Results 231 to 240 of about 522,323 (249)

RoaD: Rollouts as Demonstrations for Closed-Loop Supervised Fine-Tuning of Autonomous Driving Policies

arXiv.org
Autonomous driving policies are typically trained via open-loop behavior cloning of human demonstrations. However, such policies suffer from covariate shift when deployed in closed loop, leading to compounding errors.
Guillermo Garcia-Cobo, Maximilian Igl, Peter Karkus, Zhejun Zhang, Michael Watson, Yuxiao Chen, B. Ivanovic, Marco Pavone   +7 more
semanticscholar   +1 more source

From Foresight to Forethought: VLM-In-the-Loop Policy Steering via Latent Alignment

Robotics
While generative robot policies have demonstrated significant potential in learning complex, multimodal behaviors from demonstrations, they still exhibit diverse failures at deployment-time.
Yilin Wu, Ran Tian, Gokul Swamy, Andrea Bajcsy   +3 more
semanticscholar   +1 more source

Closed-Loop Supervised Fine-Tuning of Tokenized Traffic Models

Computer Vision and Pattern Recognition
Traffic simulation aims to learn a policy for traffic agents that, when unrolled in closed-loop, faithfully recovers the joint distribution of trajectories observed in the real world. Inspired by large language models, tokenized multi-agent policies have
Zhejun Zhang, Peter Karkus, Maximilian Igl, Wenhao Ding, Yuxiao Chen, B. Ivanovic, Marco Pavone   +6 more
semanticscholar   +1 more source

Improving the Transparency of Robot Policies Using Demonstrations and Reward Communication

ACM Trans. Hum. Robot Interact.
Demonstrations are a powerful way to teach robot decision-making to humans. Although informative demonstrations may be selected a priori using the machine teaching framework, student learning may deviate from the pre-selected curriculum in situ.
Michael S. Lee, Reid G. Simmons, H. Admoni   +2 more
semanticscholar   +1 more source

Closed and Open Loop Oil Taxation Policies in New Mexico

SSRN Electronic Journal, 2023
Saeed Langarudi, Sajid Al Noor, Kyle Eagar   +2 more
openaire   +1 more source

LoRD: Adapting Differentiable Driving Policies to Distribution Shifts

IEEE International Conference on Robotics and Automation
Distribution shifts between operational domains can severely affect the performance of learned models in self-driving vehicles (SDVs). While this is a well-established problem, prior work has mostly explored naive solutions such as fine-tuning, focusing ...
Christopher Diehl, Peter Karkus, Shushant Veer, Marco Pavone, Torsten Bertram   +4 more
semanticscholar   +1 more source

Open‐loop Stackelberg learning solution for hierarchical control problems

, 2019
K. Vamvoudakis, F. Lewis, W. Dixon
semanticscholar   +1 more source

Quality control in a single state production system: open and closed loop policies

, 1990
D. Murthy, I. Djamaludin
semanticscholar   +1 more source

Minimum-time open-loop and closed-loop optimal guidance with GA-PSO and neural fuzzy for Samarai MAV flight

IEEE Aerospace and Electronic Systems Magazine, 2015
M. Samani, M. Tafreshi, I. Shafieenejad, A. Nikkhah   +3 more
semanticscholar   +1 more source

Kyoto and Beyond Kyoto Climate Policy: Comparison of Open-Loop and Feedback Game Outcomes

2004
Juan Carlos Císcar, Antonio Soria
openaire   +1 more source