Hierarchical Sequence Iteration for Heterogeneous Question Answering

Ruiyi Yang, Hao Xue, Imran Razzak, Hakim Hacid, Flora D. Salim

Keywords: Retrieval-Augmented Generation (RAG); Iterative Retrieval; Hierarchical Sequence (HSEQ); Heterogeneous Question Answering; Multi-hop Reasoning; Agentic LLMs;

TL;DR: Unifies text, tables, and knowledge graphs into a hierarchical sequence and performs head-guided, budget-aware iteration retrieval and reasoning to collect minimal sufficient evidence before answer synthesis.

Abstract: Retrieval-augmented generation (RAG) remains brittle on multi-step questions and heterogeneous evidence sources, trading accuracy against latency and token/tool budgets. This paper introduces **Hierarchical Sequence (HSEQ) Iteration** for **Heterogeneous Question Answering**, a unified framework that (i) linearize documents, tables, and knowledge graphs into a reversible hierarchical sequence with lightweight structural tags, and (ii) perform structure-aware iteration to collect just-enough evidence before answer synthesis. A Head Agent provides guidance that leads retrieval, while an Iteration Agent selects and expands HSeq via structure-respecting actions (e.g., parent/child hops, table row/column neighbors, KG relations); Finally the head agent composes canonicalized evidence to generate the final answer, with an optional refinement loop to resolve detected contradictions. Experiments on HotpotQA (text), HybridQA/TAT-QA (table+text), and MetaQA (KG) show consistent EM/F1 gains over strong single-pass, multi-hop, and agentic RAG baselines, alongside higher efficiency. Beyond aggregate metrics, HSEQ exhibits three key advantages: (1) a **format-agnostic unification** that enables a single policy to operate across text, tables, and KGs without per-dataset specialization; (2) **guided, budget-aware iteration** that reduces unnecessary hops, tool calls, and tokens while preserving answer quality; and (3) **evidence canonicalization for reliable QA**, improving consistency and auditability of the generated answers.

Primary Area: foundation or frontier models, including LLMs

Submission Number: 5875