TCAR-Gen: Temporal Graph Retrieval with Evidence Fusion for Knowledge-Grounded Generation

Sidra Nasir; Muhammad Noman Zahid; Rizwan Ahmed Khan

cs.CLApr 15, 2026

TCAR-Gen: Temporal Graph Retrieval with Evidence Fusion for Knowledge-Grounded Generation

Authors: Sidra Nasir, Muhammad Noman Zahid, Rizwan Ahmed Khan

Organizations: Dipartimento di Informatica, Università di Verona, Italy · School of Advanced Studies, University of Camerino, Italy · Department of Computer Science, School of Mathematics and Computer Science, Institute of Business Administration (IBA), Karachi, Pakistan

Abstract

Retrieval-augmented generation systems struggle with temporal reasoning and evidence fusion when answering complex questions over historical criminal case narratives. Existing approaches either retrieve independently of query semantics or fail to integrate multiple evidence sources coherently. We propose Temporal Context Augmented Retrieval Generation (TCAR-Gen), a framework that combines query-conditioned graph neural networks, temporal evidence fusion, and chain-of-trees reasoning to ground answer generation in retrieved evidence. On the Victorian Crime Diaries benchmark, TCAR-Gen achieves 0.3738 Recall@5, outperforming Vanilla RAG, Temporal RAG, GraphRAG-C, and GraphRAG-T across seven query types including multi-hop reasoning and counterfactual questions. Ablation studies reveal that the context graph, temporal penalty mechanism, and query conditioning are critical components. Cross-model evaluation across five language model (GPT-OSS 20B to TinyLlama 1.1B) demonstrates that TCAR-Gen maintains robust retrieval coverage at smaller model scales, though generation quality degrades substantially with reduced model capacity. Our work shows that explicit temporal modelling and multi-branch evidence fusion are essential for faithful, reasoning-intensive question answering over knowledge-grounded corpora.

Figures & tables

Figure 1: Schematic overview of prior work—covering text-based RAG, graph-based reasoning, temporal KG, and LLM inference and the positioning of TCAR-Gen

Figure 2: Overview of the proposed Temporal Context Augmented Retrieval Generation (TCAR-Gen) framework for temporally grounded, graph-aware retrieval and reasoning.

Figure 3: Global temporal knowledge graph and query-specific subgraph with query-conditioned attention. The top panel shows the full Victorian Crime Diaries knowledge graph with typed entities, relations, and temporal links across cases. The bottom panel shows the query-induced subgraph for a poisoning–fraud comparison query, where node relevance scores ( $\beta$ ) highlight the most important entities and retrieved evidence chunks used for reasoning.

Figure 4: Comparative evaluation of TCAR-Gen and four baselines on the Victorian Crime Diaries benchmark.

Query Type	Hops	TCAR-Gen	Vanilla RAG	Temporal RAG	GraphRAG-C	GraphRAG-T	Best Baseline	Margin
Multi-Hop	3	0.350	0.000	0.000	0.000	0.000	—	$+0.350$
Entity Lookup	1	0.500	0.000	0.000	0.200	0.000	GraphRAG-C	$+0.300$
Counterfactual	2	0.350	0.100	0.100	0.050	0.100	Vanilla RAG	$+0.250$
Evidence	1	0.300	0.100	0.100	0.000	0.100	Vanilla RAG	$+0.200$
Temporal	2	0.267	0.000	0.000	0.100	0.000	GraphRAG-C	$+0.167$
Narrative	1	0.300	0.067	0.067	0.167	0.067	GraphRAG-C	$+0.133$

Table 1: Recall@5 disaggregated by query type for TCAR-Gen and all baselines, ordered by descending absolute margin over the best baseline. Bold : best result per row. Reasoning hops and difficulty level are defined in Section 4.2 .

Configuration	Ablated	Recall@5	$\Delta$ Recall@5
Full TCAR-Gen $\star$	All	0.3738	—
Best Baseline (GraphRAG-C)	Baseline	0.0738	$-0.250$
A1: Entity-only Context Graph	Context Graph	0.3524	$-0.021$
A2: No Context Graph	Context Graph	0.0000	$-0.374$
A6: Semantic Only	Hybrid Retrieval	0.3429	$-0.031$
A7: Semantic + Graph	Hybrid Retrieval	0.3429	$-0.031$

Table 2: Ablation study on TCAR-Gen (GPT-OSS-20B). Upper panel: retrieval-focused ablations (Recall@5). Lower panel: generation-focused ablations (Faithfulness).

\star

denotes the full system. Bold : best result per column.

Retrieval Metrics
Metric	Stage	GPT-OSS 20B	LLaMA-3.1 13B	Mistral 7B	Phi-3-mini 3.8B	TinyLlama 1.1B
Recall@3	Retrieval	0.1857	0.1689	0.1474	0.1289	0.0981
Recall@5	Retrieval	0.3738	0.2945	0.2570	0.2247	0.1711
Recall@10	Retrieval	0.5476	0.4981	0.4347	0.3800	0.2893
NDCG@5	Retrieval	0.2354	0.2141	0.1869	0.1634	0.1244
MRR	Retrieval	0.3357	0.3053	0.2665	0.2330	0.1773

Table 3: TCAR-Gen performance across five LLM models. Bold : best result per row.

Figure 5: Radar Chart: Recall@5 by query type across five LLM models.

Graph Retrieval-Augmented Generation

Graph Retrieval-Augmented Generation connects language models to knowledge graphs to retrieve relational evidence for grounded, multi-hop answers.

See all 232 papers

IA-RAG: Interval-Algebra-Driven Temporal Reasoning for Dynamic Knowledge Retrieval
Jun 4, 2026Xiaoman Wang, Yaoze Zhang, Wenzhuo Fan +7Retrieval-Augmented Generation Temporal IR
East China Normal University · Shanghai Artificial Intelligence Laboratory · University of Shanghai for Science and Technology +1
CARA: Causally Guided Retrieval Agent for Question Answering
Oct 15, 2026Weicheng Liu, Xiangcheng Wu, Depeng Xu +2Agentic Retrieval Graph-Based Retrieval
University of North Carolina at Charlotte
AutoPrunedRetriever: Persistent Reasoning Graphs for Efficient and Extensible Retrieval-Augmented Generation
Feb 4, 2026Ning Wang, Kuanyan Zhu, Daniel Yuehwoon Yee +4RAG Context Compression Retrieval-Augmented Generation
Cornell University · University of Cambridge · The University of Hong Kong +2

Jun 4, 2026J/K move · Enter open · S save

IA-RAG: Interval-Algebra-Driven Temporal Reasoning for Dynamic Knowledge Retrieval

Xiaoman Wang, Yaoze Zhang, Wenzhuo Fan, Hongwei Zhang+6

East China Normal University · Shanghai Artificial Intelligence Laboratory · University of Shanghai for Science and Technology+1

Retrieval-Augmented Generation Temporal IR Question Answering Temporal KGs Temporal QA

Open paper AskPDF arXiv

Retrieval-Augmented Generation (RAG) has shown strong effectiveness in grounding Large Language Models (LLMs) with external knowledge. However, existing RAG and Graph RAG frameworks largely treat knowledge as static or associate time with coarse-grained timestamps or metadata, failing to capture rich temporal structures such as duration, overlap, and containment. We propose IA-RAG, a hierarchical temporal RAG framework that models knowledge as time intervals and performs retrieval under formal temporal constraints. IA-RAG represents facts as Interval Event Units (IEUs) and organizes them into a hierarchical Thematic Forest, where temporal dependencies are governed by Allen's Interval Algebra. To handle incomplete or uncertain temporal boundaries, IA-RAG further introduces a Sub-graph Time Tightening mechanism that refines fuzzy intervals through logical constraints within connected event subgraphs. In addition, IA-RAG supports implicit temporal semantic retrieval through interval-algebra-guided traversal. Experiments on multiple temporal question answering benchmarks, including TimeQA, TempReason, and ComplexTR, demonstrate that IA-RAG achieves strong temporal retrieval and reasoning performance, particularly on complex compositional temporal reasoning tasks. Our code is released at https://github.com/xiaoAugenstern/LogicalRAG_TemporalQA.

Abstract

Graph Retrieval-Augmented Generation

Explore similar work

IA-RAG: Interval-Algebra-Driven Temporal Reasoning for Dynamic Knowledge Retrieval

CARA: Causally Guided Retrieval Agent for Question Answering

AutoPrunedRetriever: Persistent Reasoning Graphs for Efficient and Extensible Retrieval-Augmented Generation