Learning by Cheating

• URL: http://arxiv.org/abs/1912.12294v1
• Date: Fri, 27 Dec 2019 18:59:04 GMT
• Title: Learning by Cheating
• Authors: Dian Chen and Brady Zhou and Vladlen Koltun and Philipp Kr\"ahenb\"uhl
• Abstract summary: We show that this challenging learning problem can be simplified by decomposing it into two stages. We use the presented approach to train a vision-based autonomous driving system that substantially outperforms the state of the art. Our approach achieves, for the first time, 100% success rate on all tasks in the original CARLA benchmark, sets a new record on the NoCrash benchmark, and reduces the frequency of infractions by an order of magnitude compared to the prior state of the art.
• Score: 72.9701333689606
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Vision-based urban driving is hard. The autonomous system needs to learn to perceive the world and act in it. We show that this challenging learning problem can be simplified by decomposing it into two stages. We first train an agent that has access to privileged information. This privileged agent cheats by observing the ground-truth layout of the environment and the positions of all traffic participants. In the second stage, the privileged agent acts as a teacher that trains a purely vision-based sensorimotor agent. The resulting sensorimotor agent does not have access to any privileged information and does not cheat. This two-stage training procedure is counter-intuitive at first, but has a number of important advantages that we analyze and empirically demonstrate. We use the presented approach to train a vision-based autonomous driving system that substantially outperforms the state of the art on the CARLA benchmark and the recent NoCrash benchmark. Our approach achieves, for the first time, 100% success rate on all tasks in the original CARLA benchmark, sets a new record on the NoCrash benchmark, and reduces the frequency of infractions by an order of magnitude compared to the prior state of the art. For the video that summarizes this work, see https://youtu.be/u9ZCxxD-UUw

Related papers

• Learn 2 Rage: Experiencing The Emotional Roller Coaster That Is Reinforcement Learning [5.962453678471195]
  This work presents the experiments and solution outline for our teams winning submission in the Learn To Race Autonomous Racing Virtual Challenge 2022 hosted by AIcrowd. The objective of the Learn-to-Race competition is to push the boundary of autonomous technology, with a focus on achieving the safety benefits of autonomous driving. We focused our initial efforts on implementation of Soft Actor Critic (SAC) variants. Our goal was to learn non-trivial control of the race car exclusively from visual and geometric features, directly mapping pixels to control actions.
  arXiv  Detail & Related papers  (2024-10-24T06:16:52Z)
• Data and Knowledge for Overtaking Scenarios in Autonomous Driving [0.0]
  The overtaking maneuver is one of the most critical actions of driving. Despite the amount of work available in the literature, just a few handle overtaking maneuvers. This work contributes in this area by presenting a new synthetic dataset whose focus is the overtaking maneuver.
  arXiv  Detail & Related papers  (2023-05-30T21:27:05Z)
• FastRLAP: A System for Learning High-Speed Driving via Deep RL and Autonomous Practicing [71.76084256567599]
  We present a system that enables an autonomous small-scale RC car to drive aggressively from visual observations using reinforcement learning (RL) Our system, FastRLAP (faster lap), trains autonomously in the real world, without human interventions, and without requiring any simulation or expert demonstrations. The resulting policies exhibit emergent aggressive driving skills, such as timing braking and acceleration around turns and avoiding areas which impede the robot's motion, approaching the performance of a human driver using a similar first-person interface over the course of training.
  arXiv  Detail & Related papers  (2023-04-19T17:33:47Z)
• Scaling Vision-based End-to-End Driving with Multi-View Attention Learning [7.14967754486195]
  We present CIL++, which improves on CILRS by both processing higher-resolution images using a human-inspired HFOV as an inductive bias and incorporating a proper attention mechanism. We propose to replace CILRS with CIL++ as a strong vision-based pure end-to-end driving baseline supervised by only vehicle signals and trained by conditional imitation learning.
  arXiv  Detail & Related papers  (2023-02-07T02:14:45Z)
• Learning from All Vehicles [1.1947714868715738]
  We present a system to train driving policies from experiences collected not just from the ego-vehicle, but all vehicles that it observes. This system uses the behaviors of other agents to create more diverse driving scenarios without collecting additional data. Our method won the 2021 CARLA Autonomous Driving challenge.
  arXiv  Detail & Related papers  (2022-03-22T17:59:04Z)
• YOLOP: You Only Look Once for Panoptic Driving Perception [21.802146960999394]
  We present a panoptic driving perception network (YOLOP) to perform traffic object detection, drivable area segmentation and lane detection simultaneously. It is composed of one encoder for feature extraction and three decoders to handle the specific tasks. Our model performs extremely well on the challenging BDD100K dataset, achieving state-of-the-art on all three tasks in terms of accuracy and speed.
  arXiv  Detail & Related papers  (2021-08-25T14:19:42Z)
• End-to-End Urban Driving by Imitating a Reinforcement Learning Coach [148.2683592850329]
  Humans are good drivers, but not good coaches for end-to-end algorithms. We train a reinforcement learning expert that maps bird's-eye view images to continuous low-level actions. Supervised by our reinforcement learning coach, a baseline end-to-end agent with monocular camera-input achieves expert-level performance.
  arXiv  Detail & Related papers  (2021-08-18T17:36:51Z)
• Learning to drive from a world on rails [78.28647825246472]
  We learn an interactive vision-based driving policy from pre-recorded driving logs via a model-based approach. A forward model of the world supervises a driving policy that predicts the outcome of any potential driving trajectory. Our method ranks first on the CARLA leaderboard, attaining a 25% higher driving score while using 40 times less data.
  arXiv  Detail & Related papers  (2021-05-03T05:55:30Z)
• End-to-End Intersection Handling using Multi-Agent Deep Reinforcement Learning [63.56464608571663]
  Navigating through intersections is one of the main challenging tasks for an autonomous vehicle. In this work, we focus on the implementation of a system able to navigate through intersections where only traffic signs are provided. We propose a multi-agent system using a continuous, model-free Deep Reinforcement Learning algorithm used to train a neural network for predicting both the acceleration and the steering angle at each time step.
  arXiv  Detail & Related papers  (2021-04-28T07:54:40Z)
• Learning Accurate and Human-Like Driving using Semantic Maps and Attention [152.48143666881418]
  This paper investigates how end-to-end driving models can be improved to drive more accurately and human-like. We exploit semantic and visual maps from HERE Technologies and augment the existing Drive360 dataset with such. Our models are trained and evaluated on the Drive360 + HERE dataset, which features 60 hours and 3000 km of real-world driving data.
  arXiv  Detail & Related papers  (2020-07-10T22:25:27Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.