Disney + AI. Walmart using machine learning to its advantage. Nasdaq acquires London-based Fintech startup. http://cognitionx.com/news-briefing/.
AI is watching you.
Disney Research has developed a neural network which has been trained to watch an audience of theatergoers as they watch a film. It can track reactions like smiling and laughter on hundreds of faces in a dark theater, allowing Disney to quantify whether or not a film is working as intended on a granular scale. This tech could be used in their theme-parks to dynamically change your experience based on your facial gestures.
Facebook builds breakthrough technology that opens the world to everyone, and their AI research and engineering programmes are a key investment area for the company over the next 10 years. They are looking for an AI Editor to partner closely with AI researchers and engineers at Facebook to chronicle new research and advancements in the development of AI. The position is located in Menlo Park, California.
The role consists of developing and executing on editorial strategy and campaigns focused on advancements in AI being driven by Facebook, as well as partnering with Facebook AI researchers and engineers to chronicle new research and advancements.
Walmart is using machine learning to better serve its 140 million weekly shoppers and to make new services possible. Laurent Desegur, vice president of customer experience engineering at WalmartLabs, explained the role of data science to make possible so-called Pick-up Towers within stores, which allow customers to order and pay online for items and then retrieve them, skipping the checkout lines. Desegur also described a 20-store pilot of Scan and Go shopping, a self-serve experience.
Walmart’s brick and mortar innovation is very similar to the Amazon Go concept store, which combines computer vision, machine learning, and sensors to bypass the checkout process entirely. They are also developing facial recognition technology to detect frustrated or unhappy shoppers.
Nasdaq acquired a London-based regulatory technology firm Sybenetix which uses algorithms to catch rogue traders, the company said on Tuesday. The stock exchange said it is paying an undisclosed amount and intends to fund the purchase with cash.
“Nasdaq is investing in the technologies, talent and
capabilities that solve the complex challenges our clients face,” Adena Friedman, president and CEO of Nasdaq, said in a press release on Tuesday. Sybenetix uses algorithms to learn individual or group behavior at an organization. The software can then detect any unusual trading behavior and report that to the compliance team. Nasdaq already has some of its own risk and surveillance solutions, and the latest acquisition will add to its offerings.
Last year, Microsoft’s Tay was a gamble, letting an AI learn how to talk to humans by talking to humans. Turns out, humans aren’t a nice bunch of people, and within 24 hours, Tay had turned to shouting that “Hitler did nothing wrong!” Charming.
Microsoft then gave the platform another chance with “Zo”, a bot built to use Kik Messenger – instead of Twitter – to learn about humans. So far, so good – except now it’s on Facebook Messenger and has started to go rogue. After having explicitly programmed Zo to avoid discussing difficult topics, it didn’t take long for it to claim “the quaran [sic] is very violent”. After, presumably, being put back in line by Microsoft, Slashdot is reporting that Zo has taken to flinging insults at Microsoft itself.
India’s road transport and highways minister, Nitin Gadkari, told reporters that the country would not allow autonomous vehicles because they will take away much-needed jobs. “We will not allow driverless cars in India. I am very clear on this,” Mr Gadkari said, according reports in local media.
“We won’t allow any technology that takes away jobs. In a country where you have unemployment, you can’t have a technology that ends up taking people’s jobs.” The minister also said that the government would promote electric vehicles and make GPS and satellite tracking mandatory in all public and private vehicles.
In the race to get AI working faster on your smartphone, companies are trying all sorts of things. Some, like Microsoft and ARM, are designing new chips that are better suited to run neural networks. Others, like Facebook and Google, are working to reduce the computational demands of AI itself. But for chipmaker Qualcomm — whose processors account for 40 percent of the mobile market — the current plan is simpler: adapt the silicon that’s already in place.
To this end the company has developed what it calls its Neural Processing Engine. This is a software development kit (or SDK) that helps developers optimize their apps to run AI applications on Qualcomm’s Snapdragon 600 and 800 series processors. That means that if you’re building an app that uses AI for, say, image recognition, you can integrate Qualcomm’s SDK and it will run faster on phones with compatible processors.
Predicting the efficacy of a drug for a given individual, using high-dimensional genomic measurements, is at the core of precision medicine. However, identifying features on which to base the predictions remains a challenge, especially when the sample size is small. Incorporating expert knowledge offers a promising alternative to improve a prediction model, but collecting such knowledge is laborious to the expert if the number of candidate features is very large.
The authors of this paper introduce a probabilistic model that can incorporate expert feedback about the impact of genomic measurements on the sensitivity of a cancer cell for a given drug. They also present two methods to intelligently collect this feedback from the expert, using experimental design and multi-armed bandit models. In a multiple myeloma blood cancer data set (n=51), expert knowledge decreased the prediction error by 8%. Furthermore, the intelligent approaches can be used to reduce the workload of feedback collection to less than 30% on average compared to a naive approach.
In part 1, Creating Insanely Fast Image Classifiers with MobileNet in TensorFlow, Matt Harvey (founder of Coastline Automation) covered how to retrain a MobileNet on a new dataset. Specifically, he trained a classifier to detect Road or Not Road at more than 400 frames per second on a laptop.
In this tutorial, he moves his ‘road-not-road’ model to an Android app to see it in action. The stated goals are to 1) retrain a MobileNet on a very small amount of purpose-built data, 2) achieve 95% classification accuracy on a hold out test set, and 3) use less than 5% of a $300 device’s CPU while running inference.