When computers speak, how human should they sound? This was a question that a team of six IBM linguists, engineers and marketers faced in 2009, when they began designing a function that turned text into speech for Watson, the company's "Jeopardy!"-playing Eighteen months later, a carefully crafted voice -- sounding not quite human but also not quite like HAL 9000 from the movie "2001: A Space Odyssey" -- expressed Watson's synthetic character in a highly publicized match in which the program defeated two of the best human "Jeopardy!" The challenge of creating a computer "personality" is now one that a growing number of software designers are grappling with as computers become portable and users with busy hands and eyes increasingly use voice interaction. Machines are listening, understanding and speaking, and not just computers and smartphones.

Amid the protests against police brutality spreading across the country, Dallas is still haunted by the death of five officers at the hands of a lone gunman. However, some were concerned by the manner in which the police ended a standoff with 25-year-old Micah X Johnson, the suspect behind the shooting. The standoff ended early Friday morning when the police detonated a bomb robot, killing the suspect. The Dallas Police Department defended its action saying that the bomb was "a last resort." In a statement published Saturday, the Dallas Police Department said, "When all attempts to negotiate with the suspect, Micah Johnson, failed under the exchange of gunfire, the Department utilized the mechanical tactical robot, as a last resort, to deliver an explosion device to save the lives of officers and citizens." Police used the Remotec, Model F-5 with claw and arm extension and an explosive device of C4 with a "Det" cord.

Convolutional Neural Networks (CNN) are biologically-inspired variants of MLPs. From Hubel and Wiesel's early work on the cat's visual cortex [Hubel68], we know the visual cortex contains a complex arrangement of cells. These cells are sensitive to small sub-regions of the visual field, called a receptive field. The sub-regions are tiled to cover the entire visual field. These cells act as local filters over the input space and are well-suited to exploit the strong spatially local correlation present in natural images.

First, there were mom-and-pop stores and boutiques. Then came malls and, with them, big retail chains. After that, we got the online experience -- where with just a click we can access thousands of products, order them, try them on and send them back if they don't suit us. Americans do over 90 percent of their shopping in physical stores, and the launch of physical stores by online companies, including Amazon and Warby Partner, is one of the latest trends in retail, according to New York data company CB Insights. So startups worldwide are also striving to bring new technologies to stores, to help improve the buying experience and sales.

Below is a series of questions that I have been asked frequently and preliminary answers. The facts remain incomplete, so these are preliminary thoughts. In the wake of the shooting of the Dallas police officers Thursday night during a peaceful protest, police cornered the shooter -- Micah Xavier Johnson -- in a parking garage. After an hours-long standoff that included exchanges of gunfire, they used a robot to deliver an explosive that killed the gunman. "We saw no other option but to use our bomb robot and place a device on its extension for it to detonate where the subject was," Dallas Police Chief David Brown said at a news conference Friday morning.

SINGAPORE: At a time when radicalism is on the rise and people are concerned that their city might join the list of those tainted by terror, safety and security are topmost on the minds of government customers, Hewlett Packard Enterprise told Channel NewsAsia. The US-based IT giant is thus hoping to tackle these issues with a tool to help governments better make sense of the huge amounts of information being shared online by citizens, in the form of a Citizen Insights Dashboard. The tool, which is currently in beta, will be showcased for the first time at the World Cities Summit (WCS) held in Singapore from Sunday (Jul 10). It was developed in the city-state since April by a team led by Mr Jason Tan, director of HSE Insights Lab Singapore spearheading the project. More than just surveillance for security, the dashboard looks to make sense of posts made by citizens on social media to understand their thoughts and what they care about.

Intelligence, defined as the ability to acquire knowledge and skills. Intelligence for the longest time possible is associated with the human brain. Artificial intelligence is basically defined as intelligence that is originating from machines. Most computer applications only make existing processes and functions faster and maybe more efficiently but cannot create new duties altogether. However, artificial intelligence has already challenged this notion.

"Really, we're building collision avoidance for industrial drones," said Alexander Harmsen, CEO and co-founder of Iris Automation. "We see this huge need for industrial drones for mining exploration, pipeline inspection, agricultural surveying, forestry, or even package delivery." Without a way to avoid mid-air collisions, drones risk crashing into a Cessna, a flock of geese or a 747. Worst case scenario: a drone gets sucked into a jet engine causing catastrophic engine failure as high-velocity bits of metal penetrate fuel tanks, hydraulic lines and the cabin. Iris Automation's solution is an AI computer that blends real-time images and 3D maps to track incoming objects.

Intelligence, defined as the ability to acquire knowledge and skills. Intelligence for the longest time possible is associated with the human brain. Artificial intelligence is basically defined as intelligence that is originating from machines. Most computer applications only make existing processes and functions faster and maybe more efficiently but cannot create new duties altogether. However, artificial intelligence has already challenged this notion.

Word embeddings have been shown to be useful across state-of-the-art systems in many natural language processing tasks, ranging from question answering systems to dependency parsing. (Herbelot and Vecchi, 2015) explored word embeddings and their utility for modeling language semantics. In particular, they presented an approach to automatically map a standard distributional semantic space onto a set-theoretic model using partial least squares regression. We show in this paper that a simple baseline achieves a +51% relative improvement compared to their model on one of the two datasets they used, and yields competitive results on the second dataset.