Maleshkova, Maria

Type 1 diabetes (T1D) management can be significantly enhanced through the use of predictive machine learning (ML) algorithms, which can mitigate the risk of adverse events like hypoglycemia. Hypoglycemia, characterized by blood glucose levels below 70 mg/dL, is a life-threatening condition typically caused by excessive insulin administration, missed meals, or physical activity. Its asymptomatic nature impedes timely intervention, making ML models crucial for early detection. This study integrates short- (up to 2h) and long-term (up to 24h) prediction horizons (PHs) within a single classification model to enhance decision support. The predicted times are 5-15 min, 15-30 min, 30 min-1h, 1-2h, 2-4h, 4-8h, 8-12h, and 12-24h before hypoglycemia. In addition, a simplified model classifying up to 4h before hypoglycemia is compared. We trained ResNet and LSTM models on glucose levels, insulin doses, and acceleration data. The results demonstrate the superiority of the LSTM models when classifying nine classes. In particular, subject-specific models yielded better performance but achieved high recall only for classes 0, 1, and 2 with 98%, 72%, and 50%, respectively. A population-based six-class model improved the results with at least 60% of events detected. In contrast, longer PHs remain challenging with the current approach and may be considered with different models.

Specific phobias, such as spider phobia, are a widespread condition, that can negatively impact the quality of life of affected people. Phobias are typically characterised by avoidance behaviour, which can be measured using a behavioural avoidance test (BAT). While behavioural avoidance tests are traditionally performed using a real stimulus (e.g., a spider), virtual reality has also gained popularity in the field of psychology. This offers a more accessible and affordable alternative to traditional diagnostic methods. Our work focuses on a comparison of BATs in vivo and in virtuo, and considering two different approach modalities, by analysing the physiological responses of participants. This study aims to investigate whether these responses during in virtuo BATs are comparable to those observed during in vivo BATs, and whether the modality influences the outcome. In this work, we present our study involving 25 participants and an initial look at the data collected.

Psychological trials, such as behavioural avoidance tests (BAT), are a fundamental part of the phobia therapeutic process. In order to link physiological reactions with specific points in time during psychological trials, it is necessary to integrate observation data with data collected automatically by sensors, such as wearable devices. To this end, this paper introduces FRAME - a framework for combining real-world events occurring during psychological trials with physiological data collected by wearables. FRAME consists of three parts, an Observation App, a data integration module and a Virtual Reality (VR) App. The Observation App captures events and their exact time of occurrence. The integration module links the observations with the respective physiological data, allowing an in-depth analysis of physiological reactions. The VR App provides a virtual scenario based on the BAT in vivo, thus enabling a BAT in virtuo. The practical applicability of FRAME is tested within a study comparing behavioural avoidance tests in vivo and in virtuo, assessing 25 patients with arachnophobia wearing an Empatica E4.

Patients with type 1 diabetes (T1D) have a higher risk of experiencing hypoglycemia, which is a severe condition of decreased blood glucose levels under 70 mg/dL and can result in coma, or death in the worst case. Prediction algorithms could improve diabetes care by enabling preventive actions, but research is limited by available multivariate datasets. Thus, this work investigates the feasibility of transfer learning between two different datasets of people with T1D and type 2 pre-diabetes using a 1 Dimensional Convolutional Neural Network (1DCNN) model. Moreover, different thresholds for defining hypoglycemia are compared for the pre-diabetes group. The results show that transfer learning could be feasible if the model is trained on T1D with a threshold of 70 mg/dL, while a threshold of 80–85 mg/dL achieved the best training performance for the pre-diabetes data.

The field of virtual reality (VR) is no longer limited to the entertainment sector, and the use also increased in the sector of healthcare and therapy. This paper attempts to compare the effectiveness of a Behavioral Avoidance Test for spider phobia in virtual reality to in vivo. To do this, two different BAT approaches are designed for both in vivo and in virtuo to be compared by a group of phobic participants. In addition, physiological data of the participants were recorded to investigate this comparison on a physiological level.

This paper addresses the task of conversational question answering (ConvQA) over knowledge graphs (KGs). The majority of existing ConvQA methods rely on full supervision signals with a strict assumption of the availability of gold logical forms of queries to extract answers from the KG. However, creating such a gold logical form is not viable for each potential question in a real-world scenario. Hence, in the case of missing gold logical forms, the existing information retrieval-based approaches use weak supervision via heuristics or reinforcement learning, formulating ConvQA as a KG path ranking problem. Despite missing gold logical forms, an abundance of conversational contexts, such as entire dialog history with fluent responses and domain information, can be incorporated to effectively reach the correct KG path. This work proposes a contrastive representation learning-based approach to rank KG paths effectively. Our approach solves two key challenges. Firstly, it allows weak supervision-based learning that omits the necessity of gold annotations. Second, it incorporates the conversational context (entire dialog history and domain information) to jointly learn its homogeneous representation with KG paths to improve contrastive representations for effective path ranking. We evaluate our approach on standard datasets for ConvQA, on which it significantly outperforms existing baselines on all domains and overall. Specifically, in some cases, the Mean Reciprocal Rank (MRR) and Hit@5 ranking metrics improve by absolute 10 and 18 points, respectively, compared to the state-of-the-art performance. © 2022 ACM.

In recent years, there have been significant developments in Question Answering over Knowledge Graphs (KGQA). Despite all the notable advancements, current KGQA systems only focus on answer generation techniques and not on answer verbalization. However, in real-world scenarios (e.g., voice assistants such as Alexa, Siri, etc.), users prefer verbalized answers instead of a generated response. This paper addresses the task of answer verbalization for (complex) question answering over knowledge graphs. In this context, we propose a multi-task-based answer verbalization framework: VOGUE (Verbalization thrOuGh mUlti-task lEarning). The VOGUE framework attempts to generate a verbalized answer using a hybrid approach through a multi-task learning paradigm. Our framework can generate results based on using questions and queries as inputs concurrently. VOGUE comprises four modules that are trained simultaneously through multi-task learning. We evaluate our framework on existing datasets for answer verbalization, and it outperforms all current baselines on both BLEU and METEOR scores. © 2021, Springer Nature Switzerland AG.

Medical knowledge graphs (KGs) constructed from Electronic Medical Records (EMR) contain abundant information about patients and medical entities. The utilization of KG embedding models on these data has proven to be efficient for different medical tasks. However, existing models do not properly incorporate patient demographics and most of them ignore the probabilistic features of the medical KG. In this paper, we propose DARLING (Demographic Aware pRobabiListic medIcal kNowledge embeddinG), a demographic-aware medical KG embedding framework that explicitly incorporates demographics in the medical entities space by associating patient demographics with a corresponding hyperplane. Our framework leverages the probabilistic features within the medical entities for learning their representations through demographic guidance. We evaluate DARLING through link prediction for treatments and medicines, on a medical KG constructed from EMR data, and illustrate its superior performance compared to existing KG embedding models.

This paper addresses the task of (complex) conversational question answering over a knowledge graph. For this task, we propose LASAGNE (muLti-task semAntic parSing with trAnsformer and Graph atteNtion nEtworks). It is the first approach, which employs a transformer architecture extended with Graph Attention Networks for multi-task neural semantic parsing. LASAGNE uses a transformer model for generating the base logical forms, while the Graph Attention model is used to exploit correlations between (entity) types and predicates to produce node representations. LASAGNE also includes a novel entity recognition module which detects, links, and ranks all relevant entities in the question context. We evaluate LASAGNE on a standard dataset for complex sequential question answering, on which it outperforms existing baseline averages on all question types. Specifically, we show that LASAGNE improves the F1-score on eight out of ten question types; in some cases, the increase in F1-score is more than 20% compared to the state of the art.

In this paper, we introduce the MLM (Multiple Languages and Modalities) dataset - a new resource to train and evaluate multitask systems on samples in multiple modalities and three languages. The generation process and inclusion of semantic data provide a resource that further tests the ability for multitask systems to learn relationships between entities. The dataset is designed for researchers and developers who build applications that perform multiple tasks on data encountered on the web and in digital archives. A second version of MLM provides a geo-representative subset of the data with weighted samples for countries of the European Union. We demonstrate the value of the resource in developing novel applications in the digital humanities with a motivating use case and specify a benchmark set of tasks to retrieve modalities and locate entities in the dataset. Evaluation of baseline multitask and single task systems on the full and geo-representative versions of MLM demonstrate the challenges of generalising on diverse data. In addition to the digital humanities, we expect the resource to contribute to research in multimodal representation learning, location estimation, and scene understanding.