The United States Department of Energy (DOE) has given three institutions a $5.5 million grant to collectively find solutions to some of the most challenging problems in artificial intelligence (AI) today. Scientists from Georgia Tech, Pacific Northwest National Laboratory, and Sandia National Laboratory will collaborate to develop technologies that are core to the DOE's priorities including cybersecurity, graph analytics, and electric grid resilience. Tushar Krishna, an assistant professor in Georgia Tech's School of Electrical and Computer Engineering and Machine Learning Center (ML@GT), will serve as a deputy director of the newly established Center for Artificial Intelligence-focused Architectures and Algorithms (ARIAA). Georgia Tech is contributing expertise in modeling and developing custom accelerators for machine learning and sparse linear algebra. The institute will also provide access to its advanced computing resources.

OnviSource announced today it has started the deployment of its AI-driven solutions powered by its new proprietary Artificial Intelligence software, called iMachine . Company's solutions are able to utilize the most optimized AI engine pertinent to their specific application. For example, Company's Intelligent Virtual Agent or smart bot, called Liaa, primarily utilizes iMachine's NLP/NLU engine; while Intellecta multichannel analytics and Automata RPA products may use iMachine's ML and DL engines for a variety of their AI-driven features. Use of iMachine by Company's solutions in analytics, RPA and IVA significantly enhances their capabilities in effectively addressing today's enterprise and contact center challenges in workforce optimization, customer experience management and business process automation; as well as automating the management of enterprise contents. Content of calls, audio files, email, chat, text, and structured or unstructured documents can be analyzed by iMachine for discovering intent, purpose, compliance, categories, sentiment, root causes and complex information otherwise undetected by analytics that do not use AI engines.

AI can be harnessed in a wide range of economic sectors and situations to contribute to managing environmental impacts and climate change.Some examples of application include: AI-infused clean distributed energy grids, precision agriculture, sustainable supply chains, environmental monitoring and enforcement, and enhanced weather and disaster prediction and response. Research by PwC UK, commissioned by Microsoft, models the economic impact of AI's application to manage the environment, across four sectors – agriculture, water, energy and transport. It estimates that using AI for environmental applications could contribute up to $5.2 trillion USD to the global economy in 2030, a 4.4% increase relative to business as usual. In parallel the application of AI levers could reduce worldwide greenhouse gas (GHG) emissions by 4% in 2030, an amount equivalent to 2.4 Gt CO2e – equivalent to the 2030 annual emissions of Australia, Canada and Japan combined. At the same time as productivity improvements, AI could create 38.2 million net new jobs across the global economy offering more skilled occupations as part of this transition.

-- Deep reinforcement learning has great potential to acquire complex, adaptive behaviors for autonomous agents automatically. However, the underlying neural network polices have not been widely deployed in real-world applications, especially in these safety-critical tasks (e.g., autonomous driving). One of the reasons is that the learned policy cannot perform flexible and resilient behaviors as traditional methods to adapt to diverse environments. In this paper, we consider the problem that a mobile robot learns adaptive and resilient behaviors for navigating in unseen uncertain environments while avoiding collisions. We present a novel approach for uncertainty-aware navigation by introducing an uncertainty-aware predictor to model the environmental uncertainty, and we propose a novel uncertainty-aware navigation network to learn resilient behaviors in the prior unknown environments. T o train the proposed uncertainty-aware network more stably and efficiently, we present the temperature decay training paradigm, which balances exploration and exploitation during the training process. Our experimental evaluation demonstrates that our approach can learn resilient behaviors in diverse environments and generate adaptive trajectories according to environmental uncertainties. Videos of the experiments are available at https://sites.google.com/view/resilient-nav/ . With the recent progress of machine learning techniques, deep reinforcement learning has been seen as a promising technique for autonomous systems to learn intelligent and complex behaviors in manipulation and motion planning tasks [1]-[3].

Every year, municipalities across South Africa lose millions of Rands from electricity theft. My work as an electrical engineer at Aurecon has led me to think deeply about coming up with ways to not only help solve this problem but consider possible preventative measures that could be put into place. Municipalities generate an enormous amount of data related to electricity distribution and consumption. When combined with real-time data analysis and machine learning algorithms, this information can be used to pick up on electricity theft at any node in the grid. As part of my Research interests, I started to create an algorithm that uses machine learning and artificial neural intelligence to detect electricity theft as well as cable theft, together with one of the Junior Electrical Engineers Tendai Matiza.

Today, large commercial and industrial (C&I) companies are setting up their own renewable energy assets, including storage and EV charging, and investing in energy efficiency in a bid to reduce their carbon footprint. As a result, oil and gas companies, and utilities and energy solution providers who have been serving them are now optimising themselves to suit these demands by setting up and managing renewable energy resources. But, as energy generation gets decentralised, demand reduction, shifting, and load management have become more complex, especially due to the intermittent availability of green energy. Companies have to look for greener solutions with intermittent supply, managing grid reliance, and integrating storage and EV charging in their energy mix. It also means delivering value-added services to these C&I customers to reduce peak loads, forecast generation, analyse plant efficiency, maintain a demand-supply balance, and automate the flow of energy.

The high model complexity of deep learning algorithms enables remarkable learning capacity in many application domains. However, a large number of trainable parameters comes with a high cost. For example, during both the training and inference phases, the numerous trainable parameters consume a large amount of resources, such as CPU/GPU cores, memory and electric power. In addition, from a theoretical statistical learning perspective, the high complexity of the network can result in a high variance in its generalization performance. One way to reduce the complexity of a network without sacrificing its accuracy is to define and identify redundancies in order to remove them. In this work, we propose a method to observe and analyze redundancies in the weights of a 2D convolutional (Conv2D) network. Based on the proposed analysis, we construct a new layer called Separable Convolutional Eigen-Filters (SCEF) as an alternative parameterization to Conv2D layers. A SCEF layer can be easily implemented using depthwise separable convolution, which are known to be computationally effective. To verify our hypothesis, experiments are conducted on the CIFAR-10 and ImageNet datasets by replacing the Conv2D layers with SCEF and the results have shown an increased accuracy using about 2/3 of the original parameters and reduce the number of FLOPs to 2/3 of the original net. Implementation-wise, our method is highly modular, easy to use, fast to process and does not require any additional dependencies.

Deep reinforcement learning (DRL) is a booming area of artificial intelligence. Many practical applications of DRL naturally involve more than one collaborative learners, making it important to study DRL in a multi-agent context. Previous research showed that effective learning in complex multi-agent systems demands for highly coordinated environment exploration among all the participating agents. Many researchers attempted to cope with this challenge through learning centralized value functions. However, the common strategy for every agent to learn their local policies directly often fail to nurture strong inter-agent collaboration and can be sample inefficient whenever agents alter their communication channels. To address these issues, we propose a new framework known as centralized training and exploration with decentralized execution via policy distillation. Guided by this framework and the maximum-entropy learning technique, we will first train agents' policies with shared global component to foster coordinated and effective learning. Locally executable policies will be derived subsequently from the trained global policies via policy distillation. Experiments show that our new framework and algorithm can achieve significantly better performance and higher sample efficiency than a cutting-edge baseline on several multi-agent DRL benchmarks.

As you leave for work, your car accesses your phone calendar to determine the destination you're headed for and already knows the shortest and fastest route to take. In case you encounter heavy traffic, the car automatically notifies your office that you are running late! While this might seem like a clip from a futuristic movie, scenarios like these are already taking shape. Enter the world of Internet of Things, popularly referred to as IoT. So far the internet had mostly connected people to information, people to people, and people to business.

We propose a novel learning framework for recommendation systems, assisting collaborative filtering with a synthetic feedback loop. The proposed framework consists of a "recommender" and a "virtual user." The recommender is formulizd as a collaborative-filtering method, recommending items according to observed user behavior. The virtual user estimates rewards from the recommended items and generates the influence of the rewards on observed user behavior. The recommender connected with the virtual user constructs a closed loop, that recommends users with items and imitates the unobserved feedback of the users to the recommended items. The synthetic feedback is used to augment observed user behavior and improve recommendation results. Such a model can be interpreted as the inverse reinforcement learning, which can be learned effectively via rollout (simulation). Experimental results show that the proposed framework is able to boost the performance of existing collaborative filtering methods on multiple datasets.