Fumbling to find flashlights during blackouts may soon be a distant memory, as quantum computing and artificial intelligence could learn to decipher an electric grid's problematic quirks and solve system hiccups so fast, humans may not notice. Rather than energy grid faults turning into giant problems--such as voltage variations or widespread blackouts--blazing fast computation blended with artificial intelligence could rapidly diagnose trouble and find solutions in tiny splits of seconds, according to Cornell research forthcoming in Applied Energy (Dec. 1, 2021). "Energy power system failures are an old problem and we are still using classic computational methods to resolve them," said Fengqi You, the Roxanne E. and Michael J. Zak Professor in Energy Systems Engineering in the College of Engineering. "Today's power systems can benefit from AI and the computational power of quantum computing, so power systems can be stable and reliable." You, along with doctoral student Akshay Ajagekar, are co-authors of "Quantum Computing-based Hybrid Deep Learning for Fault Diagnosis in Electrical Power Systems."

Fully Connected Neural Network (FCNN) is a class of Artificial Neural Networks widely used in computer science and engineering, whereas the training process can take a long time with large datasets in existing many-core systems. Optical Network-on-Chip (ONoC), an emerging chip-scale optical interconnection technology, has great potential to accelerate the training of FCNN with low transmission delay, low power consumption, and high throughput. However, existing methods based on Electrical Network-on-Chip (ENoC) cannot fit in ONoC because of the unique properties of ONoC. In this paper, we propose a fine-grained parallel computing model for accelerating FCNN training on ONoC and derive the optimal number of cores for each execution stage with the objective of minimizing the total amount of time to complete one epoch of FCNN training. To allocate the optimal number of cores for each execution stage, we present three mapping strategies and compare their advantages and disadvantages in terms of hotspot level, memory requirement, and state transitions. Simulation results show that the average prediction error for the optimal number of cores in NN benchmarks is within 2.3%. We further carry out extensive simulations which demonstrate that FCNN training time can be reduced by 22.28% and 4.91% on average using our proposed scheme, compared with traditional parallel computing methods that either allocate a fixed number of cores or allocate as many cores as possible, respectively. Compared with ENoC, simulation results show that under batch sizes of 64 and 128, on average ONoC can achieve 21.02% and 12.95% on reducing training time with 47.85% and 39.27% on saving energy, respectively.

Graph neural networks are emerging as promising methods for modeling molecular graphs, in which nodes and edges correspond to atoms and chemical bonds, respectively. Recent studies show that when 3D molecular geometries, such as bond lengths and angles, are available, molecular property prediction tasks can be made more accurate. However, computing of 3D molecular geometries requires quantum calculations that are computationally prohibitive. For example, accurate calculation of 3D geometries of a small molecule requires hours of computing time using density functional theory (DFT). Here, we propose to predict the ground-state 3D geometries from molecular graphs using machine learning methods. To make this feasible, we develop a benchmark, known as Molecule3D, that includes a dataset with precise ground-state geometries of approximately 4 million molecules derived from DFT. We also provide a set of software tools for data processing, splitting, training, and evaluation, etc. Specifically, we propose to assess the error and validity of predicted geometries using four metrics. We implement two baseline methods that either predict the pairwise distance between atoms or atom coordinates in 3D space. Experimental results show that, compared with generating 3D geometries with RDKit, our method can achieve comparable prediction accuracy but with much smaller computational costs. Our Molecule3D is available as a module of the MoleculeX software library (https://github.com/divelab/MoleculeX).

The crucial component making smart technologies possible – from something as small as a ring to as large as an entire city – is the IoT. Although there are varying definitions, the term IoT is mainly used for previously'dumb' devices that didn't have an Internet connection, but that now communicate with the network independently of human action. For this reason, a smartphone isn't explicitly defined as an IoT device – although it's crammed with sensors. A connected refrigerator or microwave oven however is. Nowadays, these smart technology devices devices include billions of objects of all shapes and sizes – coffee machines, lightbulbs, driver-less trucks, wearable fitness devices, jet engines and children's smart toys – all equipped with sensors and communicating data through the Internet.

Lithium-ion (Li-ion) batteries commonly used in electric vehicles, small appliances and electronic storage systems are rechargeable and energy-efficient. As the demand for Li-ion batteries escalates, the elements needed to create them, such as cobalt, nickel and lithium, are in short supply. Jodie Lutkenhaus, professor in the Texas A&M University Artie McFerrin Department of Chemical Engineering, and Daniel Tabor, assistant professor in the Department of Chemistry, are using machine learning techniques to optimize polymers needed for developing metal-free, recyclable, organic batteries. The research is funded by the National Science Foundation (NSF) and in collaboration with Juan De Pablo and Stuart Rowan from the University of Chicago. With the approaching Li-ion battery shortage, metal-free batteries offer great potential. In theory, organic batteries could be locally sourced, decreasing demands on supply chains.

Fumbling to find flashlights during blackouts may soon be a distant memory, as quantum computing and artificial intelligence could learn to decipher an electric grid's problematic quirks and solve system hiccups so fast, humans may not notice. Rather than energy grid faults turning into giant problems – such as voltage variations or widespread blackouts – blazing fast computation blended with artificial intelligence could rapidly diagnose trouble and find solutions in tiny splits of seconds, according to Cornell research forthcoming in Applied Energy (Dec. 1, 2021). "Energy power system failures are an old problem and we are still using classic computational methods to resolve them," said Fengqi You, the Roxanne E. and Michael J. Zak Professor in Energy Systems Engineering in the College of Engineering. "Today's power systems can benefit from AI and the computational power of quantum computing, so power systems can be stable and reliable." You, along with doctoral student Akshay Ajagekar, are co-authors of "Quantum Computing-based Hybrid Deep Learning for Fault Diagnosis in Electrical Power Systems."

By constructing digital twins (DT) of an integrated energy system (IES), one can benefit from DT's predictive capabilities to improve coordinations among various energy converters, hence enhancing energy efficiency, cost savings and carbon emission reduction. This paper is motivated by the fact that practical IESs suffer from multiple uncertainty sources, and complicated surrounding environment. To address this problem, a novel DT-based day-ahead scheduling method is proposed. The physical IES is modelled as a multi-vector energy system in its virtual space that interacts with the physical IES to manipulate its operations. A deep neural network is trained to make statistical cost-saving scheduling by learning from both historical forecasting errors and day-ahead forecasts. Case studies of IESs show that the proposed DT-based method is able to reduce the operating cost of IES by 63.5%, comparing to the existing forecast-based scheduling methods. It is also found that both electric vehicles and thermal energy storages play proactive roles in the proposed method, highlighting their importance in future energy system integration and decarbonisation.

One of the key challenges to deep reinforcement learning (deep RL) is to ensure safety at both training and testing phases. In this work, we propose a novel technique of unsupervised action planning to improve the safety of on-policy reinforcement learning algorithms, such as trust region policy optimization (TRPO) or proximal policy optimization (PPO). We design our safety-aware reinforcement learning by storing all the history of "recovery" actions that rescue the agent from dangerous situations into a separate "safety" buffer and finding the best recovery action when the agent encounters similar states. Because this functionality requires the algorithm to query similar states, we implement the proposed safety mechanism using an unsupervised learning algorithm, k-means clustering. We evaluate the proposed algorithm on six robotic control tasks that cover navigation and manipulation. Our results show that the proposed safety RL algorithm can achieve higher rewards compared with multiple baselines in both discrete and continuous control problems. The supplemental video can be found at: https://youtu.be/AFTeWSohILo.

The paper presents numerical experiments and some theoretical developments in prediction with expert advice (PEA). One experiment deals with predicting electricity consumption depending on temperature and uses real data. As the pattern of dependence can change with season and time of the day, the domain naturally admits PEA formulation with experts having different ``areas of expertise''. We consider the case where several competing methods produce online predictions in the form of probability distribution functions. The dissimilarity between a probability forecast and an outcome is measured by a loss function (scoring rule). A popular example of scoring rule for continuous outcomes is Continuous Ranked Probability Score (CRPS). In this paper the problem of combining probabilistic forecasts is considered in the PEA framework. We show that CRPS is a mixable loss function and then the time-independent upper bound for the regret of the Vovk aggregating algorithm using CRPS as a loss function can be obtained. Also, we incorporate a ``smooth'' version of the method of specialized experts in this scheme which allows us to combine the probabilistic predictions of the specialized experts with overlapping domains of their competence.

Switching dynamical systems provide a powerful, interpretable modeling framework for inference in time-series data in, e.g., the natural sciences or engineering applications. Since many areas, such as biology or discrete-event systems, are naturally described in continuous time, we present a model based on an Markov jump process modulating a subordinated diffusion process. We provide the exact evolution equations for the prior and posterior marginal densities, the direct solutions of which are however computationally intractable. Therefore, we develop a new continuous-time variational inference algorithm, combining a Gaussian process approximation on the diffusion level with posterior inference for Markov jump processes. By minimizing the path-wise Kullback-Leibler divergence we obtain (i) Bayesian latent state estimates for arbitrary points on the real axis and (ii) point estimates of unknown system parameters, utilizing variational expectation maximization. We extensively evaluate our algorithm under the model assumption and for real-world examples.