PREDICT: Preference Reasoning by Evaluating Decomposed preferences Inferred from Candidate Trajectories

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Aroca-Ouellette, Stephane, Mackraz, Natalie, Theobald, Barry-John, Metcalf, Katherine
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917798012256256
author Aroca-Ouellette, Stephane
Mackraz, Natalie
Theobald, Barry-John
Metcalf, Katherine
author_facet Aroca-Ouellette, Stephane
Mackraz, Natalie
Theobald, Barry-John
Metcalf, Katherine
contents Accommodating human preferences is essential for creating AI agents that deliver personalized and effective interactions. Recent work has shown the potential for LLMs to infer preferences from user interactions, but they often produce broad and generic preferences, failing to capture the unique and individualized nature of human preferences. This paper introduces PREDICT, a method designed to enhance the precision and adaptability of inferring preferences. PREDICT incorporates three key elements: (1) iterative refinement of inferred preferences, (2) decomposition of preferences into constituent components, and (3) validation of preferences across multiple trajectories. We evaluate PREDICT on two distinct environments: a gridworld setting and a new text-domain environment (PLUME). PREDICT more accurately infers nuanced human preferences improving over existing baselines by 66.2\% (gridworld environment) and 41.0\% (PLUME).
format Preprint
id arxiv_https___arxiv_org_abs_2410_06273
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle PREDICT: Preference Reasoning by Evaluating Decomposed preferences Inferred from Candidate Trajectories
Aroca-Ouellette, Stephane
Mackraz, Natalie
Theobald, Barry-John
Metcalf, Katherine
Artificial Intelligence
Human-Computer Interaction
Accommodating human preferences is essential for creating AI agents that deliver personalized and effective interactions. Recent work has shown the potential for LLMs to infer preferences from user interactions, but they often produce broad and generic preferences, failing to capture the unique and individualized nature of human preferences. This paper introduces PREDICT, a method designed to enhance the precision and adaptability of inferring preferences. PREDICT incorporates three key elements: (1) iterative refinement of inferred preferences, (2) decomposition of preferences into constituent components, and (3) validation of preferences across multiple trajectories. We evaluate PREDICT on two distinct environments: a gridworld setting and a new text-domain environment (PLUME). PREDICT more accurately infers nuanced human preferences improving over existing baselines by 66.2\% (gridworld environment) and 41.0\% (PLUME).
title PREDICT: Preference Reasoning by Evaluating Decomposed preferences Inferred from Candidate Trajectories
topic Artificial Intelligence
Human-Computer Interaction
url https://arxiv.org/abs/2410.06273