Abstract:This paper focuses on cyber-security simulations in networks modeled as a Markov game with incomplete information and stochastic elements. The resulting game is an adversarial sequential decision making problem played with two agents, the attacker and defender. The two agents pit one reinforcement learning technique, like neural networks, Monte Carlo learning and Q-learning, against each other and examine their effectiveness against learning opponents. The results showed that Monte Carlo learning with the Softmax exploration strategy is most effective in performing the defender role and also for learning attacking strategies.

As molecular dynamics simulations increase in complexity, new analysis tools are necessary to facilitate interpreting the results. Lipids, for instance, are known to form many complicated morphologies, because of their amphipathic nature, becoming more intricate as the particle count increases. A few lipids might form a micelle, where aggregation of tens of thousands could lead to vesicle formation. Millions of lipids comprise a cell and its organelle membranes, and are involved in processes such as neurotransmission and transfection. To study such phenomena, it is useful to have analysis tools that understand what is meant by emerging entities such as micelles and vesicles. Studying such systems at the particle level only becomes extremely tedious, counterintuitive, and computationally expensive. To address this issue, we developed a method to track all the individual lipid leaflets, allowing for easy and quick detection of topological changes at the mesoscale. By using a voxel-based approach and focusing on locality, we forego costly geometrical operations without losing important details and chronologically identify the lipid segments using the Jaccard index. Thus, we achieve a consistent sequential segmentation on a wide variety of (lipid) systems, including monolayers, bilayers, vesicles, inverted hexagonal phases, up to the membranes of a full mitochondrion. It also discriminates between adhesion and fusion of leaflets. We show that our method produces consistent results without the need for prefitting parameters, and segmentation of millions of particles can be achieved on a desktop machine.

Computational chemistry has become an important tool to predict and understand molecular properties and reactions. Even though recent years have seen a significant growth in new algorithms and computational methods that speed up quantum chemical calculations, the bottleneck for trajectorybased methods to study photo-induced processes is still the huge number of electronic structure calculations. In this work, we present an innovative solution, in which the amount of electronic structure calculations is drastically reduced, by employing machine learning algorithms and methods borrowed from the realm of artificial intelligence. However, applying these algorithms effectively requires finding optimal hyperparameters, which remains a challenge itself. Here we present an automated user-friendly framework, HOAX, to perform the hyperparameter optimisation for neural networks, which bypasses the need for a lengthy manual process. The neural network-generated potential energy surfaces (PESs) reduce the computational costs compared to the ab initio-based PESs. We perform a comparative investigation on the performance of different hyperparameter optimiziation algorithms, namely grid search, simulated annealing, genetic algorithm, and Bayesian optimizer in finding the optimal hyperparameters necessary for constructing the well-performing neural network in order to fit the PESs of small organic molecules. Our results show that this automated toolkit not only facilitate a straightforward way to perform the hyperparameter optimisation but also the resulting neural networks-based generated PESs are in reasonable agreement with the ab initio-based PESs.

Computational chemistry has become an important tool to predict and understand molecular properties and reactions. Even though recent years have seen a significant growth in new algorithms and computational methods that speed up quantum chemical calculations, the bottleneck for trajectory-based methods to study photoinduced processes is still the huge number of electronic structure calculations. In this work, we present an innovative solution, in which the amount of electronic structure calculations is drastically reduced, by employing machine learning algorithms and methods borrowed from the realm of artificial intelligence. However, applying these algorithms effectively requires finding optimal hyperparameters, which remains a challenge itself. Here we present an automated user-friendly framework, HOAX, to perform the hyperparameter optimization for neural networks, which bypasses the need for a lengthy manual process. The neural network generated potential energy surfaces (PESs) reduces the computational costs compared to the ab initio-based PESs. We perform a comparative investigation on the performance of different hyperparameter optimiziation algorithms, namely grid search, simulated annealing,genetic algorithm, and bayesian optimizer in finding the optimal hyperparameters necessary for constructing the well-performing neural network in order to fit the PESs of small organic molecules. Our results show that this automated toolkit not only facilitates a straightforward way to perform the hyperparameter optimization but also the resulting neural networks-based generated PESs are in reasonable agreement with the ab initio-based PESs.