An international team based in Paris, which conducts AI research for Valeo automotive applications, in collaboraton with world-class academics. Our main research is towards better, clearer & safer automotive AI. See our papers, projects, codes, posts and tweets.
Hedi Ben-younes, research scientist | scholar | twitter
Florent Bartoccioni, PhD student
Alexandre Boulch, research scientist | page | scholar| github | twitter
Andrei Bursuc, research scientist | page | scholar | twitter
Laura Calem, PhD student | page | twitter
Charles Corbière, PhD student | page | scholar | twitter
Matthieu Cord, principal scientist | page | scholar | twitter
Spyros Gidaris, research scientist | scholar
David Hurych, research scientist | scholar
Himalaya Jain, research scientist | page | scholar
Renaud Marlet, principal scientist | page | scholar
Arthur Ouaknine, PhD student | twitter
Patrick Pérez, scientific director | page | scholar
Gilles Puy, research scientist | page | scholar
Julien Rebut, research scientist
Simon Roburin, PhD student | page
Antoine Saporta, PhD student | scholar
Tristan Schultz, research engineer
Marin Toromanoff, PhD student | scholar
Huy Van Vo, PhD student | scholar | github
Tuan-Hung Vu, research scientist | page | scholar | twitter
Emilie Wirbel, research scientist | scholar
Eloi Zablocki, research scientist | scholar | twitter
Multi-sensor perception — Automated driving relies first on a variety of sensors, like Valeo’s fish-eye cameras, LiDARs, radars and ultrasonics. Exploiting at best the outputs of each of these sensors at any instant is fundamental to understand the complex environment of the vehicle and gain robustness. To this end, we explore various machine learning approaches where sensors are considered either in isolation (as radar in Carrada at ICPR’20) or collectively (as in xMUDA at CVPR’20).
3D perception — Each sensor delivers information about the 3D world around the vehicle. Making sense of this information in terms of drivable space and important objects (road users, curb, obstacles, street furnitures) in 3D is required for the driving system to plan and act in the safest and most confortable way. This encompasses several challenging tasks, in particular detection and segmentation of objects in point coulds as in FKAConv at ACCV’20.
Frugal learning — Collecting diverse enough data, and annotating it precisely, is complex, costly and time-consuming. To reduce dramatically these needs, we explore various alternatives to fully-supervised learning, e.g, training that is unsupervised (as rOSD at ECCCV’20), self-supervised (as BoWNet at CVPR’20), semi-supervised, active, zero-shot (as ZS3 at NeurIPS’19) or few-shot. We also investigate training with fully-synthetic data (in combination with unsupervised domain adaptation) and with GAN-augmented data.
Domain adaptation — Deep learning and reinforcement learning are key technologies for autonomous driving. One of the challenges they face is to adapt to conditions which differ from those met during training. To improve systems’ performance in such situations, we explore so-called “domain adaptation” techniques, as in AdvEnt at CVPR’19 and DADA its extension at ICCV’19.
Reliability — When the unexpected happens, when the weather badly degrades, when a sensor gets blocked, the embarked perception system should continue working or, at least, diagnose the situation to react accordingly, e.g., by calling an alternative system or the human driver. With this in mind, we investigate ways to improve the robustness of neural nets to input variations, including to adversarial attacks, and to predict automatically the performance and the confidence of their predictions as in ConfidNet at NeurIPS’19.
Driving in action — Getting from sensory inputs to car control goes either through a modular stack (perception > localization > forecast > planning > actuation) or, more radically, through a single end-to-end model. We work on both strategies, more specificaly on action forecasting, automatic interpretation of decisions taken by a driving system, and reinforcement / imitation learning for end-to-end systems (as in RL work at CVPR’20).
Core Deep Learning — Deep learning being now a key component of AD systems, it is important to get a better understanding of its inner workings, in particular the link between the specifics of the learning optimization and the key properities (performance, regularity, robustness, generalization) of the trained models. Among other things, we investigate the impact of popular batch normalization on standard learning procedures and the ability to learn through unsupervised distillation.
- ESL: Entropy-guided Self-supervised Learning for Domain Adaptation in Semantic Segmentation (CVPRw’20)
- FLOT: Scene flow on point clouds guided by optimal transport (ECCV’20)
- AdamSRT: Adam exploiting BN-induced pherical invariance of CNN (arXiv 2020)
- LightConvPoint: Convolution for points (ACCV’20)
- xMUDA: Cross-modal UDA for 3D semantic segmentation (CVPR’20)
- LearningByCheating: End-to-End driving using implicit affordances (CVPR’20)
- rOSD: Unsupervised object discovery at scale (ECCV’20)
- ConvPoint Convolutions for unstructured point clouds (Computer \& Graphics 2020)
- ZS3: Zero-Shot Semantic Segmentation (NeurIPS’19)
- BF3S: Boosting few-shot visual learning with self-supervision (ICCV’19)
- ConfidNet: Addressing failure prediction by learning model confidence (NeurIPS’19)
- Rainbow-IQN Ape-X: effective RL combination for Atari games
- DADA: Depth-aware Domain Adaptation in Semantic Segmentation (ICCV’19)
- AdvEnt: Adversarial Entropy minimization for domain adaptation in semantic segmentation (CVPR’19)
- OSD: Unsupervised object discovery as optimization (CVPR’19)
CNAM (Nicolas Thome)
CTU Prague (Josef Sivic)
EPFL (Alexandre Alahi)
INRIA (Jean Ponce, Karteek Alahari)
MPI (Christian Theobalt)
Ponts (Mathieu Aubry)
Sorbonne (Matthieu Cord)
Télécom (Florence Tupin, Alasdair Newson, Florence d’Alché-Buc)
- 11/2020: IV2020 workshop on 3D Deep Learning for Autonomous Driving (3D-DLAD)
- 07/2020: Code for our ECCV’20 paper “FLOT: Scene flow on point clouds guided by optimal transport” (FLOT).
- 06/2020: Seven papers accepted at ECCV’20 (27% acceptance rate)
- 06/2020: Four team members (Alexandre, Andrei, Matthieu and Renaud) acknowledged as Outstanding Reviewers at CVPR’20
- 06/2020: Valeo.ai participates to (virtual) CVPR’20, presenting 5 papers in main conference, delivering tutorials on annotation-efficient learning, co-organizing the OmniCV workshop and presenting keynote at the SAIAD workshop.
- 06/2020: Maximilian Jaritz defends his PhD at Inria Paris on “2D-3D Scene Understanding for Autonomous Driving” (reviewers: V. Lepetit and G. Brostow, external examiners: A. Dai and F. Jurie).
- 06/2020: Code for our CVPR’20 paper “xMUDA: Cross-Modal Unsupervised Domain Adaptation for 3D Semantic Segmentation” (xMUDA).
- 04/2020: Eloi Zablocki receives the Second Best Thesis Prize from the French Association for AI.
- 02/2020: Code for our NeurIPS’19 paper “Zero-Shot Semantic Segmentation” (ZS3).
- 02/2020: Five papers accepted at CVPR’20 (22% acceptance rate), inc. one oral.
- 01/2020: Spyros Gidaris, Andrei Bursuc and Karteek Alahari (Inria) to deliver a tutorial on Few-Shot, Self-Supervised, and Incremental Learning at CVPR’20.
- 01/2020: Pedestrian monitoring demo at CES, Las Vegas.
- 12/2019: Medium post: Is deep Reinforcement Learning really superhuman on Atari?
- 12/2019: Code for our ICCV’19 paper “Boosting few-shot visual learning with self-supervision” (BF3S).
- 12/2019: Code for our ICCV’19 paper “DADA: Depth-Aware Domain Adaptation in semantic segmentation” (DADA).
- 11/2019: Spyros Gidaris receives the Thesis Prize from Université Paris Est.
- 10/2019: Valeo.ai researchers present 5 papers at ICCV’19 in Seoul, Korea, and Valeo participates to associated workshops on Autonomous Driving and on Autonomous Navigation in Unconstrained Environments.
- 10/2019: PRAIRIE research institute is officially launched, with a nice day of talks and pannels (inc. one on future mobility, program in French). Followed by PRAIRIE AI Summer School (PAISS), where Patrick Pérez delivered a lecture (slides).
- 09/2019: Code of our NeurIPS’19 paper “Addressing failure prediction by learning model confidence” (ConfidNet).
- 09/2019: Work of Marin Toromanoff et al. discussed by Andrew Ng in The Batch (deeplearning.ai newsletter).
- 09/2019: Two papers accepted at NeurIPS’19 (21% acceptance rate).
- 09/2019: Matthieu Cord among NeurIPS’19 top contributors.
- 07/2029: Medium post: ADVENT: Adversarial entropy minimization for domain adaptation in semantic segmentation.
- 07/2019: Three papers accepted at ICCV’19 (24% acceptance rate), including one oral (4.6% acceptance rate).
- 07/2019: Code of our CVPR’19 paper “AdvEnt: Adversarial Entropy Minimization for Domain Adaptation in Semantic Segmentation” is available on valeo.ai github.
- 07/2019: Marin Toromanoff (PhD student with Mines ParisTech, Valeo DAR and Valeo.ai) ranks 1st on Track 2 of Carla Challenge 2019, and 2nd on Track 1.
- 06/2019: Spyros Gidaris receives the Best Thesis Prize from Ponts Foundation.
- 06/2019: Valeo.ai researchers present 8 papers (25% acceptance rate), including 4 orals (5.6% acceptance rate), at CVPR’19. Patrick Pérez delivers keynote on “Sustainable supervision with application to autonomous driving” at the Safe AI for Automated Driving (SAIAD) CVPR’19 Workshop.
- 05/2019: Hedi Ben-younes defends his PhD at Sorbonne Université, committee: M. Cord, V. Ferrari, Y. LeCun, P. Pérez, L. Soulier, N. Thome, J. Verbeek, Ch. Wolf.
- 05/2019: Himalaya Jain receives the Best Thesis Prize from Rennes 1 Foundation.
- 05/2019: Valeo is proud to be part of Prairie, the new Paris Interdisciplinary Artificial Intelligence Institute. Stay tuned.
- With seven other France’s global industry players, Valeo signs a Manifesto on AI for Industry.
- Jacques Aschenbroich (Chairman and CEO of Valeo) on AI and valeo.ai.
- Valeo Drive4u automated car in Paris center, a world premiere!
- Patrick Pérez on the start of Valeo.ai.
- Valeo Cruise4u automated car, 24h on Parisian ringroad.
Maxime Bucher, research scientist (page,scholar), now at Augustus Intelligence
Maximilian Jaritz, PhD student (page, scholar), now at Amazon
Gabriel de Marmiesse, research engineer (github), now at EarthCube