Safe and trustworthy Artificial Intelligence (AI) is central in the deployment of any AI system in major application areas, such as medicine and autonomous vehicles. Its major keystone requirements in Machine Learning (ML) have been recently investigated by researchers of the ML group in the LITIS Lab, including robustness, explainability and fairness. The current internship aims to address anomaly detection with explainable models/results, which is a major ingredient of robust ML for Safe and trustworthy AI.

 

The broad interest in deep neural networks has driven recent advances in anomaly detection, also called out-of-distribution or novelty detection. Deep anomaly detection methods fall within three major categories: Deep one-class, variational autoencoders (VAEs) and generative adversarial networks (GANs) [1, 2]. While VAEs and GANs do not allow an exact evaluation of the probability density of new samples, they also suffer from notorious training instability (mode collapse, posterior collapse, vanishing gradients and non-convergence), as corroborated by many research studies [3]. For these reasons, we will investigate Normalizing Flows (NF), an emerging class of generative models where both sampling and density evaluation are efficient and exact, and where the latent representation is learned through an invertible transformation [4]. NF provide explainable models, are interconnected with Optimal Transport and have solid foundations for probabilistic modeling and statistical inference [5].

 

The goal of this internship is to explore Normalizing Flows for anomaly detection on time series. While NF have been previously explored with success for anomaly detection in images, they were seldom investigated for time series. The tasks to be carried out by the intern are as follows: The intern will first study relevant work on NF for anomaly detection, and then revisit them in the light of time series. She/he will explore two contexts: detection from a batch of time series data, and online detection on streaming data. For the latter, a particular attention will be paid to sequential detection. The intern will implement the different NF-based models and conduct experiments on real time series.

 

This internship may lead to a PhD thesis.

 

The intern will conduct her/his research within the Machine Learning group in the LITIS Lab, under the supervision of Prof. Paul Honeine. This internship is within a research project gathering 9 permanent researchers of the LITIS Lab and the intern will also interact with several PhD students and interns also working on deep anomaly detection for time series.

 

Applicant Profile:

-Student in final year of Master or Engineering School, in data science, artificial intelligence, applied mathematics, or related fields.

-Strong skills in advanced statistics and Machine Learning, including Deep Learning

-Good programming experience in Python

 

Location: LITIS Lab, University of Rouen Normandy, Saint Etienne du Rouvray (Rouen, France).

 

Terms: 5 to 6 months, starting in February or March 2024.

 

Application: Applicants are invited to send their CV and grade transcripts by email to paul.honeine@univ-rouen.fr.

 

[1] L. Ruff, J. R. Kauffmann, R. A. Vandermeulen, G. Montavon, W. Samek, M. Kloft, T. G. Dietterich, and K.-R. Müller, “A Unifying Review of Deep and Shallow Anomaly Detection,” Proceedings of the IEEE, vol. 109, no. 5, pp. 756–795, 2021.

[2] G. Pang, C. Shen, L. Cao, and A. V. D. Hengel, “Deep learning for anomaly detection: A review,” ACM Computing Surveys, vol. 54, no. 2, pp. 1–38, 2021.

[3] D. Saxena and J. Cao, “Generative adversarial networks (GANs) challenges, solutions, and future directions,” ACM Computing Surveys, vol. 54, no. 3, pp. 1–42, 2021.

[4] I. Kobyzev, S. J. Prince, and M. A. Brubaker, “Normalizing Flows: An Introduction and Review of Current Methods,” IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 43, no. 11, pp. 3964–3979, 2021.

[5] G. Papamakarios, E. Nalisnick, D. J. Rezende, S. Mohamed, and B. Lakshminarayanan, “Normalizing Flows for Probabilistic Modeling and Inference,” Journal of Machine Learning Research, vol. 22, no. 57, pp. 1–64, 2021.