Collaborating Authors

Energy Conservation & Efficiency

Chinese tech giant unveils a smart home speaker to compete with Amazon


Chinese internet search giant Baidu on Thursday introduced a speaker and two robots as part of its Raven series in a serious push into the highly competitive smart home market.

MOL Utilizes A.I. To Estimate Vessel Speed And Fuel Consumption


Mitsui O.S.K. Lines, Ltd. (MOL) today announced that it teamed up with Fujitsu Laboratories Ltd., and Tokyo University of Marine Science and Technology to verify the accuracy of technology to estimate vessel performance at sea by applying Fujitsu's artificial intelligence (AI) technology, "FUJITSU Human Centric AI Zinrai." This project is a part of MOL's initiative to assess the effectiveness of AI technology, and aims to reduce fuel consumption and vessels' environmental impact by verifying the accuracy of the technology, using Fujitsu's AI Technology to estimate vessel performance at sea. MOL provided actual voyage data collected from MOL fleet in operation to Fujitsu Laboratories, which, along with Tokyo University of Marine Science and Technology, verified the data by using their jointly developed machine learning method. Learned the correlation of each item of operation data using Fujitsu's unique AI technology and high-dimensional statistics analysis technology, and established the technology that estimates vessel performance. Estimated the ship speed from the data other than the speed and verified the comparison between that estimated value and actual operation data, in case to assess allowance of speed.

A unified decision making framework for supply and demand management in microgrid networks Artificial Intelligence

This paper considers two important problems - on the supply-side and demand-side respectively and studies both in a unified framework. On the supply side, we study the problem of energy sharing among microgrids with the goal of maximizing profit obtained from selling power while meeting customer demand. On the other hand, under shortage of power, this problem becomes one of deciding the amount of power to be bought with dynamically varying prices. On the demand side, we consider the problem of optimally scheduling the time-adjustable demand - i.e., of loads with flexible time windows in which they can be scheduled. While previous works have treated these two problems in isolation, we combine these problems together and provide for the first time in the literature, a unified Markov decision process (MDP) framework for these problems. We then apply the Q-learning algorithm, a popular model-free reinforcement learning technique, to obtain the optimal policy. Through simulations, we show that our model outperforms the traditional power sharing models.

Government announces £84m robotics and smart energy innovation funding


A total of £84m in government funding to support robotics technology research and the development of smart energy systems has been announced by the government as part of a package of fresh support for...

Energy Data Insights: The Missing "Smart Step" to Better Building Performance


A large and necessary step in achieving the Paris Agreement requires a transition to a highly efficient building stock in terms of real energy performance. This is perhaps nowhere more true than in Europe, where it is often stated that "all buildings built before 1990 are inefficient" and that up to 75% need renovating in order to reach a higher energy efficiency standard. In order to decarbonise EU building stock by 2050, a vision laid out in the Clean Energy for All Europeans communication (2016), the majority of buildings must be highly energy efficient, meaning they should comply with an Energy Performance Certificate (EPC) label A. Unfortunately, this might prove more difficult than expected. New research from the BPIE shows that although building performance is constantly improving in the EU, only after 2010 was the average building was built to an efficient standard (0.49 W/m2 K for the building envelope) in the European Union. That means only 3% of building stock in the EU does actually qualifies for the A-label, so 97% (not 75% as typically stated) should therefore be upgraded.

Kakao AI speaker begins official sales


Kakao has begun official sales of its Kakao Mini AI speaker, the company said. The speaker goes up for sale on Kakao's gift market, available on its chat app KakaoTalk. The speaker costs 119,000 won ($107), but subscribers to Kakao's music streaming service Melon can get it for 49,000 won ($44). Kakao said Kakao Mini will understand the context to answer queries and will sync with Melon's database to suggest music. The compnay will add other services to the speaker at a later date, including translations, ordering food, and calling cabs -- all of which are all services Kakao provides on mobile.

The Eufy Genie is here to grant all of your smart house-related wishes — and it's 58% off


Just to let you know, if you buy something featured here, Mashable might earn an affiliate commission. The Eufy Genie smart speaker is basically an Amazon Echo Dot smart clone, wrote Mashable's Ray Wong when the device first debuted. Amazon is currently selling the smart home gadget for 58% off. Eufy works as a smart speaker -- just ask Alexa to play music for you and services including Spotify, Amazon Music, and Pandora are at your fingertips -- but Eufy can do much more with Alexa. You can hook Eufy up to any compatible smart device including lights, coffee machines, light bulbs, and even smart vacuums so that they can all activate at your leisure.

Bayesian Compression for Deep Learning Machine Learning

Compression and computational efficiency in deep learning have become a problem of great significance. In this work, we argue that the most principled and effective way to attack this problem is by adopting a Bayesian point of view, where through sparsity inducing priors we prune large parts of the network. We introduce two novelties in this paper: 1) we use hierarchical priors to prune nodes instead of individual weights, and 2) we use the posterior uncertainties to determine the optimal fixed point precision to encode the weights. Both factors significantly contribute to achieving the state of the art in terms of compression rates, while still staying competitive with methods designed to optimize for speed or energy efficiency.

Artificial Intelligence and the future of energy – WePower – Medium


With the rise of cloud computing and the ever-decreasing costs associated with computations, now and in the future this technology will be more and more widely available. One of the most process heavy steps in AI systems is model training and validation. Being able to pay per minute or even second for the use of computing power removes the need for large upfront investment and data centre maintenance costs. With Google Cloud, IBM Bluemix and Amazon Cloud the power to perform highly complex computations is readily available for everyone today [11]. The systems architecture for machine learning which underpins artificial intelligence is also seamlessly provided by cloud solutions.