Luc Etienne Pommé

PhD in Computer Science

LaBRI, University of Bordeaux, France

About Me

Starting from 1st November 2020, I am a PhD Student at the University of Bordeaux in France, under the supervision of David Auber. I am working in colaboration with SUEZ, Le LyRE on the topic of ``Visualization for the Explanation of Deep Neural Networks''.

I am a member of the BKB (``Back to Knowledge and Beyond'') team at the LaBRI institute, and I usually work with Romain Bourqui and Romain Giot.

I obtained my Master Degree in Computer Sciences (Image and Sound processing) in 2020 at the University of Bordeaux.

I was awarded the Isabelle Attali prize in 2018.


Research

Keywords

Information Visualization, Deep Learning, XAI (eXplainable Artificial Intelligence), Explainable pruning, Computer Vision

Interests

The understanding of the world that surrounds us requires the observation and analysis of the phenomena that govern it. The observation of these phenomena requires the systematic collection of large quantities of data to build models representative of the reality. Eventually, the models are used to extract knowledge and properties about the object of study.

The more complex the data, the more difficult it becomes for a human to find processing rules or concepts to extract. This is where AI (e.g. machine / deep learning) algorithms come in to help. The superior performance of these algorithms being no longer in question, they are widely used in matters where decision making is required.

In addition to the worrying ecological footprint of these tools, the ethical aspect is also at stake. Indeed, these algorithms are often considered as ``black boxes'', making arbitrary decisions without justification. The question of responsability, in case of bad decision arises. It becomes necessary to build systems enabling to understand the behavior (either good or bad) of such algorithms in order to improve the trust we can put in their decisions. This problem has given rise to a field of research called XAI (eXplainable AI), which is currently in full expansion.

Within this research field, I am particularly interested in building new visualizations to highlight AIs behavior, and to try to provide justifications to their decisions. On the other hand, I am interested in tackling the over-parameterization problem of neural networks. Considering neural networks as directed graphs (where nodes are parameters/filters/neurons, and edges are input/output data), my work consists in studying the role of graph nodes and remove those that do not have a significant importance in the decisions. The induced sub-graphs represent sub-networks that achieve comparable performance on a task, with significantly less parameters, operations required to process input data, computation resources, and thus significantly less energy consumption.


Publications

2023

2022


Teaching

Master


Licence

IUT (Universitary Technical Institute)


Project supervision


Resources

ORCID [ 0000-0002-4690-7260 ]

HAL [ luc-etienne-pomme ]

Scholar [ Luc-Etienne POMMÉ ]


Contact Me