Corpus-Level End-to-End Exploration for Interactive Systems

TL;DR

This paper introduces CE3, a novel corpus-level exploration method for interactive AI systems that improves document retrieval and exploration in dynamic search tasks by leveraging a global low-dimensional corpus representation.

Contribution

The paper presents a new corpus-level end-to-end exploration approach that enhances control and recovery in dynamic search agents through a global corpus embedding and a linear retrieval function.

Findings

CE3 outperforms existing dynamic search systems on TREC DD Track.

Global corpus embedding enables better exploration of under-explored areas.

Linear retrieval function allows end-to-end document manipulation.

Abstract

A core interest in building Artificial Intelligence (AI) agents is to let them interact with and assist humans. One example is Dynamic Search (DS), which models the process that a human works with a search engine agent to accomplish a complex and goal-oriented task. Early DS agents using Reinforcement Learning (RL) have only achieved limited success for (1) their lack of direct control over which documents to return and (2) the difficulty to recover from wrong search trajectories. In this paper, we present a novel corpus-level end-to-end exploration (CE3) method to address these issues. In our method, an entire text corpus is compressed into a global low-dimensional representation, which enables the agent to gain access to the full state and action spaces, including the under-explored areas. We also propose a new form of retrieval function, whose linear approximation allows end-to-end manipulation of documents. Experiments on the Text REtrieval Conference (TREC) Dynamic Domain (DD) Track show that CE3 outperforms the state-of-the-art DS systems.