"Empathy won't necessarily save your job from artificial intelligence." "What one has to realise is that notions of trust and empathy are relevant to the current way of delivering professional services, which is one-to-one consultation. But if you can find different ways of producing and distributing knowledge and expertise in society, the features that were important in the advisory model, like empathy and trust, may be less important. What you're maybe looking for is a reliable outcome, rather than an empathetic service.'"

"Richard Susskind, author of The Future of Professions: How Technology Will Transform the Work of Human Experts, gives the example of the specialist medical practitioner." "If it transpires that in the future AI systems will be more reliable than human doctors and specialists, then a fairly simple choice might need to be made: do we go for a far more reliable online diagnosis, or do we go to a demonstrably less reliable service which is more empathetic?"

"The good news for those betting on AI is that graphics chips have so far managed to defy gravity. At the recent conference of leading graphics chipmaker Nvidia, CEO Jensen Huang displayed a chart showing how his chips' performance has continued to accelerate exponentially while growth in the performance of general purpose processors, or CPUs, has slowed."

"Doug Burger, a distinguished engineer at Microsoft's NExT division that works on commercializing new technology, says a similar gap is opening between conventional and machine-learning software. 'You're starting to see a [performance] plateau for general software -- it has stopped improving at historical rates -- but this AI stuff is still increasing rapidly,' he says."

"In the longer term, more radical changes in how computer chips work will be required to keep AI getting more powerful."

"You may think you're good at reading other people's faces and moods. But when it comes to telling the difference between a person who is suicidal and one who is merely depressed, the signs are much more subtle. To tease out the differences, the researchers looked at a handful of facial gestures displayed by three groups of people -- those with suicidal ideation, depressed patients, and a medical control group. During interviews with these groups, they recorded gestures that included smiling, frowning, eye brow raising, and head motioning. The data was then fed into a machine-learning algorithm that looked for correlations between different gestures, alone or in combination, and patient groups."

"Remarkably, rather than frowning being the most telling feature in discriminating between depressed and suicidal patients, it was smiling. Specifically, Duchenne smiles versus non-Duchenne smiles held the key to differentiating the groups."

Why AI works. +Monica Anderson elaborates on the distinction between model-based and model-free systems and how the AI community's rapid progress since 2012 has largely been due to shifting from model-based approaches (like your brain's "slow" or system 2 thinking, in the parlance of Daniel Kahneman's "Thinking Fast and Slow") to model-free deep learning approaches (like your brain's "fast" and intuitive system 1 thinking) in the inaugural post of her new Artificial Understanding blog.

"To an outsider, it may look like an older program or phone app might be 'Understanding' whatever the app is doing, but that Understanding really only happened in the mind of the programmer creating the app. The programmer first simplified the problem in their own head by discarding a lot of irrelevant detail using 'Programmer's Understanding'. This simplified mental 'Model' of the problem domain could then be explained to a computer in the form of a computer program. What is changing is that computers are now making these Models themselves."

Garry Kasparov, the chess legend who got beat by Deep Blue in 1998, and Demis Hassabis, leader of the team behind AlphaGo which recently beat the best human Go player, have a chat. Experiences playing chess computers, before, during, and after Deep Blue. How players learn differently now. Machines and humans working together. The Moravec paradox: computers are good at what humans are bad at and vice-versa.

A new method of ageing facial images has been developed. "A key feature of the method is that it teaches the machine how humans age by feeding the algorithm facial feature data from a large database of individuals at various ages. Consequently, the method improves on existing techniques, achieving greater level of accuracy."

"As a test case, the researchers chose to work on the case of Ben Needham. Ben disappeared on the Greek island of Kos on 24th July 1991, when he was only 21 months old. He has never been found, but several images have been produced by investigators showing how Ben might look at ages 11-14 years, 17-20 years, and 20-22 years. The team used their method to progress the image of Ben Needham to the ages of 6, 14 and 22 years. The resulting images show very different results, which the researchers believe more closely resemble what Ben might look like today."

The highest possible score on Ms. Pac-Man, 999990, has been achieved by an AI, higher than any score ever achieved by any human. It was achieved by breaking the game into 4 sub-problems, and developing a separate reinforcement learning algorithm for each problem. The 4 problems are: one that is rewarded for eating a pellet, one for eating a fruit, one for eating blue ghost, and one with a large negative reward if Ms. Pac-Man gets eaten by a ghost. An aggregator looks at all 4 to decide what action Ms. Pac-Man takes. The AI was developed by Maluuba, recently acquired by Microsoft and now part of Microsoft Research.

Chatbots with the ability to negotiate have been developed by Facebook. "The Facebook Artificial Intelligence Research (FAIR) researchers studied negotiation on a multi-issue bargaining task. Two agents are both shown the same collection of items (say two books, one hat, three balls) and are instructed to divide them between themselves by negotiating a split of the items. Each agent is provided its own value function, which represents how much it cares about each type of item (say each ball is worth 3 points to agent 1). As in life, neither agent knows the other agent's value function and must infer it from the dialog (you say you want the ball, so you must value it highly)."

"FAIR researchers created many such negotiation scenarios, always ensuring that it is impossible for both agents to get the best deal simultaneously. Furthermore, walking away from the negotiation (or not agreeing on a deal after 10 rounds of dialog) resulted in 0 points for both agents. Simply put, negotiation is essential, and good negotiation results in better performance."

"The FAIR researchers' key technical innovation in building such long-term planning dialog agents is an idea called dialog rollouts." "FAIR has developed dialog rollouts as a novel technique where an agent simulates a future conversation by rolling out a dialog model to the end of the conversation, so that an utterance with the maximum expected future reward can be chosen."

"After years toiling away in secret on a car project, Apple Inc. Chief Executive Officer Tim Cook has for the first time elaborated on the company's plans in the automotive market."

"We're focusing on autonomous systems. It's a core technology that we view as very important." "He likened the effort to 'the mother of all AI projects,' saying it's 'probably one of the most difficult AI projects to work on.'"