Iterative Compilation Optimization Based on Metric Learning and Collaborative Filtering

Pass selection and phase ordering are two critical compiler auto-tuning problems. Traditional heuristic methods cannot effectively address these NP-hard problems especially given the increasing number of compiler passes and diverse hardware architectures. Recent research efforts have attempted to address these problems through machine learning. However, the large search space of candidate pass sequences, the large numbers of redundant and irrelevant features, and the lack of training program instances make it difficult to learn models well. Several methods have tried to use expert knowledge to simplify the problems, such as using only the compiler passes or subsequences in the standard levels (e.g., -O1, -O2, and -O3) provided by compiler designers. However, these methods ignore other useful compiler passes that are not contained in the standard levels. Principal component analysis (PCA) and exploratory factor analysis (EFA) have been utilized to reduce the redundancy of feature data. However, these unsupervised methods retain all the information irrelevant to the performance of compilation optimization, which may mislead the subsequent model learning. To solve these problems, we propose a compiler pass selection and phase ordering approach, called Iterative Compilation based on Metric learning and Collaborative filtering (ICMC) . First, we propose a data-driven method to construct pass subsequences according to the observed collaborative interactions and dependency among passes on a given program set. Therefore, we can make use of all available compiler passes and prune the search space. Then, a supervised metric learning method is utilized to retain useful feature information for compilation optimization while removing both the irrelevant and the redundant information. Based on the learned similarity metric, a neighborhood-based collaborative filtering method is employed to iteratively recommend a few superior compiler passes for each target program. Last, an iterative data enhancement method is designed to alleviate the problem of lacking training program instances and to enhance the performance of iterative pass recommendations. The experimental results using the LLVM compiler on all 32 cBench programs show the following: (1) ICMC significantly outperforms several state-of-the-art compiler phase ordering methods, (2) it performs the same or better than the standard level -O3 on all the test programs, and (3) it can reach an average performance speedup of 1.20 (up to 1.46) compared with the standard level -O3.

Analysis of the Effectiveness of Different Types of Distance Learning

<p style="text-align: justify;">The purpose of this paper is to determine the relevance of turning the camera on or off during distance learning as an argument for active or passive student participation. Seventy-five (75) students participated in the study and were divided into five groups (1-5) according to teaching method (i.e., synchronous instruction online with camera (1) and without camera (2), synchronous transmission of the recording online with camera (3) and without camera (4) and received the online instruction (5)) only. In the beginning and at the end, all students were tested with the same adapted test to determine general physical and motor status. All groups had the same training program twice a week for 45 minutes for 7 weeks. The first training of the week was dedicated to strength development, the second to endurance. In the end, all participants completed a questionnaire to determine their additional physical activity and how they felt about using a camera. The camera being turned on was identified as a factor that made participants uncomfortable but contributed significantly to the effectiveness of the course. However, 94.6 % of all participants cited non-camera methods as their favourite.</p>

Comprehensive supervised heavy training program versus home training regimen in patients with subacromial impingement syndrome: a randomized trial

Background There is no consensus on the best training regimen for subacromial impingement syndrome (SIS). Several have been suggested, but never tested. The purpose of the study is to compare a comprehensive supervised training regimen (STR) based on latest evidence including heavy slow resistance training with a validated home-based regimen (HTR). We hypothesized that the STR would be superior to the HTR. Methods Randomised control trial with blinded assessor. 126 consecutive patients with SIS were recruited and equally randomised to 12 weeks of either supervised training regimen (STR), or home-based training regimen (HTR). Primary outcomes were Constant Score (CS) and Shoulder Rating Questionnaire (SRQ) from baseline and 6 months after completed training. Results were analyzed according to intention-to treat principles. The study was retrospectively registered in ClinicalTrials.gov. Date of registration: 07/06/2021. Identification number: NCT04915430. Results CS improved by 22.7 points for the STR group and by 23,7 points for the HTR (p = 0.0001). The SRQ improved by 17.7 and 18.1 points for the STR and the HTR groups respectively (p = 0.0001). The inter-group changes were non-significant. All secondary outcomes (passive and active range of motion, pain on impingement test, and resisted muscle tests) improved in both groups, without significant inter-group difference. Conclusion We found no significant difference between a comprehensive supervised training regimen including heavy training principles, and a home-based training program in patients with SIS.

Effects of a training program for occupational health professionals on the cognitions and perceptions of workers: a randomized controlled trial

Purpose To evaluate the effects of a training program for occupational health professionals (OHPs) on their ability to identify the cognitions and perceptions of workers with a chronic disease that may hinder work participation, and on their ability to recommend evidence-based interventions aimed at the identified cognitions and perceptions. Methods A randomized controlled trial was conducted in which OHPs were randomly assigned to a training program on the cognitions and perceptions of workers with a chronic disease (n = 29) or to a control group that did not receive training (n = 30). Participants received home assignments in which they had to identify the cognitions and perceptions of workers in video vignettes and had to indicate which interventions they would recommend to foster work participation. A generalized linear model repeated measures ANOVA was conducted to study the effects of the training program. Results The results of the analyses showed an increase in the ability to identify the cognitions and perceptions of workers of OHPs who received the training compared to the control group (p < 0.001). The results also showed an increased ability to recommend evidence-based interventions aimed at these cognitions and perceptions (p < 0.001) as a result of participation in the training. Conclusion The training program helps OHPs to identify cognitions and perceptions and to recommend evidence-based interventions. This can support them in their activities to increase the work participation of workers with a chronic disease.

EMERGING ADULTHOOD: REPRODUCTIVE HEALTH ANDREPRODUCTIVEATTITUDES

The objective assessments of the reproductive health of students of both sexes, enrolled in the bachelor's training program, belonging to the group «emerging adulthood», obtained from the results of medical examinations, are considered. The data on reproductive behavior and reproductive attitudes of students, obtained in the course of a pilot sociological study, are analyzed. A comparison is made with the available literature data, and further in-depth studies are substantiated

Brief online mindfulness training for medical students: a randomized control study

BACKGROUNDMedical students experience high levels of stress during their training. Literature suggests that mindfulness can reduce stress and increase self-compassion levels in medical students. However, most mindfulness training programs are delivered face-to-face and require significant time commitments, which can be challenging for rurally-based students with heavy academic workloads and limited support networks. PURPOSETo evaluate the feasibility and efficacy of a brief online Mindfulness training program for medical students based in rural areas, with regards to reducing stress, increasing self-compassion, mindfulness and study engagement. METHODSThis is a non-registered randomised control trial. Participants included both urban and rural medical students from UWA, University of Notre Dame and the RCSWA from 2018-2020. Participants were randomised to the intervention group, an 8-week online mindfulness training program, or the control group. Using quantitative-qualitative mixed-methods approach, we measured the frequency, duration and quality of the participants mindfulness meditation practice, and assessed changes in their perceived stress, self-compassion, mindfulness and study engagement levels. Further, the intervention group recorded a weekly reflective journal documenting their experience of the program. RESULTS114 participants were recruited to the study. 61 were randomised to the intervention, and 53 to the control. Quantitative analysis of the frequency, duration and quality of mindfulness meditation practice and changes in stress, self-compassion, mindfulness and study engagement is currently being conducted. Preliminary qualitative results reveal that participants experienced increased self-awareness, more mindfulness of their day-to-day activities, improved emotional regulation and increased productivity, while also facing difficulties with making time for their mindfulness practice. CONCLUSIONWe anticipate that this study will demonstrate that an online mindfulness training program tailored to reach rurally located medical students is feasible and effective in modifying their stress levels and psychological wellbeing.

The Effect of 16 Weeks of Lower-Limb Strength Training in Jumping Performance of Ballet Dancers

Jumping ability is considered a determinant of performance success. It is identified as one of the predictors and talent identification in many sports and dance. This study aimed to investigate the effect of 16 weeks of lower-limb strength training on the jumping performance of ballet dancers. A total of 24 participants from the same dance school were randomly selected in the control group [CG; n = 10; aged 13.00 (1.49) years; 43.09 (9.48) kg and 1.53 (0.11) m] and the intervention group [IG; n = 14; aged 12.43 (1.45) years; 38.21 (4.38) kg and 1.51 (0.07) m], evaluated before and after the applied strength training program mainly using the body weight of each participant. Jump performance was assessed using MyJump2, a scientifically validated mobile phone app. Intergroup and intragroup comparisons were assessed, and the magnitude of change was calculated using the effect size (ES). While CG significantly decreased the relative power over time (p &lt; 0.001, ES = −0.29: small), results from the intragroup comparisons suggest that IG significantly increased the countermovement jump (CMJ) height (p &lt; 0.001, ES = 1.21: large), the relative force (p &lt; 0.001, ES = 0.86: moderate), maximal velocity (p &lt; 0.001, ES = 1.15: moderate), and relative power (p &lt; 0.001, ES = 1.37: large). We concluded that a 16-week strength training program of lower limbs is an effective way to improve CMJ height in young dancers. Supplementary strength training appears to be the determinant for the improvement of the jumping performance of ballet dancers.