Search Results for author: Yorie Nakahira

Found 10 papers, 4 papers with code

Generalizable Physics-Informed Learning for Stochastic Safety-Critical Systems

no code implementations11 Jul 2024 Zhuoyuan Wang, Albert Chern, Yorie Nakahira

Accurate estimate of long-term risk is critical for safe decision-making, but sampling from rare risk events and long-term trajectories can be prohibitively costly.

Decision Making

Learning to Stabilize Unknown LTI Systems on a Single Trajectory under Stochastic Noise

no code implementations31 May 2024 Ziyi Zhang, Yorie Nakahira, Guannan Qu

We study the problem of learning to stabilize unknown noisy Linear Time-Invariant (LTI) systems on a single trajectory.

Myopically Verifiable Probabilistic Certificates for Safe Control and Learning

no code implementations23 Apr 2024 Zhuoyuan Wang, Haoming Jing, Christian Kurniawan, Albert Chern, Yorie Nakahira

When the target probability is defined using long-term trajectories, this technique can be used to design myopic conditions/controllers with assured long-term safe probability.

Decision Making

Physics-informed RL for Maximal Safety Probability Estimation

1 code implementation25 Mar 2024 Hikaru Hoshino, Yorie Nakahira

Motivated by this, we study how to estimate the long-term safety probability of maximally safe actions without sufficient coverage of samples from risky states and long-term trajectories.

reinforcement-learning

An Analytic Solution to Covariance Propagation in Neural Networks

1 code implementation24 Mar 2024 Oren Wright, Yorie Nakahira, José M. F. Moura

Uncertainty quantification of neural networks is critical to measuring the reliability and robustness of deep learning systems.

Uncertainty Quantification

Context-aware LLM-based Safe Control Against Latent Risks

no code implementations18 Mar 2024 Quan Khanh Luu, Xiyu Deng, Anh Van Ho, Yorie Nakahira

It is challenging for autonomous control systems to perform complex tasks in the presence of latent risks.

Sample-Optimal Zero-Violation Safety For Continuous Control

no code implementations9 Mar 2024 Ritabrata Ray, Yorie Nakahira, Soummya Kar

In this paper, we study the problem of ensuring safety with a few shots of samples for partially unknown systems.

Continuous Control

Physics-Informed Representation and Learning: Control and Risk Quantification

1 code implementation17 Dec 2023 Zhuoyuan Wang, Reece Keller, Xiyu Deng, Kenta Hoshino, Takashi Tanaka, Yorie Nakahira

Optimal and safety-critical control are fundamental problems for stochastic systems, and are widely considered in real-world scenarios such as robotic manipulation and autonomous driving.

Autonomous Driving Dimensionality Reduction

A Generalizable Physics-informed Learning Framework for Risk Probability Estimation

1 code implementation10 May 2023 Zhuoyuan Wang, Yorie Nakahira

In this paper, we develop an efficient method to evaluate the probabilities of long-term risk and their gradients.

A Learning and Control Perspective for Microfinance

no code implementations26 Jul 2022 Christian Kurniawan, Xiyu Deng, Adhiraj Chakraborty, Assane Gueye, Niangjun Chen, Yorie Nakahira

Although many methods in regular finance can estimate credit scores and default probabilities, these methods are not directly applicable to microfinance due to the following unique characteristics: a) under-explored (developing) areas such as rural Africa do not have sufficient prior loan data for microfinance institutions (MFIs) to establish a credit scoring system; b) microfinance applicants may have difficulty providing sufficient information for MFIs to accurately predict default probabilities; and c) many MFIs use group liability (instead of collateral) to secure repayment.

Decision Making Fairness

Cannot find the paper you are looking for? You can Submit a new open access paper.