Chronological Knowledge Retrieval: A Retrieval-Augmented Generation Approach to Construction Project Documentation

Ioannis-Aris Kostis; Natalia Sanchiz; Steeve De Schryver; François Denis; Pierre Schaus

cs.CLMar 25, 2026

Chronological Knowledge Retrieval: A Retrieval-Augmented Generation Approach to Construction Project Documentation

Authors: Ioannis-Aris Kostis, Natalia Sanchiz, Steeve De Schryver, François Denis, Pierre Schaus

Organizations: Université Catholique de Louvain, ICTEAM/INGI, Place Sainte Barbe 2, Louvain-la-Neuve, 1348, Belgium · Buildwise, Avenue P. Holoﬀe 21, Limelette, 1342, Belgium · BPC Group, Avenue Edmond Van Nieuwenhuyse 30, Brussels, 1160, Belgium

Abstract

In large-scale construction projects, the continuous evolution of decisions generates extensive records, most often captured in meeting minutes. Since decisions may override previous ones, professionals often need to reconstruct the history of specific choices. Retrieving such information manually from raw archives is both labor-intensive and error-prone. From a user perspective, we address this challenge by enabling conversational access to the whole set of project meeting minutes. Professionals can pose natural-language questions and receive answers that are both semantically relevant and explicitly time-annotated, allowing them to follow the chronology of decisions. From a technical perspective, our solution employs a Retrieval-Augmented Generation (RAG) framework that integrates semantic search with large language models to ensure accurate and context-aware responses. We demonstrate the approach using an anonymized, industry-sourced dataset of meeting minutes from a completed construction project by a large company in Belgium. The dataset is annotated and enriched with expert-defined queries to support systematic evaluation. Both the dataset and the open-source implementation are made available to the community to foster further research on conversational access to time-annotated project documentation.

Figures & tables

Panel of Figure 2 — Figure 2: Dataset composition and temporal statistics. (a)-(c) Summarize document structure: page, passage, and word counts. (d) Illustrates the temporal distribution of documents over the project’s time span.

Figure 3: Evolution of a subject in the meeting minutes . The initial decision is amended over meetings and truncated for brevity toward the end of the project. Decision text translated from French.

Figure 4: System implementation overview . Schematic representation of the implemented system designed to meet the use case specifications. Dashed lines denote offline operations, while solid lines represent online processes. Modules enclosed within the gray dotted frame are optional and not essential for core system functionality.

$n_{\text{batch}}$	$t_{\text{retrieve}}$	$t_{\text{rerank}}$	$t_{\text{total}}$	P@5	R@5	F1@5
1	5.77	190.9	275.4	$0.34\pm 0.10$	$0.92\pm 0.09$	$0.48\pm 0.11$
2	3.19	95.98	133.3	$0.49\pm 0.15$	$0.85\pm 0.12$	$0.59\pm 0.13$
6	1.42	31.13	55.2	$0.73\pm 0.20$	$0.57\pm 0.17$	$0.60\pm 0.18$
10	1.08	19.48	27.9	$0.71\pm 0.23$	$0.43\pm 0.15$	$0.47\pm 0.14$
12	0.98	16.01	22.8	$0.72\pm 0.15$	$0.43\pm 0.16$	$0.47\pm 0.15$
30	0.72	6.15	10.6	$0.88\pm 0.20$	$0.27\pm 0.21$	$0.35\pm 0.19$

Table 1: Execution times and retrieval performance as a function of batch size

n_{\text{batch}}

. Times (

t_{\text{retrieve}}

t_{\text{rerank}}

, and

t_{\text{total}}

) are in seconds and averaged over 8 queries. Precision@5 (P@5), Recall@5 (R@5), and F1@5 metrics summarize the retrieval quality.

Figure 5: System Graphical User Interface . Screenshot of the application interface showing the bottom input box for user queries, the left panel with the decision timeline and system replies, and the right panel displaying retrieved source pages grouped by time span and document.

Figure 6: Query - document relevance heatmap . Heatmap showing the distribution of pages containing relevant passages for each query in the benchmarking set. Rows correspond to evaluated queries, and columns represent documents ordered chronologically (dates shown on the top axis). Color intensity indicates the number of relevant pages cited per document.

		$t_{\text{total}}$ (s)		P@5		R@5		F1@5
	Rerank:	Off	top-5	Off	top-5	Off	top-5	Off	top-5
	10	8.61	16.16	0.68	0.74	0.40	0.46	0.45	0.50
				$\pm$ 0.19	$\pm$ 0.23	$\pm$ 0.11	$\pm$ 0.14	$\pm$ 0.10	$\pm$ 0.15
top- $k$	20	12.28	26.28	0.66	0.73	0.39	0.45	0.44	0.49
				$\pm$ 0.20	$\pm$ 0.22	$\pm$ 0.09	$\pm$ 0.14	$\pm$ 0.10	$\pm$ 0.14
	50	24.18	57.38	0.69	0.72	0.42	0.41	0.46	0.47

Table 2: Comparison of runtime and retrieval performance for different top-

k

values . Batch size

n_{\text{batch}}=10

. Precision, Recall, and F1 metrics include standard deviation (reported below the mean).

Panel of Figure 7 — Figure 7: Retrieval performance metrics at query and global levels . Evaluation of retrieval metrics at different cutoff values $k_{\text{eval}}=2,3,4,5$ , comparing system performance across queries and in aggregate.

Appendix figures & tables1 asset

Supplementary material from the paper’s appendix.

Appendix

Figure 9: System output comparison for different batch sizes . Comparison of system-generated outputs for batch sizes

n_{\texttt{batch}}=6,10,\text{and }60

. Each sample reply illustrates how varying the batch size affects the completeness and consistency of the retrieved and synthesized responses.

Document QA

Document question answering uses AI to answer questions from text, tables, images, and structured content in documents.

See all 92 papers

Topic-to-Timestamp Alignment by Constrained Evidence Selection
Jun 18, 2026Zeynep Yılbırt, Marina Litvak, Michael FärberRetrieval-Augmented Generation Temporal IR
TU Dresden, Germany · SpeechMind GmbH, Germany · Shamoon College of Engineering, Israel +1
TimelineAgent: An LLM-based Framework for Query-Centric Event Timeline Extraction Across Multiple Documents
Oct 19, 2026Qingkai Min, Yunxin Bao, Yifeng Lu +4Event Extraction LLM Information Extraction
Zhejiang University · School of Engineering, Westlake University · Institute of Advanced Technology, Westlake Institute for Advanced Study +3
GADR: Gathering Architecture Decision Records from Meeting Transcriptions
Aug 18, 2026Lucas Daniel Costa da Silva, Kiev GamaRAG Citation Faithfulness Document Information Extraction
Universidade Federal de Pernambuco (UFPE) Recife, Brasil

Jun 18, 2026J/K move · Enter open · S save

Topic-to-Timestamp Alignment by Constrained Evidence Selection

Zeynep Yılbırt, Marina Litvak, Michael Färber

TU Dresden, Germany · SpeechMind GmbH, Germany · Shamoon College of Engineering, Israel+1

Retrieval-Augmented Generation Temporal IR Information Retrieval

Open paper AskPDF arXiv

Meeting archives are difficult to search when users remember what was discussed but not when. We study topic-to-timestamp alignment: given a natural-language topic and a timestamped meeting transcript, the goal is to return the time at which the topic is discussed. A standard RAG setup can retrieve relevant transcript excerpts, but still asks the language model to generate a timestamp, which can produce unsupported or invalid timecodes. We therefore recast timestamp prediction as constrained temporal candidate selection: the system retrieves timestamped transcript chunks, and the model selects the candidate that best grounds the topic instead of generating a timecode. On 420 topic-timestamp queries from 200 municipal meeting transcripts, this increases Recall@5 from 31.9% to 50.0%, reduces MAE from 837.0 seconds to 761.0 seconds with Mistral-7B-Instruct, and increases the number of parseable outputs from 373 to 419 of 420 queries. The results suggest that temporal grounding in long transcripts depends strongly on retrieval quality and output design, not only on the choice of the language model.

Abstract

Document QA

Explore similar work

Topic-to-Timestamp Alignment by Constrained Evidence Selection

TimelineAgent: An LLM-based Framework for Query-Centric Event Timeline Extraction Across Multiple Documents

GADR: Gathering Architecture Decision Records from Meeting Transcriptions