Contributions published at Dynamic and Distributed Information Systems (Abraham Bernstein)

The Semantic Web is distributed yet interoperable: Distributed since resources are created and published by a variety of producers, tailored to their specific needs and knowledge; Interoperable as entities are linked across resources, allowing to use resources from different providers in concord. Complementary to the explicit usage of Semantic Web resources, embedding methods made them applicable to machine learning tasks. Subsequently, embedding models for numerous tasks and structures have been developed, and embedding spaces for various resources have been published. The ecosystem of embedding spaces is distributed but not interoperable: Entity embeddings are not readily comparable across different spaces. To parallel the Web of Data with a Web of Embeddings, we must thus integrate available embedding spaces into a uniform space. Current integration approaches are limited to two spaces and presume that both of them were embedded with the same method — both assumptions are unlikely to hold in the context of a Web of Embeddings. In this paper, we present FedCoder— an approach that integrates multiple embedding spaces via a latent space. We assert that linked entities have a similar representation in the latent space so that entities become comparable across embedding spaces. FedCoder employs an autoencoder to learn this latent space from linked as well as non-linked entities. Our experiments show that FedCoder substantially outperforms state-of-the-art approaches when faced with different embedding models, that it scales better than previous methods in the number of embedding spaces, and that it improves with more graphs being integrated whilst performing comparably with current approaches that assumed joint learning of the embeddings and were, usually, limited to two sources. Our results demonstrate that FedCoder is well adapted to integrate the distributed, diverse, and large ecosystem of embeddings spaces into an interoperable Web of Embeddings.

The automatic annotation of higher-level semantic information in long-form video content is still a challenging task. The Deep Video Understanding (DVU) Challenge aims at catalyzing progress in this area by offering common data and tasks. In this paper, we present our contribution to the 3rd DVU challenge. Our approach consists of multiple information streams extracted from both the visual and the audio modality. The streams can build on information generated by previous streams to increase their semantic descriptiveness. Finally, the output of all streams can be aggregated in order to produce a graph representation of the input movie to represent the semantic relationships between the relevant characters.

During the last 10 years of Video Browser Showdown (VBS), there were many different approaches tested for known-item search and ad-hoc search tasks. Undoubtedly, teams incorporating state-of-the-art models from the machine learning domain had an advantage over teams focusing just on interactive interfaces. On the other hand, VBS results indicate that effective means of interaction with a search system is still necessary to accomplish challenging search tasks. In this tutorial, we summarize successful deep models tested at the Video Browser Showdown as well as interfaces designed on top of corresponding distance/similarity spaces. Our broad experience with competition organization and evaluation will be presented as well, focusing on promising findings and also challenging problems from the most recent iterations of the Video Browser Showdown.

The retrieval of multimedia content remains a difficult problem where a high accuracy or specificity can often only be achieved interactively, with a user working closely and iteratively with a retrieval system. While there exist several venues for the exchange of insights in the area of information retrieval in general and multimedia retrieval specifically, there is little discussion on such interactive retrieval approaches. The Workshop on Interactive Multimedia Retrieval offers such a venue. Held for the 2nd time in 2022, it attracted a diverse set of contributions, six of which were accepted for presentation. The following provides a brief overview of the workshop itself as well as the contributions of 2022.

Evaluation campaigns for interactive multimedia retrieval, such as the Video Browser Shodown (VBS) or the Lifelog Search Challenge (LSC), so far imposed constraints on both simultaneity and locality of all participants, requiring them to solve the same tasks in the same place, at the same time and under the same conditions. These constraints are in contrast to other evaluation campaigns that do not focus on interactivity, where participants can process the tasks in any place at any time. The recent travel restrictions necessitated the relaxation of the locality constraint of interactive campaigns, enabling participants to take place from an arbitrary location. Born out of necessity, this relaxation turned out to be a boon since it greatly simplified the evaluation process and enabled organisation of ad-hoc evaluations outside of the large campaigns. However, it also introduced an additional complication in cases where participants were spread over several time zones. In this paper, we introduce an evaluation scheme for interactive retrieval evaluation that relaxes both the simultaneity and locality constraints, enabling participation from any place at any time within a predefined time frame. This scheme, as implemented in the Distributed Retrieval Evaluation Server (DRES), enables novel ways of conducting interactive retrieval evaluation and bridged the gap between interactive campaigns and non-interactive ones.

The Covid 19 pandemic demonstrated the importance of access to data to en- sure that authorities can make informed decisions in the event of a crisis. How- ever, there are currently three types of barriers preventing access and use of data: (i) technical barriers, especially the lack of uniform data formats and semantics; (ii) legal barriers, especially data protection that limits the use of personal data; and (iii) societal barriers, especially the lack of data literacy and trust. This call for action presents ways to overcome these barriers and pro- poses new governance mechanisms for the access and use of data in public health crises.

Scientific datasets often have hierarchical structure: for example, in surveys, individual participants (samples) might be grouped at a higher level (units) such as their geographical region. In these settings, the interest is often in exploring the structure on the unit level rather than on the sample level. Units can be compared based on the distance between their means, however this ignores the within-unit distribution of samples. Here we develop an approach for exploratory analysis of hierarchical datasets using the Wasserstein distance metric that takes into account the shapes of within-unit distributions. We use t-SNE to construct 2D embeddings of the units, based on the matrix of pairwise Wasserstein distances between them. The distance matrix can be efficiently computed by approximating each unit with a Gaussian distribution, but we also provide a scalable method to compute exact Wasserstein distances. We use synthetic data to demonstrate the effectiveness of our Wasserstein t-SNE, and apply it to data from the 2017 German parliamentary election, considering polling stations as samples and voting districts as units. The resulting embedding uncovers meaningful structure in the data.

In this paper, we present the iteration of the multimedia retrieval system vitrivr participating at LSC 2022. vitrivr is a general-purpose retrieval system which has previously participated at LSC. We describe the system architecture and functionality, and show initial results based on the test and validation topics.

For the fifth time since 2018, the Lifelog Search Challenge (LSC) facilitated a benchmarking exercise to compare interactive search systems designed for multimodal lifelogs. LSC'22 attracted nine participating research groups who developed interactive lifelog retrieval systems enabling fast and effective access to lifelogs. The systems competed in front of a hybrid audience at the LSC workshop at ACM ICMR'22. This paper presents an introduction to the LSC workshop, the new (larger) dataset used in the competition, and introduces the participating lifelog search systems.

null

null

null

Answering multi-relation questions over knowledge graphs is a challenging task as it requires multi-step reasoning over a huge number of possible paths. Reasoning-based methods with complex reasoning mechanisms, such as reinforcement learning-based sequential decision making, have been regarded as the default pathway for this task. However, these mechanisms are difficult to implement and train, which hampers their reproducibility and transferability to new domains. In this paper, we propose QAGCN - a simple but effective and novel model that leverages attentional graph convolutional networks that can perform multi-step reasoning during the encoding of knowledge graphs. As a consequence, complex reasoning mechanisms are avoided. In addition, to improve efficiency, we retrieve answers using highly-efficient embedding computations and, for better interpretability, we extract interpretable paths for returned answers. On widely adopted benchmark datasets, the proposed model has been demonstrated competitive against state-of-the-art methods that rely on complex reasoning mechanisms. We also conducted extensive experiments to scrutinize the efficiency and contribution of each component of our model.

Due to the Semantic Web’s decentralised nature, ontology engineers rarely know all applications that leverage their ontology. Consequently, they are unaware of the full extent of possible consequences that changes might cause to the ontology. Our goal is to lessen the gap between ontology engineers and users by investigating ontology engineers’ understanding of ontology changes’ impact at editing time. Hence, this paper introduces the Protégé plugin ChImp which we use to reach our goal. We elicited requirements for ChImp through a questionnaire with ontology engineers. We then developed ChImp according to these requirements and it displays all changes of a given session and provides selected information on said changes and their effects. For each change, it computes a number of metrics on both the ontology and its materialisation. It displays those metrics on both the originally loaded ontology at the beginning of the editing session and the current state to help ontology engineers understand the impact of their changes. We investigated the informativeness of materialisation impact measures, the meaning of severe impact, and also the usefulness of ChImp in an online user study with 36 ontology engineers. We asked the participants to solve two ontology engineering tasks – with and without ChImp (assigned in random order) – and answer in-depth questions about the applied changes as well as the materialisation impact measures. We found that ChImp increased the participants’ understanding of change effects and that they felt better informed. Answers also suggest that the proposed measures were useful and informative. We also learned that the participants consider different outcomes of changes severe, but most would define severity based on the amount of changes to the materialisation compared to its size. The participants also acknowledged the importance of quantifying the impact of changes and that the study will affect their approach of editing ontologies.

High level semantic video understanding deals with the problem of analyzing basic insights from movies like interpersonal relationships, relationships to other entities or interpersonal interactions. The Deep Video Understanding Challenge has focused on this issue and organizes an annual competition in which a set of queries is created which should be answered by the participants. This thesis is written in the context of the Deep Video Understanding Challenge 2021 and describes a pipeline that is able to answer the set of queries on a movie- and scene-level. The pipeline consists of a scene segmentation engine which cuts the scenes into single keyframes and shots. After that, they are processed by two streams, which consists of several feature extraction models. One stream focuses on the visual component, while the other stream focuses on the audio component. After that, the features are combined and processed. Numerous classifiers are trained and used to predict the interpersonal relationships, relationships with other entities or interpersonal interactions. At the movie-level, a knowledge graph is then created, reflecting all the relationships between all the entities of a movie. This is used to answer the queries at movie-level. There, 8% of all questions could be answered correctly. The queries from scene-level could be answered to 1.5% correctly. The other pipelines from the DVU Challenge 2021 achieves better results as the worst result on movie level is 17% of correctly answered queries and the worst result on scene-level is 27% of correctly answered queries.

While artificial intelligence (AI) is increasingly applied for decision- making processes, ethical decisions pose challenges for AI applica- tions. Given that humans cannot always agree on the right thing to do, how would ethical decision-making by AI systems be perceived and how would responsibility be ascribed in human-AI collabora- tion? In this study, we investigate how the expert type (human vs. AI) and level of expert autonomy (adviser vs. decider) influence trust, perceived responsibility, and reliance. We find that partici- pants consider humans to be more morally trustworthy but less capable than their AI equivalent. This shows in participants’ re- liance on AI: AI recommendations and decisions are accepted more often than the human expert’s. However, AI team experts are per- ceived to be less responsible than humans, while programmers and sellers of AI systems are deemed partially responsible instead.

In recent years, top-K recommender systems with implicit feed- back data gained interest in many real-world business scenarios. In particular, neural networks have shown promising results on these tasks. However, while traditional recommender systems are built on datasets with frequent user interactions, insurance recommenders often have access to a very limited amount of user interactions, as people only buy a few insurance products. In this paper, we shed new light on the problem of top-K recommendations for interaction-sparse recommender problems. In particular, we analyze six different recommender algorithms, namely a popularity-based baseline and compare it against two matrix factorization methods (SVD++, ALS), one neural network approach (JCA) and two combinations of neural network and factorization machine approaches (DeepFM, NeuFM). We evaluate these algorithms on six different interaction-sparse datasets and one dataset with a less sparse interaction pattern to elucidate the unique behavior of interaction-sparse datasets. In our experimental evaluation based on real-world insurance data, we demonstrate that DeepFM shows the best performance followed by JCA and SVD++, which indicates that neural network approaches are the dominant technologies. However, for the remaining five datasets we observe a different pattern. Overall, the matrix factorization method SVD++ is the winner. Surprisingly, the simple popularity-based approach comes out second followed by the neural network approach JCA. In summary, our experimental evaluation for interaction-sparse datasets demonstrates that in general matrix factorization methods outperform neural network approaches. As a consequence, traditional well- established methods should be part of the portfolio of algorithms to solve real-world interaction-sparse recommender problems.

Research in the area of human information interaction (HII) typically represents viewpoints on debated topics in a binary fashion, as either against or in favor of a given topic (e.g., the feminist movement). This simple taxonomy, however, greatly reduces the latent richness of viewpoints and thereby limits the potential of research and practical applications in this field. Work in the communication sciences has already demonstrated that viewpoints can be represented in much more comprehensive ways, which could enable a deeper understanding of users’ interactions with debated topics online. For instance, a viewpoint's stance usually has a degree of strength (e.g., mild or strong), and, even if two viewpoints support or oppose something to the same degree, they may use different logics of evaluation (i.e., underlying reasons). In this paper, we draw from communication science practice to propose a novel, two-dimensional way of representing viewpoints that incorporates a viewpoint's stance degree as well as its logic of evaluation. We show in a case study of tweets on debated topics how our proposed viewpoint label can be obtained via crowdsourcing with acceptable reliability. By analyzing the resulting data set and conducting a user study, we further show that the two-dimensional viewpoint representation we propose allows for more meaningful analyses and diversification interventions compared to current approaches. Finally, we discuss what this novel viewpoint label implies for HII research and how obtaining it may be made cheaper in the future.

In this paper, we present BAM, a unified Benchmark for Argument Mining (AM). We propose a method to homogenize both the evaluation process and the data to provide a common view in order to ultimately produce comparable results. Built as a four stage and end-to-end pipeline, the benchmark allows for the direct inclusion of additional argument miners to be evaluated. First, our system pre-processes a ground truth set used both for training and testing. Then, the benchmark calculates a total of four measures to assess different aspects of the mining process. To showcase an initial implementation of our approach, we apply our procedure and evaluate a set of systems on a corpus of scientific publications. With the obtained comparable results we can homogeneously assess the current state of AM in this domain.

The Video Browser Showdown addresses difficult video search challenges through an annual interactive evaluation campaign attracting research teams focusing on interactive video retrieval. The campaign aims to provide insights into the performance of participating interactive video retrieval systems, tested by selected search tasks on large video collections. For the first time in its ten year history, the Video Browser Showdown 2021 was organized in a fully remote setting and hosted a record number of sixteen scoring systems. In this paper, we describe the competition setting, tasks and results and give an overview of state-of-the-art methods used by the competing systems. By looking at query result logs provided by ten systems, we analyze differences in retrieval model performances and browsing times before a correct submission. Through advances in data gathering methodology and tools, we provide a comprehensive analysis of ad-hoc video search tasks, discuss results, task design and methodological challenges. We highlight that almost all top performing systems utilize some sort of joint embedding for text-image retrieval and enable specification of temporal context in queries for known-item search. Whereas a combination of these techniques drive the currently top performing systems, we identify several future challenges for interactive video search engines and the Video Browser Showdown competition itself.