Competition-level code generation with AlphaCode.

Programming is a powerful and ubiquitous problem-solving tool. Systems that can assist programmers or even generate programs themselves could make programming more productive and accessible. Recent transformer-based neural network models show impressive code generation abilities yet still perform poorly on more complex tasks requiring problem-solving skills, such as competitive programming problems. Here, we introduce AlphaCode, a system for code generation that achieved an average ranking in the top 54.3% in simulated evaluations on recent programming competitions on the Codeforces platform. AlphaCode solves problems by generating millions of diverse programs using specially trained transformer-based networks and then filtering and clustering those programs to a maximum of just 10 submissions. This result marks the first time an artificial intelligence system has performed competitively in programming competitions.

Grandmaster level in StarCraft II using multi-agent reinforcement learning.

Many real-world applications require artificial agents to compete and coordinate with other agents in complex environments. As a stepping stone to this goal, the domain of StarCraft has emerged as an important challenge for artificial intelligence research, owing to its iconic and enduring status among the most difficult professional esports and its relevance to the real world in terms of its raw complexity and multi-agent challenges. Over the course of a decade and numerous competitions1-3, the strongest agents have simplified important aspects of the game, utilized superhuman capabilities, or employed hand-crafted sub-systems4. Despite these advantages, no previous agent has come close to matching the overall skill of top StarCraft players. We chose to address the challenge of StarCraft using general-purpose learning methods that are in principle applicable to other complex domains: a multi-agent reinforcement learning algorithm that uses data from both human and agent games within a diverse league of continually adapting strategies and counter-strategies, each represented by deep neural networks5,6. We evaluated our agent, AlphaStar, in the full game of StarCraft II, through a series of online games against human players. AlphaStar was rated at Grandmaster level for all three StarCraft races and above 99.8% of officially ranked human players.

Effect of Attentional Focus Strategies on the Biomechanical Performance of the Drop Jump.

Comyns, TM, Brady, CJ, and Molloy, J. Effect of attentional focus strategies on the biomechanical performance of the drop jump. J Strength Cond Res 33(3): 626-632, 2019-Motor performance can be influenced by focusing an athlete's attention through the use of verbal instructions. There is limited research on the effect of internal, neutral, and external attentional focus strategies on drop jump (DJ) performance aimed at maximizing height jumped (HJ) and minimizing ground contact time (CT). The purpose of this study was to assess the effect of attentional focus strategies on biomechanical variables related to efficient DJ performance, namely HJ, CT, reactive strength index (RSI), leg-spring stiffness, and peak and relative peak ground reaction force (GRF). Seventeen male recreationally trained subjects performed 2 DJs after listening to instructions designed to evoke an internal, external, or neutral attentional focus. In total, 6 DJs were performed in the testing session, and the order of the instructions was randomly assigned. Significance was set at p ≤ 0.05. Results indicated that, compared with the neutral strategy, the external focus resulted in significantly higher RSI (p = 0.046), peak GRF (p = 0.025), relative GRF (p = 0.02), and leg-spring stiffness (p = 0.02). No significant difference was seen in DJ CT and HJ between all 3 conditions (p ≥ 0.05). These results indicate that the use of an external focus of attention may potentially result in a more effective and efficient fast stretch-shortening cycle performance because of the augmentation of RSI and leg stiffness. More research is warranted, however, because of the lack of significant results pertaining to CT and HJ.

Obesity should not be considered a contraindication to medial Oxford UKA: long-term patient-reported outcomes and implant survival in 1000 knees.

PURPOSE: Some health providers ration knee arthroplasty on the basis of body mass index (BMI). There is no long-term data on the outcome of medial mobile-bearing unicompartmental knee arthroplasty (UKA) in different BMI groups. This study aimed to determine the effect of patient body mass index (BMI) on patient-reported outcomes and long-term survival of medial UKA in a large non-registry cohort. Our hypothesis is that increasing BMI would be associated with worse outcomes. METHODS: Data were analysed from a prospective cohort of 1000 consecutive medial mobile-bearing Oxford UKA with mean 10-year follow-up. Patients were grouped: BMI < 25, BMI 25 to < 30, BMI 30 to < 35 and BMI 35+. Oxford Knee Score (OKS) and Tegner Activity Score were assessed at 1, 5 and 10 years. Kaplan-Meier survivorship was calculated and compared between BMI groups. RESULTS: All groups had significant improvement in OKS and Tegner scores. BMI 35 + kg/m2 experienced the greatest overall increase in mean OKS of 17.3 points (p = 0.02). There was no significant difference in ten-year survival, which was, from lowest BMI group to highest 92%, 95%, 94% and 93%. CONCLUSION: There was no difference in implant survival between groups, and although there was no consistent trend in postoperative OKS, the BMI 35+ group benefited the most from UKA. Therefore, when UKA is used for appropriate indications, high BMI should not be considered to be a contraindication. Furthermore rationing based on BMI seems unjustified, particularly when the commonest threshold (BMI 35) is used. LEVEL OF EVIDENCE: III.