It is good practice to gather a population of results when comparing two different machine learning algorithms or when comparing the same algorithm with different configurations. Repeating each experimental run 30 or more times gives you a population of results from which you can calculate the mean expected performance, given the stochastic nature of most machine learning algorithms. If the mean expected performance from two algorithms or configurations are different, how do you know that the difference is significant, and how significant? Statistical significance tests are an important tool to help to interpret the results from machine learning experiments. Additionally, the findings from these tools can help you better and more confidently present your experimental results and choose the right algorithms and configurations for your predictive modeling problem.

Machine learning has been increasingly capable of accurately diagnosing a range of physical and mental health conditions in recent years. I've written previously about algorithms that monitor things like speech to detect the onset of conditions such as Alzheimer's, whilst applications have also used mobile phone data to detect changes in lifestyle and possible depression in individuals. A recent study from the University of Texas at Austin takes a slightly different tact by using AI to spot vulnerability to depression from brain imaging. The researchers worked with a supercomputer to train the algorithm to detect commonalities in MRI scans, genomic data and various other datasets relevant to depression and anxiety. It aims to improve upon previous work by researchers who have studied mental disorders via the relationship between brain function and structure in neuroimaging data.

The Xprize foundation has whittled its latest competition down from 10,000 sign ups to 147 teams in its first-ever "open" competition. Announced in June of last year, the IBM Watson AI Xprize is the first not to identify a specific problem. Rather, teams are asked to define their own issue and address it using artificial intelligence. "Unlike other prizes that have a very specific end-point in mind, we're asking teams to be more free, in terms of the problems they'll be solving," competition leader Amir Banifatemi told TechCrunch in a conversation this week. "Because AI is involved in everything, we're asking teams to frame the problems that are grand challenges, and use those grand challenges as a backdrop and try to come up with a solution address them."

At universities, "people learn how to approach new things, ask questions and find answers, deal with new situations," wrote Uta Russmann, a professor of communications at the FHWien University of Applied Sciences in Vienna. "All this is needed to adjust to ongoing changes in work life. Special skills for a particular job will be learned on the job." Schools will also need to teach traits that machines can't yet easily replicate, like creativity, critical thinking, emotional intelligence, adaptability and collaboration. The problem, many respondents said, is that these are not necessarily easy to teach.

File- In this June 5, 2014 file photo, a Border Patrol agent uses a headset and computer to conduct a long distance interview by video from a facility in San Diego. A government report says inefficient computer systems and a lack of exit screening hamper U.S. authorities' efforts to track the number of foreigners who overstay their visas. The Department of Homeland Security's Office of Inspector General said Thursday, May 4, 2017, that immigration agents and analysts need better and more integrated computer systems to check visitors' immigration status.

Elon Musk's electric car firm Tesla has revealed first-quarter revenue that more than doubled and said its upcoming Model 3 was on schedule for July - but downplayed the mass-market vehicle and gave a sales pitch for its more expensive Model S. The firms first-quarter loss widened 17 percent to $330 million as it ramped up spending ahead of the launch of its Model 3 sedan and its solar panel business. Elon Musk has likened future versions of Tesla's Model 3 production line to an'alien dreadnaught' - and last week the first pictures of the factory emerged. Elon Musk's electric car firm Tesla has revealed first-quarter revenue that more than doubled and said its upcoming Model 3 was on schedule for July Chief Executive Elon Musk's bold approach to cars, space exploration and clean energy has fueled investor enthusiasm for Tesla, although skeptics are waiting to see if Musk can fulfill his promise of producing 500,000 cars per year in 2018, six times Tesla's 2016 production. Tesla's comments underscored the additional challenge of keeping up demand for its older models. Shares were down about 1 percent in after-hours trade following the results.

With advances in machine learning and AI, the CFO will finally be able to shift their focus back to strategic thinking. Many CFOs are loath to consider AI because they see it as a risk. According to Intacct's Chief Technology Officer Aaron Harris, however, automation will be a key resource that the finance industry must embrace. FEI Daily spoke with Harris about where automation is intersecting with the financial industry today and how CFOs can refocus on strategy by embracing AI. FEI Daily: Why is an understanding or AI and machine learning so important for financial executives?

Artificial intelligence can predict Supreme Court decisions better than some experts. "See you in the Supreme Court!" President Donald Trump tweeted last week, responding to lower court holds on his national security policies. But is taking cases all the way to the highest court in the land a good idea? Artificial intelligence may soon have the answer. A new study shows that computers can do a better job than legal scholars at predicting Supreme Court decisions, even with less information.

KUKA also had good Q1/17 earnings as did Intuitive Surgical. As a result, iRobot adjusted upward their 2017 revenue forecast. According to CEO Colin Angle, iRobot has shipped over 15 million robotic home floor cleaners through 2016. Last year iRobot divested its Defense and Security Division for $45 million to a VC which shortly thereafter launched Endeavor Robotics. Thus iRobot is now exclusively a commercial robotics provider.

Recurrent Neural Networks (RNN) are a type of statistical model designed to handle sequential data. The model reads a sequence one symbol at a time. Each symbol is processed based on information collected from the previous symbols. With existing RNN architectures, each symbol is processed using only information from the previous processing step. To overcome this limitation, we propose a new kind of RNN model that computes a recurrent weighted average (RWA) over every past processing step. Because the RWA can be computed as a running average, the computational overhead scales like that of any other RNN architecture. The approach essentially reformulates the attention mechanism into a stand-alone model. The performance of the RWA model is assessed on the variable copy problem, the adding problem, classification of artificial grammar, classification of sequences by length, and classification of the MNIST images (where the pixels are read sequentially one at a time). On almost every task, the RWA model is found to outperform a standard LSTM model.