OBJECTIVE: Federated Learning (FL) enables collaborative model training while keeping data locally. Currently, most FL studies in radiology are conducted in simulated environments due to numerous hurdles impeding its translation into practice. The few existing real-world FL initiatives rarely communicate specific measures taken to overcome these hurdles. To bridge this significant knowledge gap, we propose a comprehensive guide for real-world FL in radiology. Minding efforts to implement real-world FL, there is a lack of comprehensive assessments comparing FL to less complex alternatives in challenging real-world settings, which we address through extensive benchmarking. MATERIALS AND METHODS: We developed our own FL infrastructure within the German Radiological Cooperative Network (RACOON) and demonstrated its functionality by training FL models on lung pathology segmentation tasks across six university hospitals. Insights gained while establishing our FL initiative and running the extensive benchmark experiments were compiled and categorized into the guide. RESULTS: The proposed guide outlines essential steps, identified hurdles, and implemented solutions for establishing successful FL initiatives conducting real-world experiments. Our experimental results prove the practical relevance of our guide and show that FL outperforms less complex alternatives in all evaluation scenarios. DISCUSSION AND CONCLUSION: Our findings justify the efforts required to translate FL into real-world applications by demonstrating advantageous performance over alternative approaches. Additionally, they emphasize the importance of strategic organization, robust management of distributed data and infrastructure in real-world settings. With the proposed guide, we are aiming to aid future FL researchers in circumventing pitfalls and accelerating translation of FL into radiological applications.

Continual learning (CL) methods designed for natural image classification often fail to reach basic quality standards for medical image segmentation. Atlas-based segmentation, a well-established approach in medical imaging, incorporates domain knowledge on the region of interest, leading to semantically coherent predictions. This is especially promising for CL, as it allows us to leverage structural information and strike an optimal balance between model rigidity and plasticity over time. When combined with privacy-preserving prototypes, this process offers the advantages of rehearsal-based CL without compromising patient privacy. We propose Atlas Replay, an atlas-based segmentation approach that uses prototypes to generate high-quality segmentation masks through image registration that maintain consistency even as the training distribution changes. We explore how our proposed method performs compared to state-of-the-art CL methods in terms of knowledge transferability across seven publicly available prostate segmentation datasets. Prostate segmentation plays a vital role in diagnosing prostate cancer, however, it poses challenges due to substantial anatomical variations, benign structural differences in older age groups, and fluctuating acquisition parameters. Our results show that Atlas Replay is both robust and generalizes well to yet-unseen domains while being able to maintain knowledge, unlike end-to-end segmentation methods. Our code base is available under https://github.com/MECLabTUDA/Atlas-Replay.

Interacting with a conversational agent (CA) is becoming a major paradigm for human-technology interaction. Yet, ways for interacting with CAs are still forming, especially in situations involving more than one human. Starting an interaction with a CA might involve a wakeword and command. Alternatively, it could become active based on implicit requests and context information. Hence, CA designers face a serious dilemma: explicit commands disturb a natural conversation flow, while implicit requests might cause inadequate CA behavior. This study explores this dilemma and discusses observations from a project featuring a CA for financial advisory services. Advisors initially envisioned a CA that ''blends with the background'' and acts on context information. However, when engaging with a CA, they used conversational interactions in one part of the encounter and command-based interactions in another. We discuss this observation and contrast it against previous literature. This insight has implications for design and research.

Conversational agents (CA) have drawn increasing interest from HCI research. They have become popular in different aspects of our lives, for example, in the form of chatbots as the primary point of contact when interacting with an insurance company online. Additionally, CA find their way into collaborative settings in education, at work, or financial advisory. Researchers and practitioners are searching for ways to enhance the customer's experience in service encounters by deploying CA. Since competence is an important treat of a financial advisor, they only accept CA in their interaction with clients if it does not harm their impression on the client. However, we do not know how the social presence of the CA affects this perceived competence. We explore this by evaluating three prototypes with different social presences. For this, we conducted a video-based online survey. In contrast to prior studies focusing on single human-computer interaction, our study explores CA in a dyadic setting of two humans and one CA. First, our results support the Computers-Are-Social-Actors paradigm as the CA with a strong social presence was perceived as more competent than the other two designs. Second, our data show a positive correlation between CA's and advisor's competence. This implies a positive impact of the CA on the service encounter as the CA and advisor can be seen as a competent team.