIVA

France Life Imaging; Grenoble; Sur place; EUR 30 000 - 40 000; Plein temps

Résumé du poste

A research organization in France is offering a post-doc position focusing on the weaknesses of deep neural networks, particularly in medical applications. Candidates with a PhD in Neurosciences or related fields, and experience in machine learning are encouraged to apply. This role will involve collaboration with experts in Neuroimaging and Statistical methods to develop resilient network architectures.

Qualifications

• PhD in Neurosciences, Neuroimaging, or related field.
• Experience with machine learning and deep neural networks.
• Familiarity with mathematical properties of graph neural networks.

Responsabilités

• Explore weaknesses of deep neural networks in medical applications.
• Represent DNN as a graph for tracking learning and predictions.
• Implement solutions for resilience against forgetting and attacks.

Description du poste

Contexte et mission

Environment: We offer a stimulating research environment gathering experts in Neurosciences & Neuroimaging and experts in Advanced Statistical and Machine Learning methods. The post-doc position will be available in the context of the Grenoble 3AI project (chair neuromorphometrics @MIAI https://miai.univ-grenoble-alpes.fr/). The postdoctoral fellow will work in close collaboration with a PhD student working on Graphs as model for brain network studies and with a Cea-List team which has developed a bioinspired architecture which could offer interesting resilient properties.

Starting date: Autumn 2021

How to apply: Send an email directly to the supervisors with your CV. Applications will be accepted up to the 31st of August. The final decision will be given by the beginning of October.

Objectifs et responsabilités

• Use mathematical properties of classical graph neural networks in order to explore unresolved specific weaknesses of deep neural networks (DNN). We will focus on catastrophic forgetting (or catastrophic interference) and adversarial attack.
• The objective is to represent the DNN as a graph and track learning and prediction under different conditions of training and attack, to study these major DNN drawbacks notably for medical application.
• The final goals are to respond to several questions: are specific hidden neurons (or layers) vulnerable to forgetting or attack? Which solutions can be implemented (introduction of penalty during the training phase, specific architectures including feedback connections, …) to design DNN more resilient to forgetting and attack?

Mots-clés

Key words: Machine learning; Multidimensional data; Neural Network