Reuters reported Tuesday that Germany's Transport Minister Alexander Dobrindt plans legislation requiring self-driving cars be outfitted with data-recording systems in an effort to bring more accountability and to give automakers and engineers a chance to learn from the systems' mistakes. The proposal would enable autonomous car users to divert some of their attention from traffic while the systems are on as long as they are seated at the wheel but require that they maintain black boxes that track when the programs request drivers take over.

We consider the problem of predicting to which classes an image belongs, where the number of classes is large (many thousands or tens of thousands) and where each image typically belongs to multiple classes that should all be properly identified: multi-label, massively multi-class classification. In such classification problems, the best practice until now (for instance in use at Google, Inc.) has been to use a deep convolutional neural network such as the ones described in [19] or [18], culminating in a logistic regression layer with a sigmoid cross-entropy loss, with target labels encoded as high-dimensional sparse binary vectors. The use of logistic regression implies an important yet oft overlooked assumption made about the label space: the classes are considered to be statistically independent, each class being treated as an independent dimension in the label space. This is generally not the case in practice: mirroring statistical dependencies found in the real world, label spaces often have a well-defined internal structure, with some labels being more likely to cooccur than other labels. For instance, "sky" and "beach" are frequently cooccurring labels, while "crane" and "manta ray" are rarely cooccurring. The sigmoid cross-entropy loss with sparse binary targets does not allow to leverage such observations about the structure of the label space. 1 There is therefore an opportunity to exploit the internal structure of the label space for gains in training speed, precision, and recall. One simple way to achieve this is to project the labels onto a lower-dimensional manifold -an embedding space-where a distance function between embedded labels would capture useful statistical dependencies. An appropriate loss function may then allow a parametric model trained via stochastic gradient descent to benefit from the structure of the manifold during training and inference.

CityPilot has taken a key early step towards fully autonomous public transportation: The Mercedes-Benz self-driving bus program saw one of its Future Bus vehicles drive 20 km (or around 12.4 miles) in the Netherlands, on a route that connected Amsterdam's Schiphol airport with the nearby town of Haarlem. To make the trip, the bus had to stop at traffic lights, pass through tunnels, and navigate among pedestrians. This is a big win for the program, which owes its origins to the transport truck-focused Highway Pilot program debuted by Mercedes two years ago. That autonomous vehicle program didn't face the added challenges of navigating an urban environment, however, which makes the Future Bus successful test run a significant achievement. Mercedes has focused on making sure that it's well-suited to the city of the future, and accordingly put a lot of time into designing the vehicle's interior.

Providing positive and profitable customer experiences means nurturing customer relationships that unfold across time and channels. Productive customer interactions drive higher levels of engagement, increasing the Customer Lifetime Value. A more engaged, satisfied customer spends more, remains loyal, and recommends a brand to others. Today, the cutting edge of an organization's customer experience process is digitized, combining Omni Channel, analytics and automation. The core of digitally transforming customer care lies with deploying Artificial Intelligence (AI), based on Machine Learning approaches to automate routine interactions, allowing agents to enhance those interactions where humans make a real difference.

Microsoft's mini Xbox games console will go on sale on August 2nd, the firm has revealed. The Xbox One S, which comes in a colour called'robot white', is 40 per cent smaller than the original Xbox One and has abandoned the bulky power external supply seen in previous consoles. It also supports 4K Ultra HD video, but will also come with up to 2TB of internal hard disk space - and boasted it will be'the first and only console that allows you to watch Blu-ray movies and stream video in stunning 4K Ultra HD with High Dynamic Range (HDR).' The Xbox One S, which comes in a colour called'robot white', is 40 per cent smaller than the original Xbox One and has abandoned the bulky power external supply seen in previous consoles. Rumours that Sony is working on a more powerful version of the PlayStation 4 have been confirmed.

A Tesla Model S electric vehicle is shown in San Francisco, California. Tesla Motors says the Autopilot system for its Model S sedan "relieves drivers of the most tedious and potentially dangerous aspects of road travel." The second part of that promise was put in doubt by the fatal crash of a Model S earlier this year, when its Autopilot system failed to recognize a tractor-trailer turning in front of the vehicle. Tesla says the driver, Joshua Brown, also failed to notice the trailer in time to prevent a collision. In Tesla's own words, "the brake was not applied"--and the car plowed under the trailer at full speed, killing Brown.

Though deals to private artificial intelligence companies -- excluding incubator/accelerator rounds -- fell 10% in Q2'16, dollar funding reached an all-time high. That was partly thanks to 3 100M mega-rounds by companies using AI: a 154M Series A round went to China-based healthcare startup iCarbonX (with the participation of Tencent, Vcanbio), a 100M growth equity round was raised by New Jersey-based Fractal Analytics (from Khazanah Nasional Berhad), and there was a 100M Series D round raised by California-based cybersecurity unicorn Cylance (from investors including Blackstone Group, Insight Venture Partners, and Khosla Ventures). Our AI category includes companies applying AI solutions to verticals like healthcare, security, advertising, and finance as well as those developing general-purpose AI tech. Nearly 70% of the deals went to startups in the United States in Q2'16. A majority of the startups raising funds were still in their early-stages: Nearly 60% of the deals went to startups raising seed/angel and Series A rounds, while mid-stage startups (Series B and C) received 12% of the deals.

Machine learning technology is poised to move from niche data analytics applications to mainstream enterprise big data campaigns over the next two years, a recent vendor survey suggests. SoftServe, a software and application development specialist based in Austin, Texas, reports that 62 percent of the medium and large organizations it polled in April said they expect to roll out machine learning tools for business analytics by 2018. That majority said real-time data analysis was the most promising big data opportunity. The survey authors argue that artificial intelligence-based technologies like machine learning are moving beyond the "hype cycle" as enterprise look to automate analytics capabilities ranging from business intelligence to security. The goal is to demonstrate that that cyber defenses can be automated as more infrastructure is networked via an Internet of Things.)

BERLIN (Reuters) - Germany plans new legislation to require manufacturers of cars equipped with an autopilot function to install a black box to help determine responsibility in the event of an accident, transport ministry sources told Reuters on Monday. The fatal crash of a Tesla Motors Inc Model S car in its Autopilot mode has increased the pressure on industry executives and regulators to ensure that automated driving technology can be deployed safely. Under the proposal from Transport Minister Alexander Dobrindt, drivers will not have to pay attention to traffic or concentrate on steering, but must remain seated at the wheel so they can intervene in the event of an emergency. Manufacturers will also be required to install a black box that records when the autopilot system was active, when the driver drove and when the system requested that the driver take over, according to the proposals. The draft is due to be sent to other ministries for approval this summer, a transport ministry spokesman said.

Every time you Instagram a picture of the foam art on your coffee or leave a comment on a YouTube video, it contributes to the more than a billion gigabytes of new data we produce every day. As this number continues to increase, we are starting to run out of space to store the data; even the tiniest data storage element uses thousands of atoms to store one piece of information. Now for the first time, scientists have created a way to store information atom by atom to produce the smallest hard disk ever made, which could lead the way to much more efficient data storage. In traditional computers, data is expressed in one of two states – known as binary bits – which are either a 1 or a 0. In this new design, each bit consists of two positions on a surface of copper atoms, and one chlorine atom that can be slid back and forth between these two positions In traditional computers, data is expressed in one of two states – known as binary bits – which are either a 1 or a 0. In this new design, each bit consists of two positions on a surface of copper atoms, and one chlorine atom that can be slid back and forth between these two positions. If the chlorine atom is in the top position, there is a hole beneath it - this is a 1.