1 - 10 of 579
-
PublicationTopic modelling for spatial insights: Uncovering space use from movement data( 2024-08-01)We present a novel approach to understanding space use by moving entities based on repeated patterns of place visits and transitions. Our approach represents trajectories as text documents consisting of sequences of place visits or transitions and applies topic modelling to the corpus of these documents. The resulting topics represent combinations of places or transitions, respectively, that repeatedly co-occur in trips. Visualisation of the results in the spatial context reveals the regions of place connectivity through movements and the major channels used to traverse the space. This enables understanding of the use of space as a medium for movement. We compare the possibilities provided by topic modelling to alternative approaches exploiting a numeric measure of pairwise connectedness. We have extensively explored the potential of utilising topic modelling by applying our approach to multiple real-world movement data sets with different data collection procedures and varying spatial and temporal properties: GPS road traffic of cars, unconstrained movement on a football pitch, and episodic movement data reflecting social media posting events. The approach successfully demonstrated the ability to uncover meaningful patterns and interesting insights. We thoroughly discuss different aspects of the approach and share the knowledge and experience we have gained with people who might be potentially interested in analysing movement data by means of topic modelling methods.
-
PublicationA global scale comparison of risk aggregation in AI assessment frameworks( 2024-05-06)AI applications bear inherent risks in various risk dimensions, such as insufficient reliability, robustness, fairness or data protection. It is well-known that trade-offs between these dimensions can arise, for example, a highly accurate AI application may reflect unfairness and bias of the real-world data, or may provide hard-to-explain outcomes because of its internal complexity. AI risk assessment frameworks aim to provide systematic approaches to risk assessment in various dimensions. The overall trustworthiness assessment is then generated by some form of risk aggregation among the risk dimensions. This paper provides a systematic overview on risk aggregation schemes used in existing AI risk assessment frameworks, focusing on the question how potential trade-offs among the risk dimensions are incorporated. To this end, we examine how the general risk notion, the application context, the extent of risk quantification, and specific instructions for evaluation may influence overall risk aggregation. We discuss our findings in the current frameworks in terms of whether they provide meaningful and practicable guidance. Lastly, we derive recommendations for the further operationalization of risk aggregation both from horizontal and vertical perspectives.
-
PublicationDeveloping trustworthy AI applications with foundation models( 2024-04)
-
PublicationMachine Learning Operations (MLOps): Grundlagen, Chancen und Herausforderungen beim MLOps-Einsatz in Unternehmen( 2024-04)
-
PublicationControlled Randomness Improves the Performance of Transformer Models( 2024-03-19)
-
PublicationTranscriptome Data Analysis Applied to Grapevine Growth Stage Identification( 2024-03-19)In agricultural production, it is fundamental to characterize the phenological stage of plants to ensure a good evaluation of the development, growth and health of crops. Phenological characterization allows for the early detection of nutritional deficiencies in plants that diminish the growth and productive yield and drastically affect the quality of their fruits. Currently, the phenological estimation of development in grapevine (Vitis vinifera) is carried out using four different schemes: Baillod and Baggiolini, Extended BBCH, Eichhorn and Lorenz, and Modified E-L. Phenological estimation requires the exhaustive evaluation of crops, which makes it intensive in terms of labor, personnel, and the time required for its application. In this work, we propose a new phenological classification based on transcriptional measures of certain genes to accurately estimate the stage of development of grapevine. There are several genomic information databases for Vitis vinifera, and the function of thousands of their genes has been widely characterized. The application of advanced molecular biology, including the massive parallel sequencing of RNA (RNA-seq), and the handling of large volumes of data provide state-of-the-art tools for the determination of phenological stages, on a global scale, of the molecular functions and processes of plants. With this aim, we applied a bioinformatic pipeline for the high-throughput quantification of RNA-seq datasets and further analysis of gene ontology terms. We identified differentially expressed genes in several datasets, and then, we associated them with the corresponding phenological stage of development. Differentially expressed genes were classified using count-based expression analysis and clustering and annotated using gene ontology data. This work contributes to the use of transcriptome data and gene expression analysis for the classification of development in plants, with a wide range of industrial applications in agriculture.
-
PublicationDeep Dynamic Language Models( 2024-03-15)This thesis investigates the domain of deep dynamic language models, focusing on the integration of temporal dynamics to enhance language modeling and its application in various tasks, such as text generation, recommendation systems, and predicting post popularity. Temporal content change, i.e., trends and themes that change with time featured in document collections such as academic journals, news articles and social media, make the traditional static language models (LMs) not an optimal solution. In order to address this limitation, several approaches to develop dynamic LMs are proposed and explored in this thesis. Initially, the impact of incorporating temporal information is explored, specifically in the context of modeling online communities. For the analysis of temporal content change in Yelp - a crowd-sourced review platform - an instantaneous language model is proposed. This model combines a temporal point process (TPP) for modeling review creation times and a LM to capture textual aspects. Empirical evaluations demonstrate that this model significantly improves the performance of LMs in terms of both language modeling and prediction of review creation time. Building upon the success of the instantaneous LM, the research in this thesis is extended to more application oriented task, such as recommender systems. Recognizing that user preferences and item reviews change over time, the proposed model here leverages users’ reviews to enhance rating predictions. By developing time-interval aware representations, the proposed model outperforms several state-of-the-art recommender systems models in real-world datasets. Additionally, the integration of dynamic topic models into LMs is explored. First, the problem of skewed topic distributions in topic modeling is addressed, which can cause models to learn more general topics present in the majority of documents, rather than rare topics present in only a few documents. A neural dynamic focused topic model is proposed as a solution, which decouples topic activities from topic proportions in documents using sequences of Bernoulli random variables. Experimental evaluations show that this model outperforms state-of-the-art topic models in generalization tasks, while employing a comparable number of parameters and converging two times faster. Furthermore, the performance of large pre-trained language models (LPLMs) in dynamic environments is explored. The empirical analysis on Reddit datasets reveals significant performance drops when predicting the popularity of future posts due to temporal distribution shifts in data. To mitigate this issue, a model is proposed that combines neural variational dynamic topic models and attention mechanisms to infer temporal LM representations. The proposed model exhibit improved performance while utilizing only a fraction of the parameters of LPLMs, and providing interpretable representations that offer insights into real-world events. In summary, this thesis emphasizes the significance of incorporating temporal dynamics into LMs and explores their application in various tasks.
-
PublicationWie Agenten und Foundation-Modelle bei der Versorgung Schwerverletzter helfen( 2024-03)
-
PublicationSuperkraft Sprachmodell?( 2024-03)Die deutsche Verwaltungslandschaft steht angesichts der notwendigen Digitalisierung und Automatisierung von bisher manuellen Prozessen vor einer grundlegenden Transformation. Der Anstieg an Anträgen für Wohngeld, BAföG oder Einbürgerungsverfahren setzt Behörden zusätzlich unter Druck. Der entstehende Rückstau trägt zu einem sinkenden Vertrauen in die Leistungsfähigkeit der öffentlichen Verwaltung bei. Gleichzeitig muss sie die rückläufigen Mitarbeitendenzahlen infolge des demografischen Wandels kompensieren. Generative Künstliche Intelligenz (GenAI) und insbesondere große Sprachmodelle (Large Language Models, LLMs) spielen hier eine wichtige Rolle, um die Mitarbeitenden zukünftig in ihren Aufgaben zu unterstützen, zu entlasten und hierdurch Freiräume zu schaffen, um sich verstärkt der direkten Interaktion mit Bürgerinnen und Bürgern zu widmen. In diesem Briefing präsentieren Fraunhofer IAIS und Deloitte drei Anwendungsbeispiele großer Sprachmodelle, von welchen die öffentliche Verwaltung schon heute profitieren kann. Bei der Betrachtung zu etablierender Rahmenbedingungen muss zwischen den behördeninternen Voraussetzungen und der staatlichen Infrastruktur unterschieden werden. Diese Publikation betrachtet die Voraussetzungen auf individueller Ebene der Behörden.
-
PublicationSenti-Sequence: Learning to Represent Texts for Sentiment Polarity Classification( 2024-01-25)The sentiment analysis task seeks to categorize opinionated documents as having overall positive or negative opinions. This task is very important to understand unstructured text content generated by users in different domains, such as online and entertainment platforms and social networks. In this paper, we propose a novel method for predicting the overall polarity in texts. First, a new polarity-aware vector representation is automatically built for each document. Then, a bidirectional recurrent neural architecture is designed to identify the emerging polarity. The attained results outperform all of the algorithms found in the literature in the binary polarity classification task.
