ME-MCTS: Online Generalization by Combining Multiple Value Estimators

This paper addresses the challenge of online generalization in tree search. We propose Multiple Estimator Monte Carlo Tree Search (ME-MCTS), with a two-fold contribution: first, we introduce a formalization of online generalization that can represent existing techniques such as "history heuristics", "RAVE", or "OMA" -- contextual action value estimators or abstractors that generalize across specific contexts. Second, we incorporate recent advances in estimator averaging that enable guiding search by combining the online action value estimates of any number of such abstractors or similar types of action value estimators. Unlike previous work, which usually proposed a single abstractor for either the selection or the rollout phase of MCTS simulations, our approach focuses on the combination of multiple estimators and applies them to all move choices in MCTS simulations. As the MCTS tree itself is just another value estimator -- unbiased, but without abstraction -- this blurs the traditional distinction between action choices inside and outside of the MCTS tree. Experiments with three abstractors in four board games show significant improvements of ME-MCTS over MCTS using only a single abstractor, both for MCTS with random rollouts as well as for MCTS with static evaluation functions. While we used deterministic, fully observable games, ME-MCTS naturally extends to more challenging settings.

Didactic materials: Mannequin for canine abdominal and vaginal palpation

Compared to the veterinary education system decades ago, ethical and welfare questions are of major relevance in the current scenario. Therefore, certain learning practices involving in vivo animal studies are no longer performed. To overcome this limitation, the use of mannequins has emerged as an interactive and immersive method of learning that simulates a real situation. The aim of this work is development mannequins for improve physiology of the reproductive apparatus of a bitch. Here, we used a plastic canine prototype to simulate the abdominal and vaginal palpation technique. An abdominal incision was made, filled with synthetic fiber, and a fictional uterus was placed. In the same mannequin, an orifice was created that correspond to the vagina. This model greatly helped in training students in performing palpation techniques, conducting the static evaluation, and assessing maternal-fetal ratio. Moreover, the physiology of the reproductive canine prototype made possible the demonstration of abdominal and vaginal palpation. This practice of vaginal palpation was found to be of great value since several students could easily take the examination and identify different stages of canine gestation, which could hardly be achieved using animals. Mannequins have received good student acceptance owing to their safety and ease to use, and resulted in a better understanding of the procedure to perform future in vivo animal studies.

How to Define Pediatric Flatfoot: Comparison of 2 Methods: Foot Posture in Static and Dynamic Conditions in Children 5 to 9 Years Old

Flexible flatfoot is the most common posture deformity among children. There are many diagnostic techniques to identify it, from clinical observation to measurements and imaging techniques, both in static and in dynamic conditions, but their reliability, validity, and accuracy are still unproven. The static and dynamic conditions differ; thus, the aim was to compare the results of the evaluation of flatfeet in 50 children (5-9 years of age) in both conditions: standing versus gait. Static evaluation. Comparison of the footprints (Harris and Beath pedograph) with the Clarke’s footprinting graphics. Dynamic evaluation. On the same day, all children underwent pedobarography during gait. Geometric measures of the feet (midfoot width, instep width, instep, foot width) were calculated together with the Arch Index. In the static condition, 87 of 100 were classified as a flatfoot, whereas during walking, there were just 56 feet classified as flat. So 35 feet classified on the basis of the clinical examination and Clarke’s footprint chart as flatfeet, according to the Arch Index calculated during walking were not flat, and 4 feet classified on the basis of Clarke’s footprint chart as normal according to the Arch Index were flat. Levels of Evidence: Prospective cohort study

An Algorithm of Tracking and Controlling Network Attack Node based on Adaptive Neural Networks

In order to obtain certain and comprehensive information for formulating network attack strategy, a complex network attack method is proposed in this paper. The attackers’ income, loss, cost and encountered risk in network attack are analyzed and index system is established to evaluate attack effect of network node with dynamic Bayesian network. This method can overcome defects of static evaluation which is relied on single index of network topology. Simulation experiment shows that this method combines more nodes and observation during the attack. It can avoid the gap between actual attack effect and theoretical expectation when attack is implemented by relying on static evaluation. In the meanwhile, it is more accurate in attack precision and of high attack efficiency

NATM TUNNELS – CONSIDERATION OF THE PARTLY DAMAGED PRIMARY LINING IMPACT FOR THE SECONDARY LINING EVALUATION

Tunnels excavated by the New Austrian Tunnelling Method (NATM) are traditionally supported by a double-shell lining. The primary lining is generated from reinforced sprayed concrete with rockbolts, the secondary lining is generated from in-situ cast concrete and it is protected by a waterproof membrane. The static evaluation of the long term behaviour generally assumes full degradation of the primary lining (the primary lining impact is not considered). Some investigations of tunnels constructed in past indicated that the primary lining degradation is slower than expected. This paper discusses possibilities of partly damaged primary lining consideration in static calculations of the secondary lining. Various options were verified by numerical modelling, and its results were compared and discussed.