


M2 Apprentissage et Algorithmes (M2A)
Master Mathématiques et Applications / Master Informatique – Sorbonne Université
PhD positions at Imperial College London on privacy and impact of algorithms on society (fully funded, deadline: Nov 1, 2020)
Our Computational Privacy Group at Imperial College London is offering fully funded PhD positions for 2021 to study privacy, data protection, and the impact of algorithms on society.
Topics of current interests include, for instance, individual privacy in large-scale behavioral datasets; re-identification attacks against privacy-preserving data systems or aggregates, privacy of machine learning models, privacy engineering solutions such as differential privacy and query-based systems, ethics and fairness in AI, and computational social science.
For full details, please consult https://cpg.doc.ic.ac.uk/openings/
Deadline: Nov 1th 2020 (first deadline)
Recommended prerequisites. MSc or MEng (4y BEng will be considered) in computer science, statistics, mathematics, physics, electrical engineering, or a related field. Experience in data science, statistics and/or machine learning is a plus.
We encourage all qualified candidates to apply, in particular women, disabled, BAME, and LGBTQIA+ candidates.
About Imperial. Imperial College London, ranked 9th globally, is one of the top universities in the world. A full-time PhD at the South Kensington Campus takes 3-4 years, is fully funded and usually starts in October or January.
Stability and robustness of Deep Learning models to process video from thermal cameras
The PhD position is funded by Foxstream, a software company, founded in 2004, that specializes in real-time automated processing of video content analysis. The PhD thesis is a collaboration with Dauphine Université (the MILES team of the LAMSADE) with a join supervision (Quentin Barthélemy from Foxstream and Alexandre Allauzen from MILES).
For a couple of decades, Deep Learning (DL) added a huge boost to the already rapidly developing field of computer vision. While for some kind of data and tasks, DL is the most successful approach, this is not the case for all applications. For instance, the analysis of video streams generated by thermal cameras is still a research challenge because of the long range perimeter and the associated geometrical issues, along with the frequent calibration change. Therefore, the stability and robustness of DL models must be better characterized and improved. The goal of the PhD is to design a Deep architecture that can explicitely deal with these peculiarities, along with providing theoritical guarantees on the stability of the prediction and the underlying invariances.
The recent work of [1] proposes an interesting mathematical tool to charaterize the stability and the generalization capacity of convolutional network. This paper is important to better explain the lack of robustness of the DL models to some kind of examples like adversarial ones [2].
The PhD student will be host in Paris (France)in Dauphine Université and frequent meeting will be scheduled to ensure a tight collaboration with the team at Foxstream. The PhD can start in January 2021 and the position is open until it is filled.
Requirements:
– Outstanding master’s degree (or an equivalent university degree) in computer science or another related disciplines (as e.g. mathematics, information sciences, computer engineering, etc.).
– Proficiency in machine learning, computer vision, or signal processing. – Fluency in spoken and written English is required.
Application:
To apply, please email alexandre.allauzen [at] dauphine.psl.eu with:
– a curriculum vitae, with contact of 2 or more referees
– a cover letter
– a research outcome (e.g. master thesis and/or published papers) of the candidate
– a transcript of grades
[1] A. Bietti and J. Mairal, Group Invariance, Stability to Deformations,and Complexity of Deep Convolutional Representations, in JMLR 2019. http://www.jmlr.org/papers/volume20/18-190/18-190.pdf
[2] Szegedy et al, Intriguing properties of neural networks, https://arxiv.org/abs/1312.6199, 2013
https://www.foxstream.fr/
https://www.lamsade.dauphine.fr/wp/miles/