Machine Learning Roundup 10/14/2016
Machine learning (ML) and deep learning (DL) content from the past 24 hours.
Note to readers: Big Analytics will rebrand as ML/DL on Monday, October 17.
— Cynthia Harvey explains the difference between AI, ML, and DL.
— Arun Krishnan asks: can algorithms reinforce our biases?
— In Nature, Kate Crawford and Ryan Calo note rising use of AI, summarize the pros and cons of three different approaches to managing the impact of the technology and propose an approach they call social systems analysis.
New Memory System Speeds Deep Learning
— Fujitsu announces the development of a GPU memory system that will enable more layers in a deep neural network without compromising speed.
Adapteva Targets Deep Learning Workloads
— Adapteva announces the Epiphany-V processor with 1,024 64-bit cores.
— In the Washington Post, Brian Fung reports that China has eclipsed the US in AI research.
Health and Medical Applications
— Teams use machine learning to mine data from wearables, monitor pregnancy complications.
— Data scientists at Intermedix use a Dataiku stack to predict patient compliance, optimize interventions.
— Scientists at Insilico Medicine use deep learning to find new indications and targets for existing drugs.
— Tinder uses machine learning to predict “swipe-rights.”
— Srini Penchikala interviews Peter Cnudde, who describes how
Yahoo Verizon uses machine learning.
— Ocado uses TensorFlow to categorize incoming messages to its customer service center.
— Venture capitalist Jim Hogan says he’s betting on deep learning.
— Thomas Wieberneit summarizes what SAP, Oracle, Salesforce, Microsoft, Adobe and a few other technology companies are doing with machine learning.
— ComplyAdvantage lands an $8.2 million “A” round. The company uses AI and ML to help companies reduce the cost of compliance in areas such as anti-money laundering, counter-financing of terrorism and Politically Exposed Persons.
— Tachyus, a startup that uses ML to help oil companies increase production and decrease costs, lands $5 million in new financing.