Jindan Huang

Papers

Total Citations

H-Index

About

Jindan Huang explores the intersection of human-robot interaction and interactive reinforcement learning (IntRL), with a focus on how human teaching dynamics shape autonomous agent development. Their work challenges conventional approaches that rely on perfect, simulated oracles for training, instead advocating for methodologies that account for real-world variability in human feedback. In their 2024 paper "Modeling Variation in Human Feedback with User Inputs," Huang introduces an exploratory framework that captures the nuanced, imperfect nature of human instruction—a critical step toward making IntRL systems more robust and practical. This work has already garnered 7 citations, signaling its early impact on the field. Huang’s subsequent study, "On the Effect of Robot Errors on Human Teaching Dynamics" (2024, 3 citations), further investigates how people naturally adapt their teaching strategies when robots make mistakes, revealing that human instructors dynamically adjust their feedback in response to agent performance. This insight has significant implications for designing more responsive, human-aware learning algorithms. By centering the human teacher’s role and variability, Huang’s research bridges the gap between idealized simulations and real-world deployment, offering a more realistic pathway for developing robots that learn effectively from natural human interaction.