Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead Rudin et al., arXiv 2019 It's pretty clear from the title alone what Cynthia Rudin would like us to do! The paper is a mix of technical and philosophical arguments and comes with two main takeaways for me: firstly, a sharpening of my understanding of the difference between explainability and interpretability, and why the former may be problematic; and secondly some great pointers to techniques for creating truly interpretable models. A model can be a black box for one of two reasons: (a) the function that the model computes is far too complicated for any human to comprehend, or (b) the model may in actual fact be simple, but its details are proprietary and not available for inspection. In explainable ML we make predictions using a complicated black box model (e.g., a DNN), and use a second (posthoc) model created to explain what the first model is doing. A classic example here is LIME, which explores a local area of a complex model to uncover decision boundaries.

Embedded vision solutions will be a key enabler for making automobiles fully autonomous. Giving an automobile a set of eyes – in the form of multiple cameras and image sensors – is a first step, but it also will be critical for the automobile to interpret content from those images and react accordingly. To accomplish this, embedded vision processors must be hardware optimized for performance while achieving low power and small area, have tools to program the hardware efficiently, and have algorithms to run on these processors. The significant automotive safety improvements in the past (e.g., shatter-resistant glass, three-point seatbelts, airbags), were passive safety measures designed to minimize damage during an accident. We now have technology that can actively help the driver avoid crashing in the first place.

In a pub in London, England, complicated technology is taking on a simple problem: waiting for a pint in lines that can sometimes be unruly. It uses facial recognition software to form a digital queue and prevent people from cutting in line. A large TV screen is mounted above the bar with a live video feed showing the people waiting for a drink. Beside the image of each customer, a number pops up to indicate where they are in the line. "We just want to make the experience more frictionless and fair," said John Wyllie, managing director of DataSparQ, the company behind the technology called A.I. Bar.

Imagine flying from Europe to Australia in just 90 minutes. This is fantasy for now, but the stratosphere is the next frontier in aviation, with supersonic flights using that high-altitude space. And one of the keys to making it happen will be the use of Artificial Intelligence (AI) to cope with the increased complexity the sector will face. "Aviation is being reshaped by a number of powerful forces that are fundamentally impacting the Air Traffic Management sector," says Beatrice Pesquet-Popescu, Research and Business Innovation Director for Air Traffic Management (ATM) at Thales. "In addition to the growth expected in traditional aircraft, we will have to cope with new vehicles such as drones and stratospheric balloons, circulating in low or high altitude airspace."

Bail is the most common area of AI integration in the U.S., and is used in jurisdictions housing more than a quarter of the U.S. population, said Thomas as part of the Data & Design Symposium put on by The Action Group on Access to Justice. The technology -- used to measure risk factors such as re-offending or failure to appear -- has been controversial. "Criminal justice data has got generations of historic bias -- systemic bias -- against African Americans and other communities in the U.S.," said Thomas. Some people argue that such technologies should not be used at all, said Thomas, because "if the data is inherently discriminatory it means the outcome is inevitably discriminatory." However, the reality is not so simple, Thomas said.

The US Navy has a new method for transporting supplies to off-shore submarines – drone delivery. The military organization has successful delivered a five-pound payload consisting of circuit cards, medical supplies and food to the Virginia-class fast-attack submarine USS Hawaii (SSN 776) while at sea. Delivering supplies by drone will eliminate the need for submarines to pull into ports for goods and allow them to spend more time in the fight. This is the first time the US Navy has employed the use of a drone to deliver goods and Lt. Cmdr. Christopher Keithley, assigned to COMSUBPAC said'What started as an innovative idea has come to fruition as a potentially radical new submarine logistics delivery capability.' 'A large percentage of parts that are needed on submarines weigh less than five-pounds, so this capability could alleviate the need for boats to pull into ports for parts or medical supplies.'

Infantry Soldiers with 1st Battalion, 8th Infantry Regiment, 3rd Armored Brigade Combat Team, 4th Infantry Division, fire an FGM-148 Javelin during a combined arms live fire exercise in Jordan on August 27, 2019, in support of Eager Lion - file photo. A small group of maneuvering infantry soldiers will soon be able to target and destroy enemy tanks at night from distances up to 4.5 kilometers (2.8 miles) -- by firing portable, man-carried Javelin Anti-Tank Missiles engineered with a new generation of targeting optics. The U.S. Army and Raytheon plan to enter production of a new Lightweight Command Launch Unit for the Javelin designed to bring a new level of "precision lethality to an infantry squad." The new Lightweight CLU unit enables much greater standoff distance for infantry attacking tanks by doubling the attack range from 2.5km to 4.5km, developers said. "You have to be able to speed up the kill chain and detect the adversary before he can detect you. You want to get a launch shot off before he knows you are there. It all starts with sensing," Tommy Boccardi, Javelin Domestic Business Development, Raytheon, told Warrior.

Mr. Robot creator Sam Esmail says portraying Elliot (Rami Malek) in a hooded sweatshirt was a deliberate choice: "That hoodie made us closer to who Elliot was." Mr. Robot creator Sam Esmail says portraying Elliot (Rami Malek) in a hooded sweatshirt was a deliberate choice: "That hoodie made us closer to who Elliot was." Editor's note: This interview contains a racial slur. Sam Esmail, the creator, lead writer and director of the TV series Mr. Robot, has always identified with computer programming and hacker culture -- in part because of his experiences with social anxiety. In college, he shied away from parties and instead took refuge in the computer lab. It felt safer to talk to people online than in person, Esmail says.

Less than 10% … that's how many artificial-intelligence test projects are estimated to be deployed into full-scale production in enterprise environments, according to a recent report from the International Institute for Analytics. There are a number of reasons for this surprisingly small amount, including an overwhelming amount of data and the lack of easy-to-use tools to analyze it. It's a problem that calls for operationalizing AI and machine learning, making it accessible and repeatable consistently. "Ultimately, if you want to get business value from those models and all of the hard work that you've done, it has to be injected into the business process," said Anant Chintamaneni (pictured), vice president and general manager of BlueData at Hewlett Packard Enterprise Co. "Operationalization of machine learning is ultimately the key, and that's the progression that enterprises have to make." Burris was joined for a digital community event by co-host Stu Miniman (@stu), and they also interviewed Nanda Vijaydev, distinguished technologist and lead data scientist at HPE; Patrick Osborne, vice president and general manager of big data, analytics, and scale-out data platforms at HPE; and Wikibon analyst James Kobielus (@jameskobielus).

For the last few years, internet security has been based on a combination of anti-virus software, isolation techniques and encryption software. Government bodies and security companies would track traffic on the internet and look for suspicious materials based upon their signature. These techniques focused on running anti-malware software after the facts. They enabled the segregation between good data and malware. But if malware was undetected, it could lurk in the background of systems for months or even years and become active later in time. The consumer world is rapidly changing.