arXiv subjects
Publications and source records attributed to Luca Rossetto.
This report summarizes the contributions and results of the 1st Asynchronous CASTLE Challenge at the Joint Egocentric Vision Workshop in conjunction with CVPR 2026.
Multimodal knowledge graphs typically treat multimedia documents as opaque, external entities. This content-agnostic approach constrains retrieval and analysis by isolating media from the graph's core structure, hindering the ability to capture and query complex relationships across media types. To address this, we introduce the MediaGraph Data Model and its prototypical implementation MeGraS, the MediaGraph Store, a novel approach that integrates multimedia content as graph nodes. This paradigm shift enables the query engine to directly access and process a document's intrinsic content, allowing for native operations such as feature-based similarity search, dynamic segmentation, and the inference of non-materialized relations. By extending the SPARQL query language, MeGraS provides a cohesive platform for the storage, management, and expressive querying of multimodal data. MeGraS is open-source software that establishes a new framework, moving the field toward content-aware multimodal knowledge graphs.
Explanations play a crucial role in creating trustworthy recommender systems (RS), yet choosing a good explanation method presents challenges. Many explanation methods exist, but little guidance exists on which is best for which setting. Existing explanation generation methods often produce abstract outputs that require further formatting to become user-friendly, with a seemingly endless pool of options. Running user-based evaluations of all possible options is usually unfeasible, while automated evaluation metrics often either assess only the explainer's abstract output or require comparison with a ground truth, which is generally unavailable. Recent studies have shown that large language models (LLMs) can serve as ``judges'' for explanation evaluation, but their reliability has not yet been thoroughly explored. This paper studies the utility of LLMs in selecting an effective explanation method for a given application. We first explore their ability to generate explanation prototypes given varying information about the RS and the user. Specifically, we generate 18 distinct explanation prototypes, which are subsequently evaluated by 14 LLMs of varying sizes across two temperature settings. We compare these against human ratings derived from a user study. Our results show that while LLMs exhibit human-like rating patterns and achieve moderate rank correlation with human raters, their absolute rating agreement is low and varies substantially by model size and evaluation construct. We derive four practical recommendations: keep explanation-generation prompts concise, prefer larger models for evaluation, pre-test evaluation constructs, and audit explanations for factual accuracy, as neither humans nor LLMs reliably detect non-factual content.
Multimodal co-embedding models, especially CLIP, have advanced the state of the art in zero-shot classification and multimedia information retrieval in recent years by aligning images and text in a shared representation space. However, such modals trained on a contrastive alignment can lack stability towards small input perturbations. Especially when dealing with manually expressed queries, minor variations in the query can cause large differences in the ranking of the best-matching results. In this paper, we present a systematic analysis of the effect of multiple classes of non-semantic query perturbations in an multimedia information retrieval scenario. We evaluate a diverse set of lexical, syntactic, and semantic perturbations across multiple CLIP variants using the TRECVID Ad-Hoc Video Search queries and the V3C1 video collection. Across models, we find that syntactic and semantic perturbations drive the largest instabilities, while brittleness is concentrated in trivial surface edits such as punctuation and case. Our results highlight robustness as a critical dimension for evaluating vision-language models beyond benchmark accuracy.
This report presents the results of the 14th Video Browser Showdown, held at the 2025 International Conference on Multimedia Modeling on the 8th of January 2025 in Nara, Japan.
The ACM Lifelog Search Challenge (LSC) is a venue that welcomes and compares systems that support the exploration of lifelog data, and in particular the retrieval of specific information, through an interactive competition format. This paper reviews the recent advances in interactive lifelog retrieval as demonstrated at the ACM LSC from 2022 to 2024. Through a detailed comparative analysis, we highlight key improvements across three main retrieval tasks: known-item search, question answering, and ad-hoc search. Our analysis identifies trends such as the widespread adoption of embedding-based retrieval methods (e.g., CLIP, BLIP), increased integration of large language models (LLMs) for conversational retrieval, and continued innovation in multimodal and collaborative search interfaces. We further discuss how specific retrieval techniques and user interface (UI) designs have impacted system performance, emphasizing the importance of balancing retrieval complexity with usability. Our findings indicate that embedding-driven approaches combined with LLMs show promise for lifelog retrieval systems. Likewise, improving UI design can enhance usability and efficiency. Additionally, we recommend reconsidering multi-instance system evaluations within the expert track to better manage variability in user familiarity and configuration effectiveness.
To the best of our knowledge, all existing methods that can generate synthetic brain magnetic resonance imaging (MRI) scans for a specific individual require detailed structural or volumetric information about the individual's brain. However, such brain information is often scarce, expensive, and difficult to obtain. In this paper, we propose the first approach capable of generating synthetic brain MRI segmentations -- specifically, 3D white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF) segmentations -- for individuals using their easily obtainable and often readily available demographic, interview, and cognitive test information. Our approach features a novel deep generative model, CSegSynth, which outperforms existing prominent generative models, including conditional variational autoencoder (C-VAE), conditional generative adversarial network (C-GAN), and conditional latent diffusion model (C-LDM). We demonstrate the high quality of our synthetic segmentations through extensive evaluations. Also, in assessing the effectiveness of the individual-specific generation, we achieve superior volume prediction, with mean absolute errors of only 36.44mL, 29.20mL, and 35.51mL between the ground-truth WM, GM, and CSF volumes of test individuals and those volumes predicted based on generated individual-specific segmentations, respectively.
Retrieval Augmented Generation (RAG) has enjoyed increased attention in the recent past and recent advancements in Large Language Models (LLMs) have highlighted the importance of integrating world knowledge into these systems. Current RAG methodologies often modify the internal architecture of pre-trained language models (PLMs) or rely on textifying knowledge graphs (KGs), which is inefficient in terms of token usage. This paper introduces ConceptFormer, a new approach to augment LLMs with structured knowledge from KGs, such as Wikidata, without altering their internal structure or relying on textual input of KGs. ConceptFormer operates in the LLM embedding vector space, creating and injecting \emph{concept vectors} that encapsulate the information of the KG nodes directly. Trained in conjunction with a frozen LLM, ConceptFormer generates a comprehensive lookup table that maps KG nodes to their respective concept vectors. The approach aims to enhance the factual recall capabilities of LLMs by enabling them to process these concept vectors natively, thus enriching them with structured world knowledge in an efficient and scalable manner. Our experiments demonstrate that the addition of concept vectors to GPT-2 0.1B substantially increases its factual recall ability (Hit@10) by up to 272\% when tested on sentences from Wikipedia and up to 348\% on synthetically generated sentences. Even injecting only a single concept vector into the prompt increases factual recall ability (Hit@10) by up to 213\% on Wikipedia sentences, significantly outperforming RAG with graph textification while consuming 130x fewer input tokens.
Egocentric video has seen increased interest in recent years, as it is used in a range of areas. However, most existing datasets are limited to a single perspective. In this paper, we present the CASTLE 2024 dataset, a multimodal collection containing ego- and exo-centric (i.e., first- and third-person perspective) video and audio from 15 time-aligned sources, as well as other sensor streams and auxiliary data. The dataset was recorded by volunteer participants over four days in a fixed location and includes the point of view of 10 participants, with an additional 5 fixed cameras providing an exocentric perspective. The entire dataset contains over 600 hours of UHD video recorded at 50 frames per second. In contrast to other datasets, CASTLE 2024 does not contain any partial censoring, such as blurred faces or distorted audio. The dataset is available via https://castle-dataset.github.io/.
This report presents the results of the 13th Video Browser Showdown, held at the 2024 International Conference on Multimedia Modeling on the 29th of January 2024 in Amsterdam, the Netherlands.
Adding explanations to recommender systems is said to have multiple benefits, such as increasing user trust or system transparency. Previous work from other application areas suggests that specific user characteristics impact the users' perception of the explanation. However, we rarely find this type of evaluation for recommender systems explanations. This paper addresses this gap by surveying 124 papers in which recommender systems explanations were evaluated in user studies. We analyzed their participant descriptions and study results where the impact of user characteristics on the explanation effects was measured. Our findings suggest that the results from the surveyed studies predominantly cover specific users who do not necessarily represent the users of recommender systems in the evaluation domain. This may seriously hamper the generalizability of any insights we may gain from current studies on explanations in recommender systems. We further find inconsistencies in the data reporting, which impacts the reproducibility of the reported results. Hence, we recommend actions to move toward a more inclusive and reproducible evaluation.
Performance evaluation in multimedia retrieval, as in the information retrieval domain at large, relies heavily on retrieval experiments, employing a broad range of techniques and metrics. These can involve human-in-the-loop and machine-only settings for the retrieval process itself and the subsequent verification of results. Such experiments can be elaborate and use-case-specific, which can make them difficult to compare or replicate. In this paper, we present a formal model to express all relevant aspects of such retrieval experiments, as well as a flexible open-source evaluation infrastructure that implements the model. These contributions intend to make a step towards lowering the hurdles for conducting retrieval experiments and improving their reproducibility.
Interactive video retrieval is a cooperative process between humans and retrieval systems. Large-scale evaluation campaigns, however, often overlook human factors, such as the effects of perception, attention, and memory, when assessing media retrieval systems. Consequently, their setups fall short of emulating realistic retrieval scenarios. In this paper, we design novel task presentation modes based on concepts in media memorability, implement the pipelines necessary for processing target video segments, and build a custom experimental platform for the final evaluation. In order to study the effects of different task representation schemes, we conduct a large crowdsourced experiment. Our findings demonstrate that the way in which the target of a video retrieval task is presented has a substantial influence on the difficulty of the retrieval task and that individuals can successfully retrieve a target video segment despite reducing or even altering the provided hints, opening up a discussion around future evaluation protocols in the domain of interactive media retrieval.
Multi-relation question answering (QA) is a challenging task, where given questions usually require long reasoning chains in KGs that consist of multiple relations. Recently, methods with explicit multi-step reasoning over KGs have been prominently used in this task and have demonstrated promising performance. Examples include methods that perform stepwise label propagation through KG triples and methods that navigate over KG triples based on reinforcement learning. A main weakness of these methods is that their reasoning mechanisms are usually complex and difficult to implement or train. In this paper, we argue that multi-relation QA can be achieved via end-to-end single-step implicit reasoning, which is simpler, more efficient, and easier to adopt. We propose QAGCN -- a Question-Aware Graph Convolutional Network (GCN)-based method that includes a novel GCN architecture with controlled question-dependent message propagation for the implicit reasoning. Extensive experiments have been conducted, where QAGCN achieved competitive and even superior performance compared to state-of-the-art explicit-reasoning methods. Our code and pre-trained models are available in the repository: https://github.com/ruijie-wang-uzh/QAGCN
In recent years, the DBLP computer science bibliography has been prominently used for searching scholarly information, such as publications, scholars, and venues. However, its current search service lacks the capability to handle complex queries, which limits the usability of DBLP. In this paper, we present NLQxform-UI, a web-based natural language interface that enables users to query DBLP directly with complex natural language questions. NLQxform-UI automatically translates given questions into SPARQL queries and executes the queries over the DBLP knowledge graph to retrieve answers. The querying process is presented to users in an interactive manner, which improves the transparency of the system and helps examine the returned answers. Also, intermediate results in the querying process can be previewed and manually altered to improve the accuracy of the system. NLQxform-UI has been completely open-sourced: https://github.com/ruijie-wang-uzh/NLQxform-UI.
Despite the rapid progress of large language models (LLMs), knowledge graph-based question answering (KGQA) remains essential for producing verifiable and hallucination-resistant answers in many real-world settings where answer trustworthiness and computational efficiency are highly valued. However, most existing KGQA methods provide only final answers in the form of KG entities. Without explicit explanations -- ideally in the form of intermediate reasoning process over relevant KG triples, the QA results are difficult to inspect and interpret. Moreover, this limitation prevents the rich and verifiable knowledge encoded in KGs, which is a key advantage of KGQA over LLMs, from being fully leveraged. However, addressing this issue remains highly challenging due to the lack of annotated intermediate reasoning process and the requirement of high efficiency in KGQA. In this paper, we propose a novel Graph Neural Network-based Two-Step Reasoning method (GNN2R) that can efficiently retrieve both final answers and corresponding reasoning subgraphs as verifiable rationales, using only weak supervision from widely-available final answer annotations. We extensively evaluated GNN2R and demonstrated that GNN2R substantially outperforms existing state-of-the-art KGQA methods in terms of effectiveness, efficiency, and the quality of generated explanations. The complete code and pre-trained models are available at https://github.com/ruijie-wang-uzh/GNN2R.
In recent years, scholarly data has grown dramatically in terms of both scale and complexity. It becomes increasingly challenging to retrieve information from scholarly knowledge graphs that include large-scale heterogeneous relationships, such as authorship, affiliation, and citation, between various types of entities, e.g., scholars, papers, and organizations. As part of the Scholarly QALD Challenge, this paper presents a question-answering (QA) system called NLQxform, which provides an easy-to-use natural language interface to facilitate accessing scholarly knowledge graphs. NLQxform allows users to express their complex query intentions in natural language questions. A transformer-based language model, i.e., BART, is employed to translate questions into standard SPARQL queries, which can be evaluated to retrieve the required information. According to the public leaderboard of the Scholarly QALD Challenge at ISWC 2023 (Task 1: DBLP-QUAD - Knowledge Graph Question Answering over DBLP), NLQxform achieved an F1 score of 0.85 and ranked first on the QA task, demonstrating the competitiveness of the system.
This paper presents an analysis of the distribution of spoken language in the V3C video retrieval benchmark dataset based on automatically generated transcripts. It finds that a large portion of the dataset is covered by spoken language. Since language transcripts can be quickly and accurately described, this has implications for retrieval tasks such as known-item search.