Enhancing Document VQA Models via Retrieval-Augmented Generation

Document Visual Question Answering (Document VQA) must cope with documents that span dozens of pages, yet leading systems still concatenate every page or rely on very large vision-language models, both of which are memory-hungry. Retrieval-Augmented Generation (RAG) offers an attractive alternative, first retrieving a concise set of relevant segments before generating answers from this selected evidence. In this paper, we systematically evaluate the impact of incorporating RAG into Document VQA through different retrieval variants - text-based retrieval using OCR tokens and purely visual retrieval without OCR - across multiple models and benchmarks. Evaluated on the multi-page datasets MP-DocVQA, DUDE, and InfographicVQA, the text-centric variant improves the "concatenate-all-pages" baseline by up to +22.5 ANLS, while the visual variant achieves +5.0 ANLS improvement without requiring any text extraction. An ablation confirms that retrieval and reranking components drive most of the gain, whereas the layout-guided chunking strategy - proposed in several recent works to leverage page structure - fails to help on these datasets. Our experiments demonstrate that careful evidence selection consistently boosts accuracy across multiple model sizes and multi-page benchmarks, underscoring its practical value for real-world Document VQA.

Few-Shot Connectivity-Aware Text Line Segmentation in Historical Documents

A foundational task for the digital analysis of documents is text line segmentation. However, automating this process with deep learning models is challenging because it requires large, annotated datasets that are often unavailable for historical documents. Additionally, the annotation process is a labor- and cost-intensive task that requires expert knowledge, which makes few-shot learning a promising direction for reducing data requirements. In this work, we demonstrate that small and simple architectures, coupled with a topology-aware loss function, are more accurate and data-efficient than more complex alternatives. We pair a lightweight UNet++ with a connectivity-aware loss, initially developed for neuron morphology, which explicitly penalizes structural errors like line fragmentation and unintended line merges. To increase our limited data, we train on small patches extracted from a mere three annotated pages per manuscript. Our methodology significantly improves upon the current state-of-the-art on the U-DIADS-TL dataset, with a 200% increase in Recognition Accuracy and a 75% increase in Line Intersection over Union. Our method also achieves an F-Measure score on par with or even exceeding that of the competition winner of the DIVA-HisDB baseline detection task, all while requiring only three annotated pages, exemplifying the efficacy of our approach. Our implementation is publicly available at: https://github.com/RafaelSterzinger/acpr_few_shot_hist.

Exposing Privacy Risks in Graph Retrieval-Augmented Generation

Retrieval-Augmented Generation (RAG) is a powerful technique for enhancing Large Language Models (LLMs) with external, up-to-date knowledge. Graph RAG has emerged as an advanced paradigm that leverages graph-based knowledge structures to provide more coherent and contextually rich answers. However, the move from plain document retrieval to structured graph traversal introduces new, under-explored privacy risks. This paper investigates the data extraction vulnerabilities of the Graph RAG systems. We design and execute tailored data extraction attacks to probe their susceptibility to leaking both raw text and structured data, such as entities and their relationships. Our findings reveal a critical trade-off: while Graph RAG systems may reduce raw text leakage, they are significantly more vulnerable to the extraction of structured entity and relationship information. We also explore potential defense mechanisms to mitigate these novel attack surfaces. This work provides a foundational analysis of the unique privacy challenges in Graph RAG and offers insights for building more secure systems.

Named Entity Recognition of Historical Text via Large Language Model

Large language models have demonstrated remarkable versatility across a wide range of natural language processing tasks and domains. One such task is Named Entity Recognition (NER), which involves identifying and classifying proper names in text, such as people, organizations, locations, dates, and other specific entities. NER plays a crucial role in extracting information from unstructured textual data, enabling downstream applications such as information retrieval from unstructured text. Traditionally, NER is addressed using supervised machine learning approaches, which require large amounts of annotated training data. However, historical texts present a unique challenge, as the annotated datasets are often scarce or nonexistent, due to the high cost and expertise required for manual labeling. In addition, the variability and noise inherent in historical language, such as inconsistent spelling and archaic vocabulary, further complicate the development of reliable NER systems for these sources. In this study, we explore the feasibility of applying LLMs to NER in historical documents using zero-shot and few-shot prompting strategies, which require little to no task-specific training data. Our experiments, conducted on the HIPE-2022 (Identifying Historical People, Places and other Entities) dataset, show that LLMs can achieve reasonably strong performance on NER tasks in this setting. While their performance falls short of fully supervised models trained on domain-specific annotations, the results are nevertheless promising. These findings suggest that LLMs offer a viable and efficient alternative for information extraction in low-resource or historically significant corpora, where traditional supervised methods are infeasible.

mKG-RAG: Multimodal Knowledge Graph-Enhanced RAG for Visual Question Answering

Recently, Retrieval-Augmented Generation (RAG) has been proposed to expand internal knowledge of Multimodal Large Language Models (MLLMs) by incorporating external knowledge databases into the generation process, which is widely used for knowledge-based Visual Question Answering (VQA) tasks. Despite impressive advancements, vanilla RAG-based VQA methods that rely on unstructured documents and overlook the structural relationships among knowledge elements frequently introduce irrelevant or misleading content, reducing answer accuracy and reliability. To overcome these challenges, a promising solution is to integrate multimodal knowledge graphs (KGs) into RAG-based VQA frameworks to enhance the generation by introducing structured multimodal knowledge. Therefore, in this paper, we propose a novel multimodal knowledge-augmented generation framework (mKG-RAG) based on multimodal KGs for knowledge-intensive VQA tasks. Specifically, our approach leverages MLLM-powered keyword extraction and vision-text matching to distill semantically consistent and modality-aligned entities/relationships from multimodal documents, constructing high-quality multimodal KGs as structured knowledge representations. In addition, a dual-stage retrieval strategy equipped with a question-aware multimodal retriever is introduced to improve retrieval efficiency while refining precision. Comprehensive experiments demonstrate that our approach significantly outperforms existing methods, setting a new state-of-the-art for knowledge-based VQA.

GHTM: A Graph based Hybrid Topic Modeling Approach in Low-Resource Bengali Language

Topic modeling is a Natural Language Processing (NLP) technique that is used to identify latent themes and extract topics from text corpora by grouping similar documents based on their most significant keywords. Although widely researched in English, topic modeling remains understudied in Bengali due to its morphological complexity, lack of adequate resources and initiatives. In this contribution, a novel Graph Convolutional Network (GCN) based model called GHTM (Graph-Based Hybrid Topic Model) is proposed. This model represents input vectors of documents as nodes in the graph, which GCN uses to produce semantically rich embeddings. The embeddings are then decomposed using Non-negative Matrix Factorization (NMF) to get the topical representations of the underlying themes of the text corpus. This study compares the proposed model against a wide range of Bengali topic modeling techniques, from traditional methods such as LDA, LSA, and NMF to contemporary frameworks such as BERTopic and Top2Vec on three Bengali datasets. The experimental results demonstrate the effectiveness of the proposed model by outperforming other models in topic coherence and diversity. In addition, we introduce a novel Bengali dataset called "NCTBText" sourced from Bengali textbook materials to enrich and diversify the predominantly newspaper-centric Bengali corpora.

Combining Language and Topic Models for Hierarchical Text Classification

Hierarchical text classification (HTC) is a natural language processing task which has the objective of categorising text documents into a set of classes from a predefined structured class hierarchy. Recent HTC approaches use various techniques to incorporate the hierarchical class structure information with the natural language understanding capabilities of pre-trained language models (PLMs) to improve classification performance. Furthermore, using topic models along with PLMs to extract features from text documents has been shown to be an effective approach for multi-label text classification tasks. The rationale behind the combination of these feature extractor models is that the PLM captures the finer-grained contextual and semantic information while the topic model obtains high-level representations which consider the corpus of documents as a whole. In this paper, we use a HTC approach which uses a PLM and a topic model to extract features from text documents which are used to train a classification model. Our objective is to determine whether the combination of the features extracted from the two models is beneficial to HTC performance in general. In our approach, the extracted features are passed through separate convolutional layers whose outputs are combined and passed to a label-wise attention mechanisms which obtains label-specific document representations by weighing the most important features for each class separately. We perform comprehensive experiments on three HTC benchmark datasets and show that using the features extracted from the topic model generally decreases classification performance compared to only using the features obtained by the PLM. In contrast to previous work, this shows that the incorporation of features extracted from topic models for text classification tasks should not be assumed beneficial.

Millions of $\text{GeAR}$-s: Extending GraphRAG to Millions of Documents

Recent studies have explored graph-based approaches to retrieval-augmented generation, leveraging structured or semi-structured information -- such as entities and their relations extracted from documents -- to enhance retrieval. However, these methods are typically designed to address specific tasks, such as multi-hop question answering and query-focused summarisation, and therefore, there is limited evidence of their general applicability across broader datasets. In this paper, we aim to adapt a state-of-the-art graph-based RAG solution: $\text{GeAR}$ and explore its performance and limitations on the SIGIR 2025 LiveRAG Challenge.

Exploring text-to-image generation for historical document image retrieval

Attribute-based document image retrieval (ABDIR) was recently proposed as an alternative to query-by-example (QBE) searches, the dominant document image retrieval (DIR) paradigm. One drawback of QBE searches is that they require sample query documents on hand that may not be available. ABDIR aims to offer users a flexible way to retrieve document images based on memorable visual features of document contents, describing document images with combinations of visual attributes determined via convolutional neural network (CNN)-based binary classifiers. We present an exploratory study of the use of generative AI to bridge the gap between QBE and ABDIR, focusing on historical documents as a use case for their diversity and uniqueness in visual features. We hypothesize that text-to-image (T2I) generation can be leveraged to create query document images using text prompts based on ABDIR-like attributes. We propose T2I-QBE, which uses Leonardo.Ai as the T2I generator with prompts that include a rough description of the desired document type and a list of the desired ABDIR-style attributes. This creates query images that are then used within the traditional QBE paradigm, which compares CNN-extracted query features to those of the document images in the dataset to retrieve the most relevant documents. Experiments on the HisIR19 dataset of historical documents confirm our hypothesis and suggest that T2I-QBE is a viable option for historical document image retrieval. To the authors' knowledge, this is the first attempt at utilizing T2I generation for DIR.

From Chaos to Automation: Enabling the Use of Unstructured Data for Robotic Process Automation

The growing volume of unstructured data within organizations poses significant challenges for data analysis and process automation. Unstructured data, which lacks a predefined format, encompasses various forms such as emails, reports, and scans. It is estimated to constitute approximately 80% of enterprise data. Despite the valuable insights it can offer, extracting meaningful information from unstructured data is more complex compared to structured data. Robotic Process Automation (RPA) has gained popularity for automating repetitive tasks, improving efficiency, and reducing errors. However, RPA is traditionally reliant on structured data, limiting its application to processes involving unstructured documents. This study addresses this limitation by developing the UNstructured Document REtrieval SyStem (UNDRESS), a system that uses fuzzy regular expressions, techniques for natural language processing, and large language models to enable RPA platforms to effectively retrieve information from unstructured documents. The research involved the design and development of a prototype system, and its subsequent evaluation based on text extraction and information retrieval performance. The results demonstrate the effectiveness of UNDRESS in enhancing RPA capabilities for unstructured data, providing a significant advancement in the field. The findings suggest that this system could facilitate broader RPA adoption across processes traditionally hindered by unstructured data, thereby improving overall business process efficiency.