Thibault Maho

Contact Publications News Home


I am a PhD student affiliated with INRIA Rennes, where I am part of the Linkmedia team led by Laurent Amsaleg. Under the supervision of Teddy FURON and Erwan LE MERRER, my research revolves around enhancing the security of neural networks. My work involves both crafting adversarial attacks to deceive AI systems and devising robust defenses to fortify them.

I hold an engineering degree from Phelma, which is a proud member of Grenoble-INP. My educational background has equipped me with proficiencies in signal and image processing as well as electronics. Prior to my doctoral journey, I spent three years as an R&D engineer at xBrain. During my tenure, I delved into the realm of deep learning and natural language processing, contributing to diverse projects encompassing text classification, part-of-speech tagging, and knowledge discovery. I had the privilege of shepherding solutions from their initial testing phases to production deployment.

This experience increased my desire for understanding AI at a deeper level, culminating in the commencement of my PhD research titled Input Space Exploration for Enhancing the Security of Neural Network Models in September 2020. For a glimpse into my academic contributions, please explore my works here.

I am poised to defend my thesis later this year and actively seeking opportunities for a Postdoctoral Position. If you have any openings related to ML security, Trustworthiness, or Privacy, please don't hesitate to get in touch.


07 / 2023 : Acceptance to ICCV 2023 for How to choose your best allies for a transferable attack? cowritten with Seyed Mohsen Moosavi-Dezfooli during my visit at Imperial College London
07 / 2023 : Acceptance to T-IFS for Fingerprinting Classifiers with Benign Inputs
10 / 2022 : Starting 5 months as a visiting research with Seyed Mohsen Moosavi-Dezfooli at Imperial College London
09 / 2022 : Presentation of the paper Randomized Smoothing Under Attack: How Good is it in Practice? at the GRETSI conference
07 / 2022 : Participation to the 16th summer school of Peyresq organized by the GRETSI
04 / 2022 : Acceptance to IEEE ICASSP 2022 for Randomized Smoothing Under Attack: How Good is it in Practice?
05 / 2021 : Acceptance to IEEE ICIP 2021 for RoBIC: A benchmark suite for assessing classifiers robustness
03 / 2021 : Acceptance to CVPR 2021 for SurFree: a fast surrogate-free black-box attack
09 / 2020 : Beginning of my PhD: Input Space Exploration for the Security of Neural Network Models


Get In Touch