No abstract

Deep neural networks have been successfully applied in learning the board games Go, chess and shogi without prior knowledge by making use of reinforcement learning. Although starting from zero knowledge has been shown to yield impressive results, it is associated with high computationally costs especially for complex games. With this paper, we present CrazyAra which is a neural network based engine solely trained in supervised manner for the chess variant crazyhouse. Crazyhouse is a game with a higher branching factor than chess and there is only limited data of lower quality available compared to AlphaGo. Therefore, we focus on improving efficiency in multiple aspects while relying on low computational resources. These improvements include modifications in the neural network design and training configuration, the introduction of a data normalization step and a more sample efficient Monte-Carlo tree search which has a lower chance to blunder. After training on 569,537 human games for 1.5 days we achieve a move prediction accuracy of 60.4 %. During development, versions of CrazyAra played professional human players. Most notably, CrazyAra achieved a four to one win over 2017 crazyhouse world champion Justin Tan (aka LM Jann Lee) who is more than 400 Elo higher rated compared to the average player in our training set. Furthermore, we test the playing strength of CrazyAra on CPU against all participants of the second Crazyhouse Computer Championships 2017, winning against twelve of the thirteen participants. Finally, for CrazyAraFish we continue training our model on generated engine games. In ten long-time control matches playing Stockfish 10, CrazyAraFish wins three games and draws one out of ten matches.

No abstract

Aim The use of artificial intelligence represents an objective approach to increase endoscopist’s adenoma detection rate (ADR) and limit interoperator variability. In this study, we evaluated a newly developed deep convolutional neural network (DCNN) for automated detection of colorectal polyps ex vivo as well as in a first in-human trial. Methods For training of the DCNN, 116 529 colonoscopy images from 278 patients with 788 different polyps were collected. A subset of 10 467 images containing 504 different polyps were manually annotated and treated as the gold standard. An independent set of 45 videos consisting of 15 534 single frames was used for ex vivo performance testing. In vivo real-time detection of colorectal polyps during routine colonoscopy by the DCNN was tested in 42 patients in a back-to-back approach. Results When analyzing the test set of 15 534 single frames, the DCNN’s sensitivity and specificity for polyp detection and localization within the frame was 90% and 80%, respectively, with an area under the curve of 0.92. In vivo, baseline polyp detection rate and ADR were 38% and 26% and significantly increased to 50% (P = 0.023) and 36% (P = 0.044), respectively, with the use of the DCNN. Of the 13 additionally with the DCNN detected lesions, the majority were diminutive and flat, among them three sessile serrated adenomas. Conclusion This newly developed DCNN enables highly sensitive automated detection of colorectal polyps both ex vivo and during first in-human clinical testing and could potentially increase the detection of colorectal polyps during colonoscopy.

No abstract

Healthcare workers (HCW) who perform aerosol-generating procedures (AGP) are at high risk of SARS-CoV-2 infection. Data on infection rates and vaccination are limited. A nationwide, cross-sectional study focusing on AGP-related specialties was conducted between 3 May 2021 and 14 June 2021. Vaccination rates among HCW, perception of infection risk, and infection rates were analyzed, focusing on the comparison of gastrointestinal endoscopy (GIE) and other AGP-related specialties (NON-GIE), from the beginning of the pandemic until the time point of the study. Infections rates among HCW developed similarly to the general population during the course of the pandemic, however, with significantly higher infections rates among the GIE specialty. The perceived risk of infection was distributed similarly among HCW in GIE and NON-GIE (91.7%, CI: 88.6–94.4 vs. 85.8%, CI: 82.4–89.0; p < 0.01) with strongest perceived threats posed by AGPs (90.8%) and close patient contact (70.1%). The very high vaccination rate (100–80%) among physicians was reported at 83.5%, being significantly more frequently reported than among nurses (56.4%, p < 0.01). GIE had more often stated very high vaccination rate compared with NON-GIE (76.1% vs. 65.3%, p < 0.01). A significantly higher rate of GIE was reported to have fewer concerns regarding infection risk after vaccination than NON-GIE (92.0% vs. 80.3%, p < 0.01).

Objective: Healthcare workers (HCW) are at high risk of SARS-CoV-2 infection due to exposure to potentially infectious material, especially during aerosol-generating procedures (AGP). We aimed to investigate the prevalence of infection among HCW in medical disciplines with AGP. Design: A nationwide questionnaire-based study in in- and outpatient settings was conducted between 12/16/2020 and 01/24/2021. Data on SARS-CoV-2 infections among HCW and potential risk factors were investigated. Results: 2,070 healthcare facilities with 25,113 employees were included in the study. Despite a higher rate of pre-interventional testing, clinics treated three times more confirmed SARS-CoV-2 cases than private practices (28.8% vs. 88.4%, p<0.001). Overall infection rate among HCW accounted for 4.7%. Multivariate analysis revealed that ZIP-regions having comparably higher incidences were significantly associated with increased risk of infection. Furthermore, clinical setting and the GIE specialty have more than double the risk of infection (OR 2.63; 95% CI 2.501-2.817, p<0.01 and OR 2.35; 95% CI 2.245-2.498, p<0.01). The number of procedures performed per day was also significantly associated with an increased risk of infection (OR 1.01; 95% CI 1.007-1.014), p<0.01). No treatment of confirmed SARS-CoV-2 cases was tending to lower the risk of infection (OR 0.72; 95% CI 0.507-1.025, p=0.068). Conclusion: HCW in GIE seem to be at higher risk of infection than those in other AGP, especially in the clinical setting. Regions having comparably higher incidences as well as the number of procedures performed per day were also significantly associated with increased risk of infection.

No abstract