This work proposes RSF-GLLM, a framework decoupling differentiable graph reasoning from answer generation, and introduces flow sparsity regularization to theoretically guarantee convergence from soft probabilities to discrete reasoning paths.
Abstract
Multi-hop Question Answering over Knowledge Graphs faces a critical challenge: traditional retrieve-then-read pipelines break differentiability, preventing the retriever from learning to bridge the semantic gap where intermediate nodes lack lexical overlap with the query. To address this, we propose RSF-GLLM, a framework decoupling differentiable graph reasoning from answer generation. Our Recurrent Soft-Flow (RSF) module employs a GRU-guided query updater to propagate continuous relevance scores, utilizing a dynamic gating mechanism to traverse semantically dissimilar bridge nodes via structural cues. We introduce flow sparsity regularization to theoretically guarantee convergence from soft probabilities to discrete reasoning paths. These paths are extracted and textualized to fine-tune a Large Language Model (LLM), ensuring generation is grounded in factual topology. Experiments on WebQSP and CWQ demonstrate that RSF-GLLM achieves competitive performance with superior inference efficiency compared to LLM based computationally expensive approaches.
Search-on-Graph-R1 (\sogrone{}), which internalizes this navigation into a compact 8B model through supervised fine-tuning (SFT) followed by reinforcement learning (RL), which surpasses every frozen frontier-LLM system in their comparison and posts the strongest results on CWQ of any system the authors compare against.
J. Sun, Hao Yu, Fengran Mo et al.· arXiv.org· 1 citation· ⚡1
Cross-document knowledge retrieval requires the retrieval and synthesis of evidence scattered across multiple passages linked by implicit or explicit inter-document references. Existing methods either treat the corpus as a flat collection of independent passages, thereby forfeiting the links that connect evidence fragments, or reconstruct those links by generating synthetic entity triples, an expensive process that introduces hallucinated graph edges. To address this, we propose BRIDGE (Bridging Retrieval with Inter-Document Graph Expansion). This four-stage passage-retrieval pipeline exploits inter-document reference signals to recover multi-hop evidence chains without requiring the construction of graphs from expensive pretrained language model calls. To bridge the gap between flat retrieval and complex query resolution, the pipeline begins by fusing lexical term frequency and the dense embedding signals via Reciprocal Rank Fusion using an INT8-quantised vector index that compresses the embedding store by $4\times $ . It then applies a cross-encoder model as an initial precision gate to discover supporting passages that share a minimal lexical overlap with the query. And subsequently, the system traverses an inter-document reference graph anchored on cross-reference structures using Personalised PageRank. Finally, a cross-encoder reranking is performed on the merged candidate pool to ensure consistent scoring across both retrieved and graph-expanded passages. We evaluate BRIDGE on three established multi-hop QA benchmarks: HotpotQA, MuSiQue, and 2WikiMultiHopQA. Without any dataset-specific tuning, our approach achieves Recall@2 of 79.8, 41.5, and 65.7, respectively, matching or exceeding baselines including HippoRAG (59.0, 41.0, 71.5) and ColBERTv2 (64.7, 37.9, 59.2) on two of three benchmarks. Our ablation analysis confirms that reference-graph expansion discovers an average of 7.0 additional relevant passages per query, while the final cross-encoder reranking prevents score displacement and maintains top-of-list precision.
Click-through rate (CTR) prediction is a critical task in personalized recommender systems. Existing methods that align collaborative information from conventional CTR models with semantic information from pre-trained language models (PLMs) have demonstrated superior performance compared to approaches relying on a single information source. However, most of them perform alignment at the embedding level, which introduces noise from heterogeneous vector spaces and limits fine-grained semantic mapping. Moreover, these models highly depend on tabular features, thereby limiting their transferability. To address these challenges, we propose to conduct Multi-scale Graph Tokens Alignment (MGTA) for CTR prediction via pre-trained language models, which enables deep cross-modal information alignment while maintaining strong generalizability. Specifically, MGTA first captures multi-scale graph tokens rich in collaborative signals by decoupling and quantizing graph structures based on graph neural networks (GNNs), and then achieves token-level alignment between collaborative signals and semantic knowledge via PLM fine-tuning. To achieve efficient transfer with MGTA, we further introduce the Cross-domain Token Adapter that enables collaborative signals adaptation by mapping graph tokens from the target domain to the source domain, which necessitates only the injection of target-domain semantic knowledge, in turn reducing fine-tuning time. Extensive experiments on three real-world datasets demonstrate the effectiveness of MGTA compared to existing baselines.
Zhongzheng Wu, Yating Ren, Shuocheng Li et al.· Proceedings of the 32nd ACM...· 0 citations
This work introduces a GLM-based retriever and investigates the comparative strengths of GLM-based, GNN-based, and traditional vector-search-based retrievers in single- and multi-hop RAG settings, and suggests that finetuned GLM retrievers generalize better out of domain.
