# Calibrated Fusion for Heterogeneous Graph-Vector Retrieval in Multi-Hop QA

arXiv: https://arxiv.org/abs/2603.28886 · HTML: https://mlsanity.com/paper/2603.28886/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 5 of its 25 references that are in the map.
- First posted on arXiv 30 Mar 2026.

## Builds on (most cited first, 5 of 5)

- [HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering](https://mlsanity.com/paper/1809.09600/) (2018)
- [Constructing A Multi-hop QA Dataset for Comprehensive Evaluation of Reasoning Steps](https://mlsanity.com/paper/2011.01060/) (2020)
- [NV-Embed: Improved Techniques for Training LLMs as Generalist Embedding Models](https://mlsanity.com/paper/2405.17428/) (2024)
- [From RAG to Memory: Non-Parametric Continual Learning for Large Language Models](https://mlsanity.com/paper/2502.14802/) (2025)
- [Can Knowledge-Graph-based Retrieval Augmented Generation Really Retrieve What You Need?](https://mlsanity.com/paper/2510.16582/) (2025)

## Built on by (most cited first, 2 of 2)

- [Per-Query Gating of LLM Rerankers for Multi-Hop Retrieval](https://mlsanity.com/paper/2609.22880/) (2026)
- [Align Before You Combine: Reference Space Calibration for Supervision Without Ground Truth](https://mlsanity.com/paper/2610.09525/) (2026)

## Abstract (arXiv)

Graph-augmented retrieval combines dense similarity with graph-based relevance signals such as Personalized PageRank (PPR), but these scores have different distributions and are not directly comparable. We study this as a score calibration problem for heterogeneous retrieval fusion in multi-hop question answering. Our method, PhaseGraph, maps vector and graph scores to a common unit-free scale using percentile-rank normalization (PIT) before fusion, enabling stable combination without discarding magnitude information. Across MuSiQue and 2WikiMultiHopQA, calibrated fusion improves held-out last-hop retrieval on HippoRAG2-style benchmarks: LastHop@10 increases from 69.1% to 71.0% on MuSiQue (15W/5L, p=0.041, n=514) and LastHop@5 from 51.7% to 53.6% on 2WikiMultiHopQA (11W/2L, p=0.023, n=491), both on independent held-out test splits. Against the official HippoRAG 2 pipeline on 2WikiMultiHopQA, calibrated fusion is ahead at LastHop@10 (+6.3pp, p<10^-3) and behind at LastHop@5 (-8.4pp), a cutoff-dependent cross-over we report in full. A theory-driven ablation shows that percentile-based calibration is directionally more robust than min-max normalization on both tune and test splits (1W/6L, p=0.125), while Boltzmann weighting performs comparably to linear fusion after calibration (0W/3L, p=0.25). These results suggest that score commensuration is a robust design choice, and the exact post-calibration operator appears to matter less on these benchmarks.