1 - 10 of 47
-
PublicationInspect, Understand, Overcome: A Survey of Practical Methods for AI Safety( 2021)The use of deep neural networks (DNNs) in safety-critical applications like mobile health and autonomous driving is challenging due to numerous model-inherent shortcomings. These shortcomings are diverse and range from a lack of generalization over insufficient interpretability to problems with malicious inputs. Cyber-physical systems employing DNNs are therefore likely to suffer from safety concerns. In recent years, a zoo of state-of-the-art techniques aiming to address these safety concerns has emerged. This work provides a structured and broad overview of them. We first identify categories of insufficiencies to then describe research activities aiming at their detection, quantification, or mitigation. Our paper addresses both machine learning experts and safety engineers: The former ones might profit from the broad range of machine learning (ML) topics covered and discussions on limitations of recent methods. The latter ones might gain insights into the specifics of modern ML methods. We moreover hope that our contribution fuels discussions on desiderata for ML systems and strategies on how to propel existing approaches accordingly.
-
PublicationGrundlagen des Maschinellen Lernens( 2021)Zu definieren, was die menschliche Intelligenz sowie intelligentes Handeln – und damit auch die Künstliche Intelligenz – ausmacht, ist außerordentlich schwer und beschäftigt Philosophen und Psychologen seit Jahrtausenden. Allgemein anerkannt istaber, dass die Fähigkeit zu lernen ein zentrales Merkmal vonIntelligenzist. So ist auchdas Forschungsgebiet desMaschinellen Lernens(engl.machine learning, ML) ein zentraler Teil der Künstlichen Intelligenz, das hinter vielen aktuellen Erfolgen von KI-Systemen steckt.
-
PublicationVisual Analytics in the Aviation and Maritime Domains( 2020)Visual analytics is a research discipline that is based on acknowledging the power and the necessity of the human vision, understanding, and reasoning in data analysis and problem solving. It develops a methodology of analysis that facilitates human activities by means of interactive visual representations of information. By examples from the domains of aviation and maritime transportation, we demonstrate the essence of the visual analytics methods and their utility for investigating properties of available data and analysing data for understanding real-world phenomena and deriving valuable knowledge. We describe four case studies in which distinct kinds of knowledge have been derived from trajectories of vessels and airplanes and related spatial and temporal data by human analytical reasoning empowered by interactive visual interfaces combined with computational operations.
-
PublicationAligning Subjective Ratings in Clinical Decision Making( 2020)In addition to objective indicators (e.g. laboratory values), clinical data often contain subjective evaluations by experts (e.g. disease severity assessments). While objective indicators are more transparent and robust, the subjective evaluation contains a wealth of expert knowledge and intuition. In this work, we demonstrate the potential of pairwise ranking methods to align the subjective evaluation with objective indicators, creating a new score that combines their advantages and facilitates diagnosis. In a case study on patients at risk for developing Psoriatic Arthritis, we illustrate that the resulting score (1) increases classification accuracy when detecting disease presence/absence, (2) is sparse and (3) provides a nuanced assessment of severity for subsequent analysis.
-
PublicationAdvanced Sensing and Human Activity Recognition in Early Intervention and Rehabilitation of Elderly People( 2020)Ageing is associated with a decline in physical activity and a decrease in the ability to perform activities of daily living, affecting physical and mental health. Elderly people or patients could be supported by a human activity recognition (HAR) system that monitors their activity patterns and intervenes in case of change in behavior or a critical event has occurred. A HAR system could enable these people to have a more independent life. In our approach, we apply machine learning methods from the field of human activity recognition (HAR) to detect human activities. These algorithmic methods need a large database with structured datasets that contain human activities. Compared to existing data recording procedures for creating HAR datasets, we present a novel approach, since our target group comprises of elderly and diseased people, who do not possess the same physical condition as young and healthy persons. Since our targeted HAR system aims at supporting elderly and diseased people, we focus on daily activities, especially those to which clinical relevance in attributed, like hygiene activities, nutritional activities or lying positions. Therefore, we propose a methodology for capturing data with elderly and diseased people within a hospital under realistic conditions using wearable and ambient sensors. We describe how this approach is first tested with healthy people in a laboratory environment and then transferred to elderly people and patients in a hospital environment. We also describe the implementation of an activity recognition chain (ARC) that is commonly used to analyse human activity data by means of machine learning methods and aims to detect activity patterns. Finally, the results obtained so far are presented and discussed as well as remaining problems that should be addressed in future research.
-
-
PublicationKünstliche Intelligenz im Krankenhaus: Potenziale und Herausforderungen - Eine Fallstudie im Bereich der NotfallversorgungKünstliche Intelligenz (KI) ist bereits fest in unserem Alltag verankert und ist auch im medizinischen Kontext nicht mehr wegzudenken. Gerade in Zeiten einer globalen Pandemie kann die Digitalisierung und Verschlankung von Prozessen im Gesundheitswesen dabei unterstützen, wertvolle Ressourcen zu sparen und Überlastungen abzufedern beispielsweise durch den Einsatz von telemedizinischen Anwendungen wie digitale Sprechstunden und intelligente Operations-Planung. Dadurch ergeben sich völlig neue Potenziale aber auch Herausforderungen für den Einsatz von Künstlicher Intelligenz. Für das vorliegende Whitepaper haben wir einen konkreten Anwendungsfall, die Notfallversorgung im Krankenhaus, im Detail untersucht und beschrieben.
-
PublicationUsing Probabilistic Soft Logic to Improve Information Extraction in the Legal Domain( 2020)Extracting information from court process documents to populate a knowledge base produces data valuable to legal faculties, publishers and law firms. A challenge lies in the fact that the relevant information is interdependent and structured by numerous semantic constraints of the legal domain. Ignoring these dependencies leads to inferior solutions. Hence, the objective of this paper is to demonstrate how the extraction pipeline can be improved by the use of probabilistic soft logic rules that reflect both legal and linguistic knowledge. We propose a probabilistic rule model for the overall extraction pipeline, which enables to both map dependencies between local extraction models and to integrate additional domain knowledge in the form of logical constraints. We evaluate the performance of the model on a German court sentences corpus.
-
PublicationImproving Word Embeddings Using Kernel PCA( 2019)Word-based embedding approaches such as Word2Vec capture the meaning of words and relations between them, particularly well when trained with large text collections; however, they fail to do so with small datasets. Extensions such as fastText reduce the amount of data needed slightly, however, the joint task of learning meaningful morphology, syntactic and semantic representations still requires a lot of data. In this paper, we introduce a new approach to warm-start embedding models with morphological information, in order to reduce training time and enhance their performance. We use word embeddings generated using both word2vec and fastText models and enrich them with morphological information of words, derived from kernel principal component analysis (KPCA) of word similarity matrices. This can be seen as explicitly feeding the network morphological similarities and letting it learn semantic and syntactic similarities. Evaluating our models on word similarity and analogy tasks in English and German, we find that they not only achieve higher accuracies than the original skip-gram and fastText models but also require significantly less training data and time. Another benefit of our approach is that it is capable of generating a high-quality representation of infrequent words as, for example, found in very recent news articles with rapidly changing vocabularies. Lastly, we evaluate the different models on a downstream sentence classification task in which a CNN model is initialized with our embeddings and find promising results.
-
PublicationNoise Reduction in Distant Supervision for Relation Extraction Using Probabilistic Soft Logic( 2019)The performance of modern relation extraction systems is to a great degree dependent on the size and quality of the underlying training corpus and in particular on the labels. Since generating these labels by human annotators is expensive, Distant Supervision has been proposed to automatically align entities in a knowledge base with a text corpus to generate annotations. However, this approach suffers from introducing noise, which negatively affects the performance of relation extraction systems. To tackle this problem, we propose a probabilistic graphical model which simultaneously incorporates different sources of knowledge such as domain experts knowledge about the context and linguistic knowledge about the sentence structure in a principled way. The model is defined using the declarati ve language provided by Probabilistic Soft Logic. Experimental results show that the proposed approach, compared to the original distantly supervised set, not only improves the quality of such generated training data sets, but also the performance of the final relation extraction model. The performance of modern relation extraction systems is to a great degree dependent on the size and quality of the underlying training corpus and in particular on the labels. Since generating these labels by human annotators is expensive, Distant Supervision has been proposed to automatically align entities in a knowledge base with a text corpus to generate annotations. However, this approach suffers from introducing noise, which negatively affects the performance of relation extraction systems. To tackle this problem, we propose a probabilistic graphical model which simultaneously incorporates different sources of knowledge such as domain experts knowledge about the context and linguistic knowledge about the sentence structure in a principled way. The model is defined using the declarati ve language provided by Probabilistic Soft Logic. Experimental results show that the proposed approach, compared to the original distantly supervised set, not only improves the quality of such generated training data sets, but also the performance of the final relation extraction model.
- 1 (current)
- 2
- 3
- 4
- 5