Researchers from Harvard and the University of Vermont have found a way to use Instagram to detect depression. Using machine learning tools, they developed a model that can predict whether a person is clinically depressed with surprising accuracy, just by looking at their Instagram photos. It's important to note right away that this model is just the beginning. The researchers themselves only go so far as to say that "these findings suggest new avenues for early screening and detection of mental illness," and "the findings reported here should not be taken as enduring facts." That said, their results are impressive.

Does the hotel search result you view on an online travel site a perfect match, just in the ballpark or even irrelevant? Six-year-old metasearch site Hipmunk quietly began offering desktop-site visitors the ability to enter natural language queries such as, "I want a cheap hotel on the Upper West Side" in a bid to make searching easier and to provide more relevant results. Travelers type natural language keywords into a search field and receive corresponding results with real-time availability and rates. The update isn't immediately obvious and requires some guesswork to locate as there's no banner or text indicating users can perform such a search. But various keyword searches we tried yielded relevant results.

In the current issue of Science, Stanford researchers propose an accurate way to identify poverty in areas previously void of valuable survey information, says the news statement. "We have a limited number of surveys conducted in scattered villages across the African continent, but otherwise we have very little local-level information on poverty," said study coauthor Marshall Burke, an assistant professor of Earth system science at Stanford and a fellow at the Center on Food Security and the Environment. "At the same time, we collect all sorts of other data in these areas – like satellite imagery – constantly." According to Stanford, the researchers sought to understand whether high-resolution satellite imagery – an unconventional but readily available data source – could inform estimates of where impoverished people live. "The difficulty was that while standard machine learning approaches work best when they can access vast amounts of data, in this case there was little data on poverty to start with," the release says.

The general perception of machine learning is that it involves sophisticated algorithms and predictive models that far surpass the understanding of the business domain expert that relies upon this technique to glean analytic insight. But it doesn't have to. According to Paxata Vice President of Corporate and Product Marketing Michele Goetz--a former analyst for Forrester--an artful combination of machine learning, data visualizations, natural language processing, and a heaping of semantic technologies can render this technology functional to business analyst end users for enhanced analytic insight. "There are analytics and visualizations that pull out of the machine learning that provide a bit of insight without having to go fully down the analytics path to get that type of insight," Goetz commented. "Just having something that you're continuously refining and tuning is interesting when you're working in a much more exploratory fashion."

Even if you haven't tried it, everyone knows about Siri, Apple's speech interpretation and recognition personal assistant. The bot is available at the touch of a button or voice activation to answer questions about tomorrow's weather or recommend a nearby coffee shop, all with its signature sassy style. Siri was just an apple in Alan Turing's eye when he developed a test in the 1950s to see if machines could think. This was a highly influential first step in the development of artificial intelligence, but when Apple wowed an auditorium of cheering spectators five years ago at its launch, Siri was one of the first technologies to set expectations for technologists and consumers about what they could actually do with these capabilities. Because of Siri and other computer assistants, consumers finally have their own notions of what A.I.-enabled technologies should be able to do.

ArXiv sanity clean for text-to-speech bookmark. How to use text-to-speech to create audio books and read science papers: https://www.youtube.com/watch?v R2b9d...

While the oil and gas industry has had its share of ups and downs over the past decade, many financial institutions are banking on a very slow growth of oil prices in 2017. Though some believe that the efficiency gains that the oil industry can capture are quickly coming to an end, this sentiment is only capturing hard technology specifically related to oil and gas. To help bring the O&G industry to the 21st century, technology from other industries needs to be incorporated, using many hard-earned years of expertise and different lines of thinking. Oilprice previously mentioned incorporating food industry technology to increase safety standards when fracking, but incorporating technology from the IT industry is something that the O&G industry as a whole can benefit from. Whether its neural networks, machine learning, fuzzy logic, case-based reasoning or expert systems, AI has the potential to transform the industry.

Germany's Interior Minister says he wants to introduce facial recognition software at train stations and airports to help identify suspects following two attacks in the country last month. In a report published on Sunday in the German newspaper Bild am Sonntag, Thomas de Maiziere said internet software was able to determine whether persons shown in photographs were celebrities or politicians. "I would like to use this kind of facial recognition technology in video cameras at airports and train stations. Then, if a suspect appears and is recognised, it will show up in the system," he told the paper. Germany's Thomas de Maiziere takes aim at face veils He said a similar system was already being tested for unattended luggage, which the camera reports after a certain number of minutes.

XGBoost is a popular implementation of Gradient Boosting because of its speed and performance. Internally, XGBoost models represent all problems as a regression predictive modeling problem that only takes numerical values as input. If your data is in a different form, it must be prepared into the expected format. In this post you will discover how to prepare your data for using with gradient boosting with the XGBoost library in Python. Data Preparation for Gradient Boosting with XGBoost in Python Photo by Ed Dunens, some rights reserved.

"Speaker: Ben Hamner This tutorial will offer an introduction machine learning and how to apply it to a Kaggle competition. We will cover methodologies that have worked well across a diverse set of problems, and then work on a current Kaggle competition together using iPython notebook and scikit-learn. We will cover concepts including feature extraction, feature selection, model evaluation, and data visualization.