1 – An AI Ophthalmologist Shows How Machine Learning May Transform Medicine
Google researchers are disrupting the medical field with a new algorithm capable of diagnosing eye disease. The team has been collaborating with the Aravind Medical Research Foundation in India to test-drive a deep-learning algorithm, trained to spot and diagnose a common cause of blindness known as diabetic retinopathy. The method overlaps with the Google’s machine learning approach for labeling an influx of web images. While AI systems have had mixed success in medical diagnostics in past trials, this new algorithm has so far matched or surpassed ophthalmologists in diagnosing the condition in patients. This type of automated detection could not only increase reliability and efficient of diagnoses, but also serve patients in places where human expertise is limited. Google Researcher Lily Peng voiced that one of the present challenges is developing a a system that can explain its findings.
2 – Welcoming a New Leader for Intel’s IoT Business and Elevating Our Automated Driving Efforts
This week, Intel announced its decision to bring on Tom Lantzsch to manage its IoT group. Lantzsch has 30 years of experience in Fortune 500 and early-state startups, and spent the last decade as executive vice president of stategy at ARM. In addition to new hires, Intel also revealed its new Automated Driving Group (ADG), which will focus 100 percent of its efforts on driver assist and autonomous driving systems. Doug Davis, previously manager of the IoT group, has been appointed the new senior general manager of the ADG group, along with Intel’s Kathy Winter as general manager. Intel’s decision to dedicate an entire arm to autonomous driving solutions speaks to the technology’s accelerating emergence and increased collaboration between automakers and advanced technology companies.
(Read the full press release at Intel Newsroom)
3 – Here are the Top 20 AI Investors … and Their Biggest 2016 Deals
AI- and machine learning-based startups and companies have attracted an increasing share of venture capitalists’ funds (particularly in the Silicon Valley), according to a report done for TechFlash Silicon Valley by PitchBook Data. Unique from previous years, there were no overlaps on AI deals amongst the top 20 investors – Lux Ventures, for example, was the only firm involved in the $300M raised for Menlo Park-based Zoox, a self-driving auto startup. The link provided below showcases the top AI investors, number of deals for the year, and their top-funded venture.
(Read the full article on Silicon Valley Business Journal)
4 – AI Songsmith Cranks Out Surprisingly Catchy Tunes
A Google Research Scientist and an Intern recently unveiled the latest version of Google’s song-writing AI, which is part of a larger AI creativity project named Magenta. Douglas Eck and Natasha Jaques trained recurrent neural networks with a reinforcement learning (RL) tuner algorithm, allowing the system to retain information learned from past experiences while also following a set of musical rules. The result is AI-composed tunes that are more “elegant and catchy” rather than “flat and mechanical”. Eck believes this type of RL algorithm will be applied to many other areas beyond music, including robotics, language translations, and recommender engines.
5 -Facebook Looks to Harness Artificial Intelligence to Weed Out Fake News
During a recent round table at Facebook’s headquarters, Director of AI Yann LeCun posed some important questions to reporters regarding the recent media clamor around Facebook’s “fake newsfeeds”:
“What’s the trade-off between filtering and censorship? Freedom of experience and decency?”
LeCun noted that Facebook has the AI expertise needed to help create systems that can filter out fake news, violent videos, and other controversial content, but that its policy and product teams are still grappling with big questions like those above. How Facebook could use AI technology is not clear, and the risk is implementing an AI filter that blocks too much content. Founder Mark Zuckerberg wrote a Facebook posting two weeks ago about the company’s approach to the issue, outlining seven projects the company is currently working on that could help defend against or disrupt fake news.
(Read the full article on The Wall Street Journal)
Image credit: BetaNews