Yes we care!-Certification for machine learning methods through the care label framework.

Machine learning applications have become ubiquitous. Their applications range from embedded control in production machines over process optimization in diverse areas (e.g., traffic, finance, sciences) to direct user interactions like advertising and recommendations. This has led to an increased effort of making machine learning trustworthy. Explainable and fair AI have already matured. They address the knowledgeable user and the application engineer. However, there are users that want to deploy a learned model in a similar way as their washing machine. These stakeholders do not want to spend time in understanding the model, but want to rely on guaranteed properties. What are the relevant properties? How can they be expressed to the stake- holder without presupposing machine learning knowledge? How can they be guaranteed for a certain implementation of a machine learning model? These questions move far beyond the current state of the art and we want to address them here. We propose a unified framework that certifies learning methods via care labels. They are easy to understand and draw inspiration from well-known certificates like textile labels or property cards of electronic devices. Our framework considers both, the machine learning theory and a given implementation. We test the implementation's compliance with theoretical properties and bounds.

Wrist Arthroscopy Is Effective for the Diagnosis and Treatment of Chronic Wrist Pain in Pediatric Patients.

Purpose: To determine the outcomes of diagnostic and therapeutic arthroscopy in patients with chronic wrist pain who are younger than 16 years of age. Methods: We retrospectively analyzed collected data from medical records of patients who had undergone wrist arthroscopy by the senior author between 2015 and 2017 for longstanding wrist pain and were 16 years old or younger. Findings from preoperative magnetic resonance imaging (MRI) were compared with the intraoperative diagnosis, and midterm results were gathered by a telephone interview. Results: Ten patients were included. Eight of 10 patients had undergone conservative therapy before surgery, and 60% had a history of single trauma. In 6 of 10 patients, the arthroscopic diagnosis correlated with the MRI findings. Eight of 10 wrists (80%) showed a tear of the triangular fibrocartilage complex (TFCC). Only 4 of the 8 TFCC tears were correctly identified by 3-Tesla MRI. Conclusions: In our study, wrist arthroscopy was an effective tool not only to diagnose but also to treat relevant TFCC lesions both in adolescents and children suffering from persistent wrist pain. A 3-Tesla MRI was neither sensitive nor specific enough to correctly diagnose lesions in small pediatric wrists. Level of Evidence: IV, Therapeutic case series.

Transition by intermittency in granular matter: from discontinuous avalanches to continuous flow.

We investigate, in the rotating drum configuration, the transition from the regime of discontinuous avalanches observed at low angular velocity to the regime of continuous flow observed at higher velocity. Instead of the hysteretic transition reported previously by Rajchenbach [Phys. Rev. Lett. 65, 2221 (1990)], with an apparent bistability of the two flow regimes in a range of drum velocities, we observe intermittency with spontaneous erratic switches from one regime to the other. Both scenarios of transition are recovered by a model dynamic equation for the avalanche flow with two sources of stochasticity: a Langevin noise during the avalanche flow and a distributed maximal stability angle at which avalanches start.

Dynamics of dry granular avalanches.

A fine analysis of the statistics of dry granular avalanches in a rotating drum setup reveals that, beyond the fluctuations, the angle at which an avalanche ends is correlated experimentally to the angle at which the avalanche starts. This correlation resulting from inertia defines an intermediate "neutral" angle that characterizes the corresponding granular pile. In addition, an intensive study of the dynamics of the avalanche shows that the time duration of the avalanche is correlated to its amplitude, being smaller for higher amplitude. The time relaxation of the pile slope during any avalanche, governed by the deviation of the starting angle from the neutral angle, follows a master curve. A simple model recovers most of the results and contributes to a better understanding of the physics of the avalanche flow.

Instantaneous velocity profiles during granular avalanches.

We perform experimental measurements of the instantaneous velocity profile of the flowing layer during granular avalanches. In the pile depth, the velocity profile follows a pure exponential decrease in contrast with steady flows that are known to exhibit a well developed upper linear part. The velocity profile in the pile width is a plug flow with two exponential boundary layers at the walls. Even though no steady state is observed during the avalanche, these velocity profiles are self-similar and build up almost instantaneously, with time independent characteristic lengths.