004 Datenverarbeitung; Informatik
Refine
Has Fulltext
- yes (78)
Is part of the Bibliography
- yes (78)
Year of publication
Document Type
- Journal article (78) (remove)
Keywords
- virtual reality (11)
- machine learning (5)
- augmented reality (3)
- immersion (3)
- Deep learning (2)
- Quadrocopter (2)
- Quadrotor (2)
- XR (2)
- artificial intelligence (2)
- automation (2)
Institute
- Institut für Informatik (78) (remove)
Neural networks have to capture mathematical relationships in order to learn various tasks. They approximate these relations implicitly and therefore often do not generalize well. The recently proposed Neural Arithmetic Logic Unit (NALU) is a novel neural architecture which is able to explicitly represent the mathematical relationships by the units of the network to learn operations such as summation, subtraction or multiplication. Although NALUs have been shown to perform well on various downstream tasks, an in-depth analysis reveals practical shortcomings by design, such as the inability to multiply or divide negative input values or training stability issues for deeper networks. We address these issues and propose an improved model architecture. We evaluate our model empirically in various settings from learning basic arithmetic operations to more complex functions. Our experiments indicate that our model solves stability issues and outperforms the original NALU model in means of arithmetic precision and convergence.
The strict restrictions introduced by the COVID-19 lockdowns, which started from March 2020, changed people’s daily lives and habits on many different levels. In this work, we investigate the impact of the lockdown on the communication behavior in the mobile instant messaging application WhatsApp. Our evaluations are based on a large dataset of 2577 private chat histories with 25,378,093 messages from 51,973 users. The analysis of the one-to-one and group conversations confirms that the lockdown severely altered the communication in WhatsApp chats compared to pre-pandemic time ranges. In particular, we observe short-term effects, which caused an increased message frequency in the first lockdown months and a shifted communication activity during the day in March and April 2020. Moreover, we also see long-term effects of the ongoing pandemic situation until February 2021, which indicate a change of communication behavior towards more regular messaging, as well as a persisting change in activity during the day. The results of our work show that even anonymized chat histories can tell us a lot about people’s behavior and especially behavioral changes during the COVID-19 pandemic and thus are of great relevance for behavioral researchers. Furthermore, looking at the pandemic from an Internet provider perspective, these insights can be used during the next pandemic, or if the current COVID-19 situation worsens, to adapt communication networks to the changed usage behavior early on and thus avoid network congestion.
Group-based communication is a highly popular communication paradigm, which is especially prominent in mobile instant messaging (MIM) applications, such as WhatsApp. Chat groups in MIM applications facilitate the sharing of various types of messages (e.g., text, voice, image, video) among a large number of participants. As each message has to be transmitted to every other member of the group, which multiplies the traffic, this has a massive impact on the underlying communication networks. However, most chat groups are private and network operators cannot obtain deep insights into MIM communication via network measurements due to end-to-end encryption. Thus, the generation of traffic is not well understood, given that it depends on sizes of communication groups, speed of communication, and exchanged message types. In this work, we provide a huge data set of 5,956 private WhatsApp chat histories, which contains over 76 million messages from more than 117,000 users. We describe and model the properties of chat groups and users, and the communication within these chat groups, which gives unprecedented insights into private MIM communication. In addition, we conduct exemplary measurements for the most popular message types, which empower the provided models to estimate the traffic over time in a chat group.
To deliver the best user experience (UX), the human-centered design cycle (HCDC) serves as a well-established guideline to application developers. However, it does not yet cover network-specific requirements, which become increasingly crucial, as most applications deliver experience over the Internet. The missing network-centric view is provided by Quality of Experience (QoE), which could team up with UX towards an improved overall experience. By considering QoE aspects during the development process, it can be achieved that applications become network-aware by design. In this paper, the Quality of Experience Centered Design Cycle (QoE-CDC) is proposed, which provides guidelines on how to design applications with respect to network-specific requirements and QoE. Its practical value is showcased for popular application types and validated by outlining the design of a new smartphone application. We show that combining HCDC and QoE-CDC will result in an application design, which reaches a high UX and avoids QoE degradation.
Virtual environments (VEs) can evoke and support emotions, as experienced when playing emotionally arousing games. We theoretically approach the design of fear and joy evoking VEs based on a literature review of empirical studies on virtual and real environments as well as video games’ reviews and content analyses. We define the design space and identify central design elements that evoke specific positive and negative emotions. Based on that, we derive and present guidelines for emotion-inducing VE design with respect to design themes, colors and textures, and lighting configurations. To validate our guidelines in two user studies, we 1) expose participants to 360° videos of VEs designed following the individual guidelines and 2) immerse them in a neutral, positive and negative emotion-inducing VEs combining all respective guidelines in Virtual Reality. The results support our theoretically derived guidelines by revealing significant differences in terms of fear and joy induction.
Climate models are the tool of choice for scientists researching climate change. Like all models they suffer from errors, particularly systematic and location-specific representation errors. One way to reduce these errors is model output statistics (MOS) where the model output is fitted to observational data with machine learning. In this work, we assess the use of convolutional Deep Learning climate MOS approaches and present the ConvMOS architecture which is specifically designed based on the observation that there are systematic and location-specific errors in the precipitation estimates of climate models. We apply ConvMOS models to the simulated precipitation of the regional climate model REMO, showing that a combination of per-location model parameters for reducing location-specific errors and global model parameters for reducing systematic errors is indeed beneficial for MOS performance. We find that ConvMOS models can reduce errors considerably and perform significantly better than three commonly used MOS approaches and plain ResNet and U-Net models in most cases. Our results show that non-linear MOS models underestimate the number of extreme precipitation events, which we alleviate by training models specialized towards extreme precipitation events with the imbalanced regression method DenseLoss. While we consider climate MOS, we argue that aspects of ConvMOS may also be beneficial in other domains with geospatial data, such as air pollution modeling or weather forecasts.
In many real world settings, imbalanced data impedes model performance of learning algorithms, like neural networks, mostly for rare cases. This is especially problematic for tasks focusing on these rare occurrences. For example, when estimating precipitation, extreme rainfall events are scarce but important considering their potential consequences. While there are numerous well studied solutions for classification settings, most of them cannot be applied to regression easily. Of the few solutions for regression tasks, barely any have explored cost-sensitive learning which is known to have advantages compared to sampling-based methods in classification tasks. In this work, we propose a sample weighting approach for imbalanced regression datasets called DenseWeight and a cost-sensitive learning approach for neural network regression with imbalanced data called DenseLoss based on our weighting scheme. DenseWeight weights data points according to their target value rarities through kernel density estimation (KDE). DenseLoss adjusts each data point’s influence on the loss according to DenseWeight, giving rare data points more influence on model training compared to common data points. We show on multiple differently distributed datasets that DenseLoss significantly improves model performance for rare data points through its density-based weighting scheme. Additionally, we compare DenseLoss to the state-of-the-art method SMOGN, finding that our method mostly yields better performance. Our approach provides more control over model training as it enables us to actively decide on the trade-off between focusing on common or rare cases through a single hyperparameter, allowing the training of better models for rare data points.
The issue of sustainability is at the top of the political and societal agenda, being considered of extreme importance and urgency. Human individual action impacts the environment both locally (e.g., local air/water quality, noise disturbance) and globally (e.g., climate change, resource use). Urban environments represent a crucial example, with an increasing realization that the most effective way of producing a change is involving the citizens themselves in monitoring campaigns (a citizen science bottom-up approach). This is possible by developing novel technologies and IT infrastructures enabling large citizen participation. Here, in the wider framework of one of the first such projects, we show results from an international competition where citizens were involved in mobile air pollution monitoring using low cost sensing devices, combined with a web-based game to monitor perceived levels of pollution. Measures of shift in perceptions over the course of the campaign are provided, together with insights into participatory patterns emerging from this study. Interesting effects related to inertia and to direct involvement in measurement activities rather than indirect information exposure are also highlighted, indicating that direct involvement can enhance learning and environmental awareness. In the future, this could result in better adoption of policies towards decreasing pollution.
Background
Information extraction techniques that get structured representations out of unstructured data make a large amount of clinically relevant information about patients accessible for semantic applications. These methods typically rely on standardized terminologies that guide this process. Many languages and clinical domains, however, lack appropriate resources and tools, as well as evaluations of their applications, especially if detailed conceptualizations of the domain are required. For instance, German transthoracic echocardiography reports have not been targeted sufficiently before, despite of their importance for clinical trials. This work therefore aimed at development and evaluation of an information extraction component with a fine-grained terminology that enables to recognize almost all relevant information stated in German transthoracic echocardiography reports at the University Hospital of Würzburg.
Methods
A domain expert validated and iteratively refined an automatically inferred base terminology. The terminology was used by an ontology-driven information extraction system that outputs attribute value pairs. The final component has been mapped to the central elements of a standardized terminology, and it has been evaluated according to documents with different layouts.
Results
The final system achieved state-of-the-art precision (micro average.996) and recall (micro average.961) on 100 test documents that represent more than 90 % of all reports. In particular, principal aspects as defined in a standardized external terminology were recognized with f 1=.989 (micro average) and f 1=.963 (macro average). As a result of keyword matching and restraint concept extraction, the system obtained high precision also on unstructured or exceptionally short documents, and documents with uncommon layout.
Conclusions
The developed terminology and the proposed information extraction system allow to extract fine-grained information from German semi-structured transthoracic echocardiography reports with very high precision and high recall on the majority of documents at the University Hospital of Würzburg. Extracted results populate a clinical data warehouse which supports clinical research.
Heat and excessive solar radiation can produce abiotic stresses during apple maturation, resulting fruit quality. Therefore, the monitoring of temperature on fruit surface (FST) over the growing period can allow to identify thresholds, above of which several physiological disorders such as sunburn may occur in apple.
The current approaches neglect spatial variation of FST and have reduced repeatability, resulting in unreliable predictions. In this study, LiDAR laser scanning and thermal imaging were employed to detect the temperature on fruit surface by means of 3D point cloud. A process for calibrating the two sensors based on an active board target and producing a 3D thermal point cloud was suggested. After calibration, the sensor system was utilised to scan the fruit trees, while temperature values assigned in the corresponding 3D point cloud were based on the extrinsic calibration. Whereas a fruit detection algorithm was performed to segment the FST from each apple.
• The approach allows the calibration of LiDAR laser scanner with thermal camera in order to produce a 3D thermal point cloud.
• The method can be applied in apple trees for segmenting FST in 3D. Whereas the approach can be utilised to predict several physiological disorders including sunburn on fruit surface.
Psycho-pathological conditions, such as depression or schizophrenia, are often accompanied by a distorted perception of time. People suffering from this conditions often report that the passage of time slows down considerably and that they are “stuck in time.” Virtual Reality (VR) could potentially help to diagnose and maybe treat such mental conditions. However, the conditions in which a VR simulation could correctly diagnose a time perception deviation are still unknown. In this paper, we present an experiment investigating the difference in time experience with and without a virtual body in VR, also known as avatar. The process of substituting a person’s body with a virtual body is called avatar embodiment. Numerous studies demonstrated interesting perceptual, emotional, behavioral, and psychological effects caused by avatar embodiment. However, the relations between time perception and avatar embodiment are still unclear. Whether or not the presence or absence of an avatar is already influencing time perception is still open to question. Therefore, we conducted a between-subjects design with and without avatar embodiment as well as a real condition (avatar vs. no-avatar vs. real). A group of 105 healthy subjects had to wait for seven and a half minutes in a room without any distractors (e.g., no window, magazine, people, decoration) or time indicators (e.g., clocks, sunlight). The virtual environment replicates the real physical environment. Participants were unaware that they will be asked to estimate their waiting time duration as well as describing their experience of the passage of time at a later stage. Our main finding shows that the presence of an avatar is leading to a significantly faster perceived passage of time. It seems to be promising to integrate avatar embodiment in future VR time-based therapy applications as they potentially could modulate a user’s perception of the passage of time. We also found no significant difference in time perception between the real and the VR conditions (avatar, no-avatar), but further research is needed to better understand this outcome.
Crowdsourced network measurements (CNMs) are becoming increasingly popular as they assess the performance of a mobile network from the end user's perspective on a large scale. Here, network measurements are performed directly on the end-users' devices, thus taking advantage of the real-world conditions end-users encounter. However, this type of uncontrolled measurement raises questions about its validity and reliability. The problem lies in the nature of this type of data collection. In CNMs, mobile network subscribers are involved to a large extent in the measurement process, and collect data themselves for the operator. The collection of data on user devices in arbitrary locations and at uncontrolled times requires means to ensure validity and reliability. To address this issue, our paper defines concepts and guidelines for analyzing the precision of CNMs; specifically, the number of measurements required to make valid statements. In addition to the formal definition of the aspect, we illustrate the problem and use an extensive sample data set to show possible assessment approaches. This data set consists of more than 20.4 million crowdsourced mobile measurements from across France, measured by a commercial data provider.
Even today, the automatic digitisation of scanned documents in general, but especially the automatic optical music recognition (OMR) of historical manuscripts, still remains an enormous challenge, since both handwritten musical symbols and text have to be identified. This paper focuses on the Medieval so-called square notation developed in the 11th–12th century, which is already composed of staff lines, staves, clefs, accidentals, and neumes that are roughly spoken connected single notes. The aim is to develop an algorithm that captures both the neumes, and in particular its melody, which can be used to reconstruct the original writing. Our pipeline is similar to the standard OMR approach and comprises a novel staff line and symbol detection algorithm based on deep Fully Convolutional Networks (FCN), which perform pixel-based predictions for either staff lines or symbols and their respective types. Then, the staff line detection combines the extracted lines to staves and yields an F\(_1\) -score of over 99% for both detecting lines and complete staves. For the music symbol detection, we choose a novel approach that skips the step to identify neumes and instead directly predicts note components (NCs) and their respective affiliation to a neume. Furthermore, the algorithm detects clefs and accidentals. Our algorithm predicts the symbol sequence of a staff with a diplomatic symbol accuracy rate (dSAR) of about 87%, which includes symbol type and location. If only the NCs without their respective connection to a neume, all clefs and accidentals are of interest, the algorithm reaches an harmonic symbol accuracy rate (hSAR) of approximately 90%. In general, the algorithm recognises a symbol in the manuscript with an F\(_1\) -score of over 96%.
Social patterns and roles can develop when users talk to intelligent voice assistants (IVAs) daily. The current study investigates whether users assign different roles to devices and how this affects their usage behavior, user experience, and social perceptions. Since social roles take time to establish, we equipped 106 participants with Alexa or Google assistants and some smart home devices and observed their interactions for nine months. We analyzed diverse subjective (questionnaire) and objective data (interaction data). By combining social science and data science analyses, we identified two distinct clusters—users who assigned a friendship role to IVAs over time and users who did not. Interestingly, these clusters exhibited significant differences in their usage behavior, user experience, and social perceptions of the devices. For example, participants who assigned a role to IVAs attributed more friendship to them used them more frequently, reported more enjoyment during interactions, and perceived more empathy for IVAs. In addition, these users had distinct personal requirements, for example, they reported more loneliness. This study provides valuable insights into the role-specific effects and consequences of voice assistants. Recent developments in conversational language models such as ChatGPT suggest that the findings of this study could make an important contribution to the design of dialogic human–AI interactions.
Plenty of theories, models, measures, and investigations target the understanding of virtual presence, i.e., the sense of presence in immersive Virtual Reality (VR). Other varieties of the so-called eXtended Realities (XR), e.g., Augmented and Mixed Reality (AR and MR) incorporate immersive features to a lesser degree and continuously combine spatial cues from the real physical space and the simulated virtual space. This blurred separation questions the applicability of the accumulated knowledge about the similarities of virtual presence and presence occurring in other varieties of XR, and corresponding outcomes. The present work bridges this gap by analyzing the construct of presence in mixed realities (MR). To achieve this, the following presents (1) a short review of definitions, dimensions, and measurements of presence in VR, and (2) the state of the art views on MR. Additionally, we (3) derived a working definition of MR, extending the Milgram continuum. This definition is based on entities reaching from real to virtual manifestations at one time point. Entities possess different degrees of referential power, determining the selection of the frame of reference. Furthermore, we (4) identified three research desiderata, including research questions about the frame of reference, the corresponding dimension of transportation, and the dimension of realism in MR. Mainly the relationship between the main aspects of virtual presence of immersive VR, i.e., the place-illusion, and the plausibility-illusion, and of the referential power of MR entities are discussed regarding the concept, measures, and design of presence in MR. Finally, (5) we suggested an experimental setup to reveal the research heuristic behind experiments investigating presence in MR. The present work contributes to the theories and the meaning of and approaches to simulate and measure presence in MR. We hypothesize that research about essential underlying factors determining user experience (UX) in MR simulations and experiences is still in its infancy and hopes this article provides an encouraging starting point to tackle related questions.
Artificial Intelligence (AI) covers a broad spectrum of computational problems and use cases. Many of those implicate profound and sometimes intricate questions of how humans interact or should interact with AIs. Moreover, many users or future users do have abstract ideas of what AI is, significantly depending on the specific embodiment of AI applications. Human-centered-design approaches would suggest evaluating the impact of different embodiments on human perception of and interaction with AI. An approach that is difficult to realize due to the sheer complexity of application fields and embodiments in reality. However, here XR opens new possibilities to research human-AI interactions. The article’s contribution is twofold: First, it provides a theoretical treatment and model of human-AI interaction based on an XR-AI continuum as a framework for and a perspective of different approaches of XR-AI combinations. It motivates XR-AI combinations as a method to learn about the effects of prospective human-AI interfaces and shows why the combination of XR and AI fruitfully contributes to a valid and systematic investigation of human-AI interactions and interfaces. Second, the article provides two exemplary experiments investigating the aforementioned approach for two distinct AI-systems. The first experiment reveals an interesting gender effect in human-robot interaction, while the second experiment reveals an Eliza effect of a recommender system. Here the article introduces two paradigmatic implementations of the proposed XR testbed for human-AI interactions and interfaces and shows how a valid and systematic investigation can be conducted. In sum, the article opens new perspectives on how XR benefits human-centered AI design and development.
In this paper we study connectivity augmentation problems. Given a connected graph G with some desirable property, we want to make G 2-vertex connected (or 2-edge connected) by adding edges such that the resulting graph keeps the property. The aim is to add as few edges as possible. The property that we consider is planarity, both in an abstract graph-theoretic and in a geometric setting, where vertices correspond to points in the plane and edges to straight-line segments.
We show that it is NP-hard to nd a minimum-cardinality augmentation that makes a planar graph 2-edge connected. For making a planar graph 2-vertex connected this was known. We further show that both problems are hard in the geometric setting, even when restricted to trees. The problems remain hard for higher degrees of connectivity. On the other hand we give polynomial-time algorithms for the special case of convex geometric graphs.
We also study the following related problem. Given a planar (plane geometric) graph G, two vertices s and t of G, and an integer c, how many edges have to be added to G such that G is still planar (plane geometric) and contains c edge- (or vertex-) disjoint s{t paths? For the planar case we give a linear-time algorithm for c = 2. For the plane geometric case we give optimal worst-case bounds for c = 2; for c = 3 we characterize the cases that have a solution.
Semantic Fusion for Natural Multimodal Interfaces using Concurrent Augmented Transition Networks
(2018)
Semantic fusion is a central requirement of many multimodal interfaces. Procedural methods like finite-state transducers and augmented transition networks have proven to be beneficial to implement semantic fusion. They are compliant with rapid development cycles that are common for the development of user interfaces, in contrast to machine-learning approaches that require time-costly training and optimization. We identify seven fundamental requirements for the implementation of semantic fusion: Action derivation, continuous feedback, context-sensitivity, temporal relation support, access to the interaction context, as well as the support of chronologically unsorted and probabilistic input. A subsequent analysis reveals, however, that there is currently no solution for fulfilling the latter two requirements. As the main contribution of this article, we thus present the Concurrent Cursor concept to compensate these shortcomings. In addition, we showcase a reference implementation, the Concurrent Augmented Transition Network (cATN), that validates the concept’s feasibility in a series of proof of concept demonstrations as well as through a comparative benchmark. The cATN fulfills all identified requirements and fills the lack amongst previous solutions. It supports the rapid prototyping of multimodal interfaces by means of five concrete traits: Its declarative nature, the recursiveness of the underlying transition network, the network abstraction constructs of its description language, the utilized semantic queries, and an abstraction layer for lexical information. Our reference implementation was and is used in various student projects, theses, as well as master-level courses. It is openly available and showcases that non-experts can effectively implement multimodal interfaces, even for non-trivial applications in mixed and virtual reality.