Search Results - "Hong, Joey"

Rules of the Road: Predicting Driving Behavior With a Convolutional Model of Semantic Interactions by Hong, Joey, Sapp, Benjamin, Philbin, James

“…We focus on the problem of predicting future states of entities in complex, real-world driving scenarios. Previous research has approached this problem via…”
Get full text

Wearing the Witch Identity as a Way of Becoming in Shirley Jackson's We Have Always Lived in the Castle by Hong, Joey Junsu

“…This thesis aims to analyze Shirley Jackson’s last novel We Have Always Lived in the Castle (1962) through the lens of 17th-century New England history and…”
Get full text

Q-SFT: Q-Learning for Language Models via Supervised Fine-Tuning by Hong, Joey, Dragan, Anca, Levine, Sergey

“…Value-based reinforcement learning (RL) can in principle learn effective policies for a wide range of multi-turn problems, from games to dialogue to robotic…”
Get full text

Zero-Shot Goal-Directed Dialogue via RL on Imagined Conversations by Hong, Joey, Levine, Sergey, Dragan, Anca

“…Large language models (LLMs) have emerged as powerful and general solutions to many natural language tasks. However, many of the most important applications of…”
Get full text

Offline RL with Observation Histories: Analyzing and Improving Sample Complexity by Hong, Joey, Dragan, Anca, Levine, Sergey

“…Offline reinforcement learning (RL) can in principle synthesize more optimal behavior from a dataset consisting only of suboptimal trials. One way that this…”
Get full text

Interactive Dialogue Agents via Reinforcement Learning on Hindsight Regenerations by Hong, Joey, Lin, Jessica, Dragan, Anca, Levine, Sergey

“…Recent progress on large language models (LLMs) has enabled dialogue agents to generate highly naturalistic and plausible text. However, current LLM language…”
Get full text

Learning to Influence Human Behavior with Offline Reinforcement Learning by Hong, Joey, Levine, Sergey, Dragan, Anca

“…When interacting with people, AI agents do not just influence the state of the world -- they also influence the actions people take in response to the agent,…”
Get full text

On the Sensitivity of Reward Inference to Misspecified Human Models by Hong, Joey, Bhatia, Kush, Dragan, Anca

“…Inferring reward functions from human behavior is at the center of value alignment - aligning AI objectives with what we, humans, actually want. But doing so…”
Get full text

Confidence-Conditioned Value Functions for Offline Reinforcement Learning by Hong, Joey, Kumar, Aviral, Levine, Sergey

“…Offline reinforcement learning (RL) promises the ability to learn effective policies solely using existing, static datasets, without any costly online…”
Get full text

Strategically Conservative Q-Learning by Shimizu, Yutaka, Hong, Joey, Levine, Sergey, Tomizuka, Masayoshi

“…Offline reinforcement learning (RL) is a compelling paradigm to extend RL's practical utility by leveraging pre-collected, static datasets, thereby avoiding…”
Get full text

Multi-Task Off-Policy Learning from Bandit Feedback by Hong, Joey, Kveton, Branislav, Katariya, Sumeet, Zaheer, Manzil, Ghavamzadeh, Mohammad

“…Many practical applications, such as recommender systems and learning to rank, involve solving multiple similar tasks. One example is learning of…”
Get full text

When Should We Prefer Offline Reinforcement Learning Over Behavioral Cloning? by Kumar, Aviral, Hong, Joey, Singh, Anikait, Levine, Sergey

“…Offline reinforcement learning (RL) algorithms can acquire effective policies by utilizing previously collected experience, without any online interaction. It…”
Get full text

Compositional Generalization and Decomposition in Neural Program Synthesis by Shi, Kensen, Hong, Joey, Zaheer, Manzil, Yin, Pengcheng, Sutton, Charles

“…When writing programs, people have the ability to tackle a new complex task by decomposing it into smaller and more familiar subtasks. While it is difficult to…”
Get full text

LMRL Gym: Benchmarks for Multi-Turn Reinforcement Learning with Language Models by Abdulhai, Marwa, White, Isadora, Snell, Charlie, Sun, Charles, Hong, Joey, Zhai, Yuexiang, Xu, Kelvin, Levine, Sergey

“…Large language models (LLMs) provide excellent text-generation capabilities, but standard prompting and generation methods generally do not lead to intentional…”
Get full text

Deep Hierarchy in Bandits by Hong, Joey, Kveton, Branislav, Katariya, Sumeet, Zaheer, Manzil, Ghavamzadeh, Mohammad

“…Mean rewards of actions are often correlated. The form of these correlations may be complex and unknown a priori, such as the preferences of a user for…”
Get full text

Hierarchical Bayesian Bandits by Hong, Joey, Kveton, Branislav, Zaheer, Manzil, Ghavamzadeh, Mohammad

“…Meta-, multi-task, and federated learning can be all viewed as solving similar tasks, drawn from a distribution that reflects task similarities. We provide a…”
Get full text

ExeDec: Execution Decomposition for Compositional Generalization in Neural Program Synthesis by Shi, Kensen, Hong, Joey, Deng, Yinlin, Yin, Pengcheng, Zaheer, Manzil, Sutton, Charles

“…When writing programs, people have the ability to tackle a new complex task by decomposing it into smaller and more familiar subtasks. While it is difficult to…”
Get full text

Thompson Sampling with a Mixture Prior by Hong, Joey, Kveton, Branislav, Zaheer, Manzil, Ghavamzadeh, Mohammad, Boutilier, Craig

“…We study Thompson sampling (TS) in online decision making, where the uncertain environment is sampled from a mixture distribution. This is relevant in…”
Get full text

Rules of the Road: Predicting Driving Behavior with a Convolutional Model of Semantic Interactions by Hong, Joey, Sapp, Benjamin, Philbin, James

“…The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019, pp. 8454-8462 We focus on the problem of predicting future states of entities in…”
Get full text

Latent Programmer: Discrete Latent Codes for Program Synthesis by Hong, Joey, Dohan, David, Singh, Rishabh, Sutton, Charles, Zaheer, Manzil

“…In many sequence learning tasks, such as program synthesis and document summarization, a key problem is searching over a large space of possible output…”
Get full text