Amir Dib

AI Scientist

dib.amir@gmail.com

I completed a PhD at ENS Paris-Saclay under the supervision of Nicolas VAYATIS and Mathilde MOUGEOT titled High dimensional pattern learning for symbolic time-series. My research focused on ML model for symbolic time series classification applied to anomaly detection. It covers related problem in statistical learning theory, bayesian inference and pattern extraction. The thesis manuscript and the PhD defense' slides are available online.

Currently helping build cutting-edge AI products at POLYGON AI at large scale for major industrial clients.

Academia

Ph.D Machine learning

2017 - 2021

ENS Paris-Saclay

M.Sc. Probability & Finance (DEA El Karoui)

2014 - 2015

Sorbonne University, Polytechnique

M.Sc. Fundamental Physics

2012 - 2015

Paris-Saclay University

Work Experience

Research Scientist

2023 - Present

POLYGON AI

Lead Research Scientist

2021 - end 2022

CITIO / RATP

Doctoral Candidate

2017 - 2021

ENS Paris-Saclay, French National Railway Company (SNCF)

Research Data Scientist.

2015 - 2017

French National Railway Company (SNCF)

Data Scientist.

2015 - 2015

Quantmetry

Recent Publications

Unified occupancy on a public transport network through combination of AFC and APC data, 2023 IEEE 26th International Conference on Intelligent Transportation Systems (ITSC).

Amir DIB , Noëlie Chierrier , Martin Graive , Baptiste Rérolle , Eglantine Schmitt

link

A Bayesian Markov Model for Station-Level Origin-Destination Matrix Reconstruction, European Conference on Machine Learning (ECML PKDD 2022)

Victor Amblard , Amir Dib , Noëlie Cherrier , Guillaume Barthe

pdf

High dimensional pattern-based learning for temporal time series, Thesis

Amir Dib

pdf slides

Epidemic Models for COVID-19 during the First Wave from February to May 2020: a Methodological Review, Preprint

Marie Garin , Limnios Myrto , Marie Garin , Bargiotas, Ioannis , Boulant Olivier , Chick Stephen , Dib Amir , Evgenious Theorodors , Fekom Mathilde , Kalogeratos Argyris

pdf

Localized pattern mining, To be submitted.

Amir Dib* , Cyrus Cousins*

Bayesian Feature Discovery for Predictive Maintenance, IEEE European Signal Processing Conference (EUSIPCO 2021)

Amir Dib , Charles Truong , Laurent Oudre , Mathilde Mougeot , Nicolas Vayatis , Heloise Nonne

code pdf slides

Quantized Variational Inference, Conference on Neural Information Processing Systems (NeurIPS 2020).

Amir DIB

code pdf poster slides

Projects

ONADAP

Le projet vise à la mise en place d’un outil d’aide à la décision pour les hôpitaux et autres unités sensibles soumis à l’attrition du personnel atteint par le COVID-19. L’objectif est de produire des rapports journaliers permettant à la fois de visualiser l’évolution de l’épidémie à l’échelle de l’unité, mais aussi de proposer des projections à quelques jours et enfin d’évaluer a priori les scénarios de redéploiement des personnels ou de réorganisation interne.

press website