Fast Neural Perception and Learning for Intelligent Vehicles and Robotics

Abstract

Recently, intelligent perception and learning techniques have obtained wide attention in the areas of autonomous vehicles and robotics. The goal is that the robots and vehicles can learn/adapt its surrounding environment via multiple types of sensors (such as optical, vision or acoustic sensors) to conduct different tasks with intelligent learning approaches. Among many intelligent learning approaches, neural networks (NNs), especially recently-proposed light-weight NNs (Li-NNs), have gained a series of success across various areas including image, lidar, decision-making as well as user-interaction data. However, there are still many intriguing research problems, such as accuracy and robustness under uncertainties, learning efficiency for real robotic environment. Compared to traditional methods, Li-NNs are able to gain efficiency and prove to achieving real-time performance by parallel computing and strong processing power. This workshop will bring together participants from academia and industry alike to share advancements and new technologies in the field of intelligent vehicles and robotics. The attendees of this workshop will be introduced to fast neural perception and learning from academic experts in the field. Experts from the industry will explain the current needs in intelligent vehicles and robotics, which will inspire researchers with challenges drawn from real use case scenarios. Experts from the academia will bring the latest advancements in the field, providing potential new solutions to real problems. The organizers and the invited speakers of this workshop have a multidisciplinary background that will stimulate interesting discussions, promote the cross-fertilization of ideas and encourage future collaborations.