【直播】【青年科学半月谈】Simultaneous State Estimation and Dynamics……
报告名称：
Simultaneous State Estimation and Dynamics Learning from Indirect Observations报告时间：
2022年11月10日 (周四) 10:00
报告嘉宾：
Jiayi Wei
主办单位：
蔻享学术
直播通道
蔻享学术直播间 |
报告人介绍
Jiayi Wei
Jiayi Wei is a 5th-year Computer Science PhD student at UT Austin, working with Isil Dillig and Greg Durrett . His current research focuses on applying deep learning techniques to programming languages tasks such as type inference and program synthesis. Previously, he also worked on diverse topics ranging from software security and verification to program synthesis and robotics. Before joining UT, he received his bachelor’s degree in Physics from the University of Science and Technology of China. He was also a former PhD intern at Meta and JetBrains.
报告简介
Accurate kinodynamic models play a crucial role in many robotics applications such as off-road navigation and high-speed driving. Many state-of-the-art approaches in learning stochastic kinodynamic models, however, require precise measurements of robot states as labeled input/output examples, which can be hard to obtain in outdoor settings due to limited sensor capabilities and the absence of ground truth. In this work, we propose a new technique for learning neural stochastic kinodynamic models from noisy and indirect observations by performing simultaneous state estimation and dynamics learning. The proposed technique iteratively improves the kinodynamic model in an expectation-maximization loop, where the E Step samples posterior state trajectories using particle filtering, and the M Step updates the dynamics to be more consistent with the sampled trajectories via stochastic gradient ascent. We evaluate our approach on both simulation and real-world benchmarks and compare it with several baseline techniques. Our approach not only achieves significantly higher accuracy but is also more robust to observation noise, thereby showing promise for boosting the performance of many other robotics applications.
推荐阅读
【青年科学半月谈】人工智能的产业化——人工智能科研的互补视角 >>
【青年科学半月谈】可原位调控的集成光学克尔非线性 >>【青年科学半月谈】机器学习预测材料性质：难点与改进策略 >>【青年科学半月谈】非厄米趋肤效应的反常动力学 >>编辑：吴良秀
蔻享学术 平台介绍
蔻享学术平台，国内领先的一站式科学资源共享平台，依托国内外一流科研院所、高等院校和企业的科研力量，聚焦前沿科学，以优化科研创新环境、传播和服务科学、促进学科交叉融合为宗旨，打造优质学术资源的共享数据平台。
识别二维码，
下载 蔻享APP 查看最新资源数据。
点击阅读原文，查看更多精彩报告！