TOKUSHIMA – The Tokushima Prefectural Government and e-books distributor Media Do Holdings Co. announced Tuesday they will launch an experiment to use AI to summarize the governor's news conferences in a bid to make more administrative documents available online more quickly. Under the experiment, which they plan to start Monday, they will use a speech recognition system to catch remarks at news conferences given by the governor, and the complete transcription will be posted on the prefectural government's website after being checked by prefecture officials. Users can choose the amount of words they want to read from 10 percent to 90 percent, and AI summarizes the documents in accordance with the users' requests. The new system will reduce the time needed to transcribe news conferences from roughly 10 hours to two hours, and will enable the prefecture to post documents about four hours after the news conferences are held, according to the officials. "Government agencies create a lot of documents but most of them are difficult to read. I hope summarized documents will help people familiarize themselves with the government," said Tokushima Gov. Kamon Iizumi, adding that the move will also help reduce prefectural officials' working hours.

Due to the lack of enough generalization in the state-space, common methods in Reinforcement Learning (RL) suffer from slow learning speed especially in the early learning trials. This paper introduces a model-based method in discrete state-spaces for increasing learning speed in terms of required experience (but not required computational time) by exploiting generalization in the experiences of the subspaces. A subspace is formed by choosing a subset of features in the original state representation (full-space). Generalization and faster learning in a subspace are due to many-to-one mapping of experiences from the full-space to each state in the subspace. Nevertheless, due to inherent perceptual aliasing in the subspaces, the policy suggested by each subspace does not generally converge to the optimal policy. Our approach, called Model Based Learning with Subspaces (MoBLeS), calculates confidence intervals of the estimated Q-values in the full-space and in the subspaces. These confidence intervals are used in the decision making, such that the agent benefits the most from the possible generalization while avoiding from detriment of the perceptual aliasing in the subspaces. Convergence of MoBLeS to the optimal policy is theoretically investigated. Additionally, we show through several experiments that MoBLeS improves the learning speed in the early trials.

Conceptual spaces are geometric representations of conceptual knowledge, in which entities correspond to points, natural properties correspond to convex regions, and the dimensions of the space correspond to salient features. While conceptual spaces enable elegant models of various cognitive phenomena, the lack of automated methods for constructing such representations have so far limited their application in artificial intelligence. To address this issue, we propose a method which learns a vector-space embedding of entities from Wikipedia and constrains this embedding such that entities of the same semantic type are located in some lower-dimensional subspace. We experimentally demonstrate the usefulness of these subspaces as (approximate) conceptual space representations by showing, among others, that important features can be modelled as directions and that natural properties tend to correspond to convex regions.

Michel Fornasier, one of the presenters of the Cybathlon, uses his bionic hand prosthesis to demonstrate one of the Cybathlon disciplines. In the movie Star Wars: The Empire Strikes Back, Luke Skywalker is given a mechanical hand that moves and perform functions as well as his real hand. Konrad Kording, an avid Star Wars fan, has no doubt that advances in brain-machine interfaces (BMIs) will make this bit of science fiction a reality; he just doesn't know when. "We have applications for one channel and a few channels," says Kording, a neuroscientist and professor of Physical Medicine and Rehabilitation, Physiology, and Biomedical Engineering at Northwestern University in Evanston, IL. "The question is, what are the BMI applications with hundreds of thousands of channels, and no one knows that at the moment." The channels he's referring to are electrical wires or optical connectors that can be attached to the brain and can be controlled and measured.

Cambit pieces can be assembled to create a dozen different imaging systems. The cameras in our phones and tablets have turned us all into avid photographers, regularly using them to capture special moments and document our lives. One notable feature of camera phones is they are compact and fully automatic, enabling us to point and shoot without having to adjust any settings. However, when we need to capture photos of high aesthetic quality, we resort to more sophisticated DSLR cameras in which a variety of lenses and flashes can be used interchangeably. This flexibility is important for spanning the entire range of real-world imaging scenarios, while enabling us to be more creative. Many developers have sought to make these cameras even more flexible through both hardware and software.

Google is training AI to identify human behavior, using clips from movies. A link has been sent to your friend's email address. A link has been posted to your Facebook feed. Google is training AI to identify human behavior, using clips from movies.

Special effects that let Hollywood filmmakers remove unwanted objects and people from their final shots could soon be available at the push of a button. Adobe has unveiled a peak of its Cloak project, which enables automatic editing of footage that would take professionals hours to complete. Users of the tool, which is currently under development, can select any item for deletion and smart software fills in the gap. The company has not yet given any word on if or when Cloak will be built into its software. Special effects that let Hollywood filmmakers remove unwanted objects and people from their final shots could soon be available at the push of a button.

We are proud to announce that our film, "Road to A.I.," has been selected as an official entry in The Block Film & Art Festival and the 2017 Alexandria Film Festival. The "Road to A.I." documentary film asks some of the most crucial questions facing the transportation and technology industries as we shift to a more autonomous future. How is artificial intelligence being programmed to make decisions we can live with? Can sharing data make self-driving cars safer? Once seen as purely science fiction, self-driving cars are becoming more than a vision for the future; they are becoming a reality.

The convergence of broadcast television and internet-delivered content services has become a reality -- one that soon will be cemented by the media industry's embrace of ATSC 3.0. In this cross-media environment, broadcast stations face many challenges, a mix of old and new. While ratings are a long-familiar concern for broadcasters, fierce competition for audience share today comes from a variety of providers on platforms ranging from linear channels to over-the-top (OTT) services to social and digital media. The new complexity of content distribution and consumption makes it hard for executives at a station to understand that station's place -- and the value of its media inventory -- within the larger marketplace. Even more difficult is to be proactive in ensuring that the station leverages that inventory to its best advantage.

Amazon started streaming 4K HDR video way back in 2015, and now it has a streaming device to match in the new Fire TV. Assuming you have a 4K HDR television, the 2017 Fire TV streams crisper, more color-rich video for the same $70 price as Roku's Streaming Stick and Google's Chromecast Ultra. It's also a bit faster than the 1080p Fire TV Stick, which remains available for $40. Despite the step forward on video quality, the new Fire TV suffers from some confounding design choices that make basic navigation a hassle. It also requires you to keep a separate remote on hand for TV volume and power, even as those controls become table stakes on other streaming players.