Véhicule autonome | EPFL Graph Search

vignette|Le Chrysler Pacifica équipé de la technologie autonome Waymo. vignette|droite|Voiture autonome dont on distingue certains capteurs sur le toit. vignette|Une voiture sans pilote Robocar en présentation au grand prix de Formule E de 2017 à New York. vignette|Les voitures de Sber Autotech sont formées sur le terrain d'entraînement. Un véhicule autonome, véhicule automatisé, véhicule à délégation de conduite ou véhicule entièrement automatisé est un véhicule automobile capable de rouler sans intervention d'un conducteur. Le concept désigne un véhicule pouvant circuler sur la voie publique dans les situations prévues de trafic sans intervention humaine. C'est une application typique du domaine de la robotique mobile dans laquelle de nombreux acteurs sont engagés. Des éléments de solutions techniques, légales, psychologiques et juridiques ont déjà été introduits, mais certaines questions restent non résolues. La notion de voiture autonome couvre, selon le contexte, un véhicule t

Human- and machine-centred designs of molecules and materials for sustainability and decarbonization

Sokseiha Muy, Yang Yu

Data-driven approaches based on high-throughput capabilities and machine learning hold promise in revolutionizing human-centred materials discovery for sustainability and decarbonization. This Review examines the strengths and limitations of different traditional and emerging approaches to demonstrate their inherent connection and highlight the evolving paradigms of materials design. Breakthroughs in molecular and materials discovery require meaningful outliers to be identified in existing trends. As knowledge accumulates, the inherent bias of human intuition makes it harder to elucidate increasingly opaque chemical and physical principles. Moreover, given the limited manual and intellectual throughput of investigators, these principles cannot be efficiently applied to design new materials across a vast chemical space. Many data-driven approaches, following advances in high-throughput capabilities and machine learning, have tackled these limitations. In this Review, we compare traditional, human-centred methods with state-of-the-art, data-driven approaches to molecular and materials discovery. We first introduce the limitations of human-centred Edisonian, model-system and descriptor-based approaches. We then discuss how data-driven approaches can address these limitations by promoting throughput, reducing cognitive overload and biases, and establishing atomistic understanding that is transferable across a broad chemical space. We examine how high-throughput capabilities can be combined with active learning and inverse design to efficiently optimize materials out of millions or an intractable number of candidates. Lastly, we pinpoint challenges to accelerate future workflows and ultimately enable self-driving platforms, which automate and streamline the optimization of molecules and materials in iterative cycles.

NATURE PORTFOLIO2022

Understanding Deep Neural Networks using Adversarial Attacks

Krishna Kanth Nakka

Deep Neural Networks (DNNs) have achieved great success in a wide range of applications, such as image recognition, object detection, and semantic segmentation. Even thoughthe discriminative power of DNNs is nowadays unquestionable, serious concerns have arised ever since DNNs have shown to be vulnerable to adversarial examples craftedby adding imperceptible perturbations to clean images. The implications of these malicious attacks are even more significant for DNNs deployed in real-world systems, e.g.,autonomous driving and biometric authentication. Consequently, an intriguing question that we aim to understand is the underlying behavior of DNNs to adversarial attacks.This thesis contributes to a better understanding of the mechanism of adversarial attacks on DNNs. Our main contributions are broadly in two directions: (1) we proposeinterpretable architectures first to understand the reasons for the success of adversarial attacks and then to improve the robustness of DNNs; (2) we design intuitive adversarialattacks to both mislead and use as a tool to expand our present understanding of DNNs' internal workings and their limitations. In the first direction, we introduce deep architectures that allow humans to interpret the reasoning process of DNNs prediction. Specifically, we incorporate Bag-of-visual-wordsrepresentations from the pre-deep learning era into DNNs using an attention scheme. We find key reasons for adversarial attack success and use these insights to propose anadversarial defense by maximally separating the latent features of discriminative regions while minimizing the contribution of non-discriminative regions in the final prediction.The second direction deals with the design of adversarial attacks to understand DNNs' limitations in a real-world environment. To begin with, we show that existing state-of-the-art semantic segmentation networks that achieve superior performance by exploiting the context are highly susceptible to indirect local attacks. Furthermore, we demonstrate the existence of universal directional perturbations that are quasi-independent of the input template but still successfully fool unknown siamese-based visual object trackers. We then identify that the mid-level filter banks across different backbones bear strong similarities and thus can be potential common ground for attack. We, therefore, learn a generator that disrupts mid-level features with high transferability across different target architectures, datasets, and tasks. In short, our attacks highlight critical vulnerabilities of DNNs, which make their deployment challenging in the real-world environment, even in the extreme case when the attacker is unaware of the target architecture or the targetdata used to train it.Furthermore, we go beyond fooling networks and demonstrate the usefulness of adversarial attacks for studying the internal disentangled representations in self-supervised 3D pose estimation networks. We observe that adversarial manipulation of appearance information in the input image alters the pose output, indicating that the pose code contains appearance information and disentanglement is far from complete. Besides the above contributions, an underlying theme that arises multiple times in this thesis is counteracting the adversarial attacks by detecting them.

EPFL2022

Pedestrian 3D Bounding Box Prediction

Alexandre Massoud Alahi, Yi Zhou Ju, Saeed Saadatnejad

Safety is still the main issue of autonomous driving, and in order to be globally deployed, they need to predict pedestrians' motions sufficiently in advance. While there is a lot of research on coarse-grained (human center prediction) and fine-grained predictions (human body keypoints prediction), we focus on 3D bounding boxes, which are reasonable estimates of humans without modeling complex motion details for autonomous vehicles. This gives the flexibility to predict in longer horizons in real-world settings. We suggest this new problem and present a simple yet effective model for pedestrians' 3D bounding box prediction. This method follows an encoder-decoder architecture based on recurrent neural networks, and our experiments show its effectiveness in both the synthetic (JTA) and real-world (NuScenes) datasets. The learned representation has useful information to enhance the performance of other tasks, such as action anticipation. Our code is available online: https://github.com/vita-epfl/bounding-box-prediction

2022

Understanding Deep Neural Networks using Adversarial Attacks

Krishna Kanth Nakka

EPFL2022