10 papers
Anomaly-Informed Confidence Calibration for Vision-Based Safety Prediction
Zhenjiang Mao, Jiawen Wu, Gabriel Wagner +2
Reliable confidence estimates are important for safely deploying vision-based controllers in autonomous racing, where safety predictions must be derived from camera images, yet mod…
Confidence over Time: Confidence Calibration with Temporal Logic for Large Language Model Reasoning
Zhenjiang Mao, Anirudhh Venkat, Artem Bisliouk +4
Large Language Models (LLMs) increasingly rely on long-form, multi-step reasoning to solve complex tasks such as mathematical problem solving and scientific question answering. Des…
Recurrent Confidence Chain: Temporal-Aware Uncertainty Quantification in Large Language Models
Zhenjiang Mao, Anirudhh Venkat
As reasoning modules, such as the chain-of-thought mechanism, are applied to large language models, they achieve strong performance on various tasks such as answering common-sense…
How Safe Will I Be Given What I Saw? Calibrated Prediction of Safety Chances for Image-Controlled Autonomy
Zhenjiang Mao, Mrinall Eashaan Umasudhan, Ivan Ruchkin
Autonomous robots that rely on deep neural network controllers pose critical challenges for safety prediction, especially under partial observability and distribution shift. Tradit…
Temporalizing Confidence: Evaluation of Chain-of-Thought Reasoning with Signal Temporal Logic
Zhenjiang Mao, Artem Bisliouk, Rohith Reddy Nama +1
Large Language Models (LLMs) have shown impressive performance in mathematical reasoning tasks when guided by Chain-of-Thought (CoT) prompting. However, they tend to produce highly…
Generalizable Image Repair for Robust Visual Control
Carson Sobolewski, Zhenjiang Mao, Kshitij Maruti Vejre +1
Vision-based control relies on accurate perception to achieve robustness. However, image distribution changes caused by sensor noise, adverse weather, and dynamic lighting can degr…
Four Principles for Physically Interpretable World Models
Jordan Peper, Zhenjiang Mao, Yuang Geng +2
As autonomous systems are increasingly deployed in open and uncertain settings, there is a growing need for trustworthy world models that can reliably predict future high-dimension…
Physically Interpretable World Models via Weakly Supervised Representation Learning
Zhenjiang Mao, Mrinall Eashaan Umasudhan, Ivan Ruchkin
Learning predictive models from high-dimensional sensory observations is fundamental for cyber-physical systems, yet the latent representations learned by standard world models lac…
Language-Enhanced Latent Representations for Out-of-Distribution Detection in Autonomous Driving
Zhenjiang Mao, Dong-You Jhong, Ao Wang +1
Out-of-distribution (OOD) detection is essential in autonomous driving, to determine when learning-based components encounter unexpected inputs. Traditional detectors typically use…
Zero-shot Safety Prediction for Autonomous Robots with Foundation World Models
Zhenjiang Mao, Siqi Dai, Yuang Geng +1
A world model creates a surrogate world to train a controller and predict safety violations by learning the internal dynamic model of systems. However, the existing world models re…