Refine
Document Type
- Article (peer reviewed) (10)
- Conference Proceeding (10)
- Part of a Book (1)
Has Fulltext
- no (21)
Is part of the Bibliography
- no (21)
Keywords
In modern medicine, Clinical Practice Guidelines (CPGs) are well-established resources for the appropriate treatment of diseases. Evidence-based CPGs contain recommendations which are based on the state of the art and which have been achieved by consensus of several experts. Nevertheless, there is a potential for problems in translating guideline documents into specific actions for physicians. Therefore we propose to formalize the treatment process in an understandable representation as UML activities together with a domain expert. This formalization serves as a basis for the transfer of knowledge into a model, in this case PROforma, which directly allows execution in an interactive assistance software. The results of this work are part of an ongoing research project on the treatment of colon cancer based on the corresponding evidence-based CPG.
Probabilistic Estimation of Human Interaction Needs in Context of a Robotic Assistance in Geriatrics
(2019)
The key purpose of assistance robots is to help people coping with work-related or everyday tasks. To ensure an intuitive and effective support by an assistance robot, its expectation conform behavior is essential. In particular, when using assistance robots in geriatrics to assist elderly patients, special attention to the human-robot interaction should be paid. In order to help elderly patients maintain their independence and abilities as much as possible, the robot should only intervene when its support is needed. Therefore, the continuous estimation of the patient’s need for interaction is of particular importance. For enabling suitable models to estimate this need, we elaborate the use of Bayesian Networks. The analysis of our results seems promising, yielding a robust and practical approach.
Der Einsatz künstlicher Intelligenz (KI) wird unsere Lebenswirklichkeit nachhaltig beeinflussen. Mittels technischer Verfahren maschinellen Lernens oder tiefer neuronaler Netze können Modelle aus existierenden Datenbeständen gelernt und Vorhersagen abgeleitet werden. Wir stellen skizzenartig zwei aktuelle Projekte vor, die sich u.a. mit den Potentialen und der menschzentrierten Gestaltung von Systemen künstlicher Intelligenz beschäftigen. Wir betten dies schlaglichtartig in den Kontext der aktuellen Debatte über die Entwicklung und Gestaltung von KI Systemen ein. Im Zuge dessen umreißen wir Herausforderungen und Chancen für die partizipative und sozialverantwortliche Technologieentwicklung von komplexen und auf Verfahren der künstlichen Intelligenz basierenden Systemen. Schließlich fokussieren wir dabei auf die Arbeitsebene und beschreiben kurz, welche konkreten Aktivitäten wir planen und stellen diese zur Diskussion.
With the digital transformation of companies, ever larger amounts of data are generated and available for analysis. Process mining techniques can be used to extract and analyze process models from these data. Related techniques have quickly developed into an important field with constantly increasing investments in recent years. Thus, the automated analysis of processes has gained an important role in many companies. In this context, graphs have been shown to be an intuitive representation of how the gathered processes are carried out using the aforementioned techniques. For the analysis of these so-called control flow graphs, we investigate the use of convolution neural networks, which are specially designed for graphs: graph convolution networks (GCNs). In our contribution, GCNs are used to perform a regression task based on individual control flows of a process in which farmers apply for specific governmental payments. The approach achieved promising results on this publicly available data set.
Artificial Intelligence (AI) is continuously moving into our surroundings. In its various forms, it has the potential to disrupt most aspects of human life. Yet, the discourse around AI has long been by experts and for experts. In this paper, we argue for a participatory approach towards designing human-AI interactions. We outline how we used design methodology to organise an interdisciplinary workshop with a diverse group of students – a workbook sprint with 45 participants from four different programs and 13 countries – to develop speculative design futures in five focus areas. We then provide insights into our findings and share our lessons learned regarding our workshop topic – AI and Space – our process, and our research. We learned that involving non-experts in complex technical discourses – such as AI – through the structural rigour of design methodology is a viable approach. We then conclude by laying out how others might use our findings and initiate their own workbook sprint to explore complex technologies in a human-centred way.
In this position paper, we lay out an approach to use participatory and co-design methodology to explore how users perceive and interact with explanations of artificially intelligent decision support systems. We describe how we intend to construct bottom-up participatory design spaces to systematically inform the design of interactive explanations in Human-AI interaction.
Oncologic patients are regarded as the population most at risk of developing a severe course of COVID-19 due to the fact that malignant diseases and chemotherapy often weaken the immune system. In the face of the ongoing SARS-CoV-2 pandemic, how particular patients deal with this infection remains an important question. In the period between the 15 and 26 April 2020, a total of 1227 patients were tested in one of seven oncologic outpatient clinics for SARS-CoV-2, regardless of symptoms, employing RT-qPCR. Of 1227 patients, 78 (6.4%) were tested positive of SARS-CoV-2. Only one of the patients who tested positive developed a severe form of COVID-19 with pneumonia (CURB-65 score of 2), and two patients showed mild symptoms. Fourteen of 75 asymptomatic but positively tested patients received chemotherapy or chemo-immunotherapy according to their regular therapy algorithm (±4 weeks of SARS-CoV-2 test), and 48 of 78 (61.5%) positive-tested patients received glucocorticoids as co-medication. None of the asymptomatic infected patients showed unexpected complications due to the SARS-CoV-2 infection during the cancer treatment. These data clearly contrast the view that patients with an oncologic disease are particularly vulnerable to SARS-CoV-2 and suggest that compromising therapies could be continued or started despite the ongoing pandemic. Moreover the relatively low appearance of symptoms due to COVID-19 among patients on chemotherapy and other immunosuppressive co-medication like glucocorticoids indicate that suppressing the response capacity of the immune system reduces disease severity.
Artificial Intelligence (AI) and Machine Learning (ML) promise to significantly enhance the capabilities of decision support systems in medicine. Yet, if these systems fail to providean understandable rationale of the decision making process the adoption of this powerful technology will be difficult. Hence, there is growing interest in Explainable Artificial Intelligence (XAI). Explanations that speak the language of the user are a cornerstone for Fair AI. In this position paper, we reflect on research about XAI and designing decision support in medicine. From there, we lay out an approach to use co-design methodology to explore how users perceive and interact with explanations of decision support systems.
German cancer registries offer a systematic approach for the collection, storage, and management of data on patients with cancer and related diseases. Much hope in research and healthcare in general is depending on such register-based analyses in order to comprehensively consider the features of a highly diverse population. Next to the data collection the cancer registries are responsible for data protection. To fulfill legal regulations, access to data has to be controlled in a strict way leading to sometimes bureaucratic and slow processes. The situation is especially complicated in Germany, since cancer data is distributed over numerous federal cancer registries. If a nationwide data evaluation is conducted a research team has to negotiate a separate contract with each cancer registry.In a joint work in progress effort of cancer registries, technical, medical, and economical experts we propose a different solution for cooperative data processing. Our approach aims for combining data in a virtual pool based on the selection criteria of individual requests from researchers. To achieve our goal, we adapt the Fraunhofer Medical Data Space as enabling technology. The architecture we propose will allow us to pool data of multiple partners regulated by data access policies. In doing so, each of the data sources can introduce its own rules and specifications on how data is used. Additionally, we add a digital consent management that will allow individual patients to decide how their data is used. Finally, we show the high potential of the cooperative analysis of distributed cancer data supported by the proposed solution in our approach.
After several years of negative phase III trials in gastric and esophageal cancer, a significant breakthrough in the treatment of metastatic adenocarcinomas of the gastroesophageal junction (GEJ) and stomach (GC) is now becoming evident with the emerging of precision oncology and implementation of molecular targets in tumor treatment. In addition, new generation studies such as umbrella and basket trials are focused on these molecular targets, which makes an early molecular diagnosis based on IHC/ISH and NGS necessary. The required companion diagnostics of Her2neu overamplification or PD-L1 expression is based on immunohistochemistry (IHC) or additionally in situ hybridization (ISH) in case of an IHC Her2neu score of 2+. However, there are investigator-dependent differences in the assessment of Her2neu amplification and different PD-L1 scoring systems obtained by IHC/ISH. The use of high-throughput technologies such as next-generation sequencing (NGS) holds the potential to standardize the analysis and thus make them more comparable. In the presented study, real-world multigene sequencing data of 72 Caucasian patients diagnosed with metastatic adenocarcinomas of GEJ and stomach were analyzed. In the clinical companion diagnostics, we found ESCAT level I molecular targets in one-third of our patients, which directly determined the therapy. In addition, we found potential targets in 14/72 patients (19.4%) who potentially qualify for precision therapies in corresponding molecular studies. The study highlights the importance of comprehensive molecular profiling for precision treatment of GEJ/GC and indicates that a biomarker evaluation should be performed for all patients with metastatic adenocarcinomas before the initiation of first-line treatment and during second-line or subsequent treatment.