Hallucination remains a major obstacle to the domain generalizability and reliability of Large Language Models (LLMs). Recent approaches address this issue by integrating Retrieval-Augmented Generation (RAG) with stepwise reasoning processes to iteratively retrieve knowledge. However, indiscriminate incorporation of external knowledge may interfere with reasoning, increasing latency and amplifying error accumulation. Moreover, existing methods rely on a unidirectional flow of external knowledge into LLMs while neglecting internal–external knowledge synergy, limiting autonomous reasoning capability. To address these limitations, we propose the Reasoning–Evidence Interactive Augmented Graph (RE-IAG), a framework that couples reasoning with evidence through a staged triggering mechanism and structured interaction. RE-IAG performs localized refinement of intermediate reasoning via adaptive branching under uncertainty and selectively triggers retrieval when internal reasoning stagnates. Crucially, it organizes both internal reasoning and retrieved evidence into aligned graph structures, enabling structure-guided verification and fine-grained refinement of intermediate conclusions. This design transforms retrieval from passive augmentation into an active constraint on reasoning, reducing error propagation, alleviating knowledge conflicts, and improving knowledge integration for hallucination mitigation. Extensive experiments on four multi-hop QA benchmarks show that RE-IAG outperforms adaptive RAG baselines, achieves competitive or superior performance to RL-based approaches, and demonstrates strong robustness and generalization across model scales and architectures.

The advancement of wearable intelligent systems presents a unique opportunity to transform how humans interact with digital content. This workshop explores the design of Interactive Multimodal Information Retrieval and Access systems specifically tailored for Heads-Up Computing environments. By leveraging multimodal inputs, such as voice, gaze, and gesture, these systems enable real-time, hands-free access to digital information, facilitating seamless and efficient interaction. The goal is to support tasks requiring rapid information access in dynamic environments while ensuring users remain "heads-up" and engaged with the real world. This half-day workshop will share research outcomes and best practices, foster community building, and facilitate discussions on key challenges. By bringing together researchers and practitioners, it aims to drive further advancements in both research and practical applications within this rapidly evolving field.

Many user behaviour models have been developed with the motivation of getting a better understanding of user behaviour and developing more personalised search services. For example, The ISE Model (Information Goal, Search Strategy, Evaluation Threshold) that was guided by the Information Foraging Theory (IFT), has shown its effectiveness in modelling users based on a small dataset. However, the effectiveness of the ISE model has not been validated in a large dataset. This study, therefore, aims to validate the effectiveness of the ISE model using a larger dataset. Further, based on the findings from further validating the ISE model, a new simplified model called Integrated User Behaviour Analysis Model (IUBAM) is proposed, which shows a new way of categorising users into Depth-Oriented, Breadth-Oriented or Mixed DB search behaviour types. The proposed IUBAM model provide an easier grouping for user behaviour analysis and understanding.

Textual descriptions for multimodal inputs entail recurrent refinement of queries to produce relevant output images. Despite efforts to address challenges such as scaling model size and data volume, the cost associated with pre-training and fine-tuning remains substantial. However, pre-trained large multimodal models (LMMs) encounter a modality gap, characterised by a misalignment between textual and visual representations within a common embedding space. Although fine-tuning can potentially mitigate this gap, it is typically expensive and impractical due to the requirement for extensive domain-driven data. To overcome this challenge, we propose a lightweight training-free approach utilising Retrieval-Augmented Generation (RAG) to extend across the modality using a linear mapping, which can be computed efficiently. Our reproducible code can be found in https://github.com/amitkumarj441/mRAG-gim. During inference, this mapping is applied to images embedded by an LMM enabling retrieval of closest textual descriptions from the training set. These textual descriptions, in conjunction with an instruction, cater as an input prompt for the language model to generate new textual descriptions. In addition, we introduce an iterative technique for distilling the mapping by generating synthetic descriptions via the language model facilitating optimisation for standard utilised image description measures. Experimental results on two benchmark multimodal datasets demonstrate significant improvements.

Conversational image search has emerged as a progressive step beyond traditional keyword-based methodologies, which addresses challenges in human-computer interaction during the information retrieval process. This paper introduces a demonstration called DoodleShoper, a forward-thinking conversational image search assistant centered around sketching, specifically tailored for online product searches. It underscores the importance of visual diversity, often eluding verbal expression while highlighting the efficacy of a sketch-based approach in enhancing user interaction. The proposed modular architecture integrates a state-of-the-art Language Model with advanced Stable Diffusion technologies in the image generation field to offer users a more intuitive and precise conversational search experience. Unlike most conventional methods that directly align prompts or sketches with images, our approach leverages a generative model to produce an intermediate search outcome. This strategic shift streamlines the search process from a zero-shot query - where the query directly corresponds to an image - to a reverse image search task, facilitating the discovery of similar images through multimodal interaction. The implemented demonstration involves refining and expanding the application to diverse user information needs and preferences, including exploring the potential of utilising sketches as an alternative or complementary search environment, a novel concept rooted in current research.

Traditional neural word embeddings are usually dependent on a richer diversity of vocabulary. However, the language models recline to cover major vocabularies via the word embedding parameters, in particular, for multilingual language models that generally cover a significant part of their overall learning parameters. In this work, we present a new compact embedding structure to reduce the memory footprint of the pre-trained language models with a sacrifice of up to 4% absolute accuracy. The embeddings vectors reconstruction follows a set of subspace embeddings and an assignment procedure via the contextual relationship among tokens from pre-trained language models. The subspace embedding structure1 calibrates to masked language models, to evaluate our compact embedding structure on similarity and textual entailment tasks, sentence and paraphrase tasks. Our experimental evaluation shows that the subspace embeddings achieve compression rates beyond 99.8% in comparison with the original embeddings for the language models on XNLI and GLUE benchmark suites.

No abstract available.

The BIRDS workshop aimed to foster the cross-fertilization of Information Science (IS), Information Retrieval (IR) and Data Science (DS). Recognising the commonalities and differences between these communities, the proposed full-day workshop brought together experts and researchers in IS, IR and DS to discuss how they can learn from each other to provide more user-driven data and infor- mation exploration and retrieval solutions. Therefore, the papers aimed to convey ideas on how to utilise, for instance, IS concepts and theories in DS and IR or DS approaches to support users in data and information exploration.