Electricity disaggregation identifies individual appliances from one or more aggregate data streams and has immense potential to reduce residential and commercial electrical waste. Since supervised learning methods rely on meticulously labeled training samples that are expensive to obtain, unsupervised methods show the most promise for wide-spread application. However, unsupervised learning methods previously applied to electricity disaggregation suffer from critical limitations. This paper introduces the concept of iterative appliance discovery, a novel unsupervised disaggregation method that progressively identifies the "easiest to find" or "most likely" appliances first. Once these simpler appliances have been identified, the computational complexity of the search space can be significantly reduced, enabling iterative discovery to identify more complex appliances. We test iterative appliance discovery against an existing competitive unsupervised method using two publicly available datasets. Results using different sampling rates show iterative discovery has faster runtimes and produces better accuracy. Furthermore, iterative discovery does not require prior knowledge of appliance characteristics and demonstrates unprecedented scalability to identify long, overlapped sequences that other unsupervised learning algorithms cannot.
Posted: 25 July 2018 Ami S Lakdawala (GSK's In-silico drug discovery unit), George Okafo (GSK's In-silico drug discovery unit), John Baldoni (GSK's In-silico drug discovery unit), Michael Palovich (GSK's In-silico drug discovery unit), Tobias Sikosek (GSK's In-silico drug discovery unit), Voshal Sahni (GSK's In-silico drug discovery unit) No comments yet Drug discovery has always been challenging; today, more so than ever. While there has been success in addressing many diseases, others remain intractable. There is a need and opportunity to explore new drug discovery approaches that harness immense datasets (public and private), which have been built upon the successes and failures of the past to guide in-silico approaches to new therapies. Advances in genetics and molecular biology have revealed potential new targets for developing medicines. Deciding which target to pursue is challenging and an area in which there is opportunity to increase productivity.
The whole point of virtual reality (VR) is to immerse yourself in unfamiliar environments, rendered in digital form. So it makes a lot of sense to apply the technology to virtual travel experiences. Discovery Communications disca is teaming with Google googl on a new VR show called "Discovery TRVLR" that will show viewers places and cultures around the world, all in the 360-degree format that's supported on devices such as the Google Daydream and Cardboard headsets. The first instalment in the 38-part series will launch Friday. North America, South America, Asia, Africa, Europe and Australia will each feature in six episodes, while Antarctica will get two.
When it comes to finding exoplanets in photos, hundreds or even thousands of eyes are better than an algorithm. The University of Geneva is creating a game to help researchers identify exoplanets with the help of gamers, it will be one of the largest citizen science projects ever, according to the MIT Technology Review. The game will operate within EVE Online servers under the title "Project Discovery." EVE is an online gaming platform based in space, developed by the video game company CCP. There is already one Project Discovery game that involves identifying patterns of protein to help advance the science and expansion of the Human Protein Atlas database.
A technology called machine learning is behind that seemingly magical ability of social networking websites to identify people in posted photos. By analyzing subtle patterns in facial features, machine learning algorithms can recognize people you've never tagged and might not even know. Machine learning is also transforming how scientists at Novartis discover and develop new drugs. Similar to how social networking websites use the technology to classify people on your computer screen, Novartis scientists use it to classify digital images of cells, each treated with different experimental compounds. Biological insights that might take months to generate using time-consuming laboratory experiments and human visual inspection can be revealed much faster using automated computer algorithms looking at pictures.