Let the Agent Search: Autonomous Exploration Beats Rigid Workflows in Temporal Question Answering

2026-03-02

AI Summary

This work addresses the challenges of multi-hop reasoning and temporal constraint handling in temporal knowledge graph question answering by proposing AT2QA, a training-free autonomous agent. Operating in a zero-shot setting, AT2QA leverages off-the-shelf large language models and general-purpose search tools through dynamic interaction, autonomously planning reasoning paths to answer complex temporal questions without human-designed pipelines or supervised fine-tuning. The study demonstrates for the first time that endowing large language models with autonomous decision-making capabilities alone can substantially enhance performance, particularly on multi-target queries. On the MultiTQ benchmark, AT2QA achieves a Hits@1 score of 88.7%, outperforming the previous state-of-the-art method by 10.7% overall and by as much as 20.1% on multi-target questions.

Technology Category

• Natural Language Processing: Question Answering
• Planning, Routing, and Scheduling: Temporal Planning
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Temporal Knowledge Graph Question Answering (TKGQA) demands multi-hop reasoning under temporal constraints. Prior approaches based on large language models (LLMs) typically rely on rigid, hand-crafted retrieval workflows or costly supervised fine-tuning. We show that simply granting an off-the-shelf LLM autonomy, that is, letting it decide what to do next, already yields substantial gains even in a strict zero-shot setting. Building on this insight, we propose AT2QA, an autonomous, training-free agent for temporal question answering that iteratively interacts with the temporal knowledge graph via a general search tool for dynamic retrieval. Experiments on MultiTQ demonstrate large improvements: AT2QA achieves 88.7% Hits@1 (+10.7% over prior SOTA), including a +20.1% gain on challenging multi-target queries, showing that agentic autonomy can decisively outperform fine-tuning for temporal question answering. Code and the full set of sampled trajectories are available on https://github.com/AT2QA-Official-Code/AT2QA-Official-Code

Problem

• Temporal Knowledge Graph Question Answering
• multi-hop reasoning
• temporal constraints
• question answering

Innovation

• autonomous agent
• temporal knowledge graph
• zero-shot reasoning
• dynamic retrieval
• LLM-based search