In an article recently published in the journal ACS Applied Materials & Interfaces, using machine learning (ML) and Wannier analysis, an effective divide-and-conquer strategy is proposed to develop Hamiltonian of the system. Study: All-Atom Nonadiabatic Dynamics Simulation of Hybrid Graphene Nanoribbons Based on Wannier Analysis and Machine Learning. Several essential excitons and electron dynamic mechanisms, such as charge carrier movement, relaxing and diffusing excitons, segregation of electrons and holes, and recombination, are critical aspects in photovoltaic conversion devices, including solar panels, field-effect semiconductors, and LEDs. The dynamics of excitons and electrons are closely related to nuclear movements in these circumstances, and hence all fall under the framework of diabatic dynamics. In general, traditional molecular dynamics does not describe fundamental quantum phenomena, but pure quantum dynamics has enormous processing requirements when a significant number of degrees of freedom (DOFs) are concerned.

The wireless network communication mode represented by the Internet of vehicles (IoV) has been widely used. However, due to the limitations of traditional network architecture, resource scheduling in wireless network environment is still facing great challenges. This paper focuses on the allocation of bandwidth resources in the virtual network environment. This paper proposes a bandwidth aware multi domain virtual network embedding algorithm (BA-VNE). The algorithm is mainly aimed at the problem that users need a lot of bandwidth in wireless communication mode, and solves the problem of bandwidth resource allocation from the perspective of virtual network embedding (VNE). In order to improve the performance of the algorithm, we introduce particle swarm optimization (PSO) algorithm to optimize the performance of the algorithm. In order to verify the effectiveness of the algorithm, we have carried out simulation experiments from link bandwidth, mapping cost and virtual network request (VNR) acceptance rate. The final results show that the proposed algorithm is better than other representative algorithms in the above indicators.

Traditional ground wireless communication networks cannot provide high-quality services for artificial intelligence (AI) applications such as intelligent transportation systems (ITS) due to deployment, coverage and capacity issues. The space-air-ground integrated network (SAGIN) has become a research focus in the industry. Compared with traditional wireless communication networks, SAGIN is more flexible and reliable, and it has wider coverage and higher quality of seamless connection. However, due to its inherent heterogeneity, time-varying and self-organizing characteristics, the deployment and use of SAGIN still faces huge challenges, among which the orchestration of heterogeneous resources is a key issue. Based on virtual network architecture and deep reinforcement learning (DRL), we model SAGIN's heterogeneous resource orchestration as a multi-domain virtual network embedding (VNE) problem, and propose a SAGIN cross-domain VNE algorithm. We model the different network segments of SAGIN, and set the network attributes according to the actual situation of SAGIN and user needs. In DRL, the agent is acted by a five-layer policy network. We build a feature matrix based on network attributes extracted from SAGIN and use it as the agent training environment. Through training, the probability of each underlying node being embedded can be derived. In test phase, we complete the embedding process of virtual nodes and links in turn based on this probability. Finally, we verify the effectiveness of the algorithm from both training and testing.

Surrogate gradient (SG) training provides the possibility to quickly transfer all the gains made in deep learning to neuromorphic computing and neuromorphic processors, with the consequent reduction in energy consumption. Evidence supports that training can be robust to the choice of SG shape, after an extensive search of hyper-parameters. However, random or grid search of hyper-parameters becomes exponentially unfeasible as we consider more hyper-parameters. Moreover, every point in the search can itself be highly time and energy consuming for large networks and large datasets. In this article we show how complex tasks and networks are more sensitive to SG choice. Secondly, we show how low dampening, high sharpness and low tail fatness are preferred. Thirdly, we observe that Glorot Uniform initialization is generally preferred by most SG choices, with variability in the results. We finally provide a theoretical solution to reduce the need of extensive gridsearch, to find SG shape and initializations that result in improved accuracy.

Many of the successes of machine learning are based on minimizing an averaged loss function. However, it is well-known that this paradigm suffers from robustness issues that hinder its applicability in safety-critical domains. These issues are often addressed by training against worst-case perturbations of data, a technique known as adversarial training. Although empirically effective, adversarial training can be overly conservative, leading to unfavorable trade-offs between nominal performance and robustness. To this end, in this paper we propose a framework called probabilistic robustness that bridges the gap between the accurate, yet brittle average case and the robust, yet conservative worst case by enforcing robustness to most rather than to all perturbations. From a theoretical point of view, this framework overcomes the trade-offs between the performance and the sample-complexity of worst-case and average-case learning. From a practical point of view, we propose a novel algorithm based on risk-aware optimization that effectively balances average- and worst-case performance at a considerably lower computational cost relative to adversarial training. Our results on MNIST, CIFAR-10, and SVHN illustrate the advantages of this framework on the spectrum from average- to worst-case robustness.

Under the presence of these two types of variables, the goal is to identify the design variables that optimize the black-box function by taking into account the uncertainty of environmental variables. In the past few years, Bayesian Optimization (BO) framework that takes the uncertain environmental variables into considerations have been studied in various setups (see §1.1). In this paper, we study one of such problems called distributionally robust chance-constrained (DRCC) problem. The DRCC problem is an instance of constrained optimization problems in an uncertain environment, which is important in a variety of practical problems in science and engineering. The goal of a CC problem is to identify the design variables that maximize the expectation of the objective function under the constraint that the probability of the constraint function exceeding a given threshold is greater than a certain level. Let f(x, w) and g(x, w) be the unknown objective and constraint functions, respectively, both of which depend on the design variables x X and the environmental variables w Ω.

The concept of smart cities is founded on sophisticated cellular networks that would not only connect humans in the future but also humans to other smart devices. However, this would also require huge energy consumption. In the wake of climate change, this can make matters worse for our environment by increasing the greenhouse gas emissions. Thus, we not only need smart cities but also greener smart cities. One way to address this issue is by switching off base stations (BSs), radio transmitters/receivers that serve as the hub of the local wireless network, when they have little to no traffic load.

Scatter due to interaction of photons with the imaged object is a fundamental problem in X-ray Computed Tomography (CT). It manifests as various artifacts in the reconstruction, making its abatement or correction critical for image quality. Despite success in specific settings, hardware-based methods require modification in the hardware, or increase in the scan time or dose. This accounts for the great interest in software-based methods, including Monte-Carlo based scatter estimation, analytical-numerical, and kernel-based methods, with data-driven learning-based approaches demonstrated recently. In this work, two novel physics-inspired deep-learning-based methods, PhILSCAT and OV-PhILSCAT, are proposed. The methods estimate and correct for the scatter in the acquired projection measurements. Different from previous works, they incorporate both an initial reconstruction of the object of interest and the scatter-corrupted measurements related to it, and use a deep neural network architecture and cost function, both specifically tailored to the problem. Numerical experiments with data generated by Monte-Carlo simulations of the imaging of phantoms reveal consistent improvement over a recent purely projection-domain deep neural network scatter correction method.

Optimization is a ubiquitous modeling tool and is often deployed in settings which repeatedly solve similar instances of the same problem. Amortized optimization methods use learning to predict the solutions to problems in these settings. This leverages the shared structure between similar problem instances. In this tutorial, we will discuss the key design choices behind amortized optimization, roughly categorizing 1) models into fully-amortized and semi-amortized approaches, and 2) learning methods into regression-based and objectivebased. We then view existing applications through these foundations to draw connections between them, including for manifold optimization, variational inference, sparse coding, meta-learning, control, reinforcement learning, convex optimization, and deep equilibrium networks. This framing enables us easily see, for example, that the amortized inference in variational autoencoders is conceptually identical to value gradients in control and reinforcement learning as they both use fully-amortized models with an objective-based loss.

The Swedish musician and AI enthusiast known as Supercomposite used an AI to create hundreds of creepy new tarot cards -- and has been blasting them on Twitter for days, in a delightful barrage of occult-flavored machine learning. The artist is using an AI called Looking Glass, which debuted last year and was made by Twitter user ai.curio. Some cards have humanoid characters with holes for faces, some feature monstrous-looking creatures in bloody shades of red, and some are creepy simply because they seem uncannily like tarot cards at first glance. These tarot cards do not exist. I generated 500 of these and I'm not stopping.