Inside | Real news, curated by real humans
Inside AI

Inside AI (Jul 10th, 2019)

1. The European Commission earmarked €35 million ($40 million) toward AI research that could boost the early detection of cancer. The E.U. wants organizations to develop image recognition systems that can help diagnose the most common types of cancer, potentially speeding up referrals and improving the timing and accuracy diagnoses. The AI systems are capable of reading MRI, microscopic, and PET scans. Applicants for the funding are asked to contribute to data stores. - SCIENCE | BUSINESS

  • Email gray
  • Permalink gray

2. Researchers used natural-language processing to analyze 3.3 million scientific abstracts on material science research. According to a new paper, published last week in the journal Nature, the word relationships found by the AI "captured fundamental knowledge within the field, including the structure of the periodic table and the way chemicals’ structures relate to their properties." They also found some chemical compounds with properties akin to thermoelectric materials but haven't been studied yet. - MIT TECH REVIEW

  • Email gray
  • Permalink gray

3. A new Genesys survey found that 70 percent of U.S. workers polled reported a positive outlook on AI tech. The Workplace Attitudes to Artificial Intelligence report, released on Wednesday, surveyed 1,000 workers and found that:

  • 75 percent said they rarely or never feel threatened by AI technology
  • 32 percent said they believe AI will have a positive impact on their job within the next five years
  • 66 percent said they believe AI will make them more efficient at their jobs

- TECHREPUBLIC

  • Email gray
  • Permalink gray

4. Voicebot.ai pitted the top smart assistants against each other to determine which one could accurately answer questions about specific brands — and Google Assistant won by a long shot. The Voicebot.ai team asked the assistants more than 4,000 questions, including queries like "How do I contact JetBlue?" and "What is the longest-lasting lipstick?" The phone-based Google Assistant correctly answered 92 percent of questions, and the smart-speaker based Google Assistant landed at 81 percent. On the other hand, Amazon's Alexa only answered 34 percent correctly, Samsung's Bixby landed at 32 percent, and Apple's Siri on the HomePod answered 28 percent correctly. - CNET

This item first appeared in Inside IoT.

  • Email gray
  • Permalink gray

5. British choreographer Wayne McGregor partnered with Google Arts & Culture to develop an AI system that can create new dance moves based on his unique style. McGregor worked with Google engineers over two years to train the "Living Archive" algorithm using thousands of hours of video from his works over the last 25 years. It learned the distinct moves performed by each of his 10 dancers and provided suggestions for future choreographed dances. The result is a program, “Living Archive: An AI Performance Experiment,” that debuts on Friday at the Music Center’s Dorothy Chandler Pavilion. - LATIMES

  • Email gray
  • Permalink gray

6. Researchers at Penn State and Columbia University created an AI tool that can detect discrimination based on attributes like race or gender. The tool employs counterfactual inference algorithms to guess answers to hypothetical questions based on the concept of causality. - PHYS.ORG

  • Email gray
  • Permalink gray

7. A new model shares machine learning applications with people who can't code. Runway ML is available for download in beta and connects designers, artists, and other creatives with an "app store" of machine learning techniques. Some examples of uses include using a neural network called “PosetNet" in place of buying expensive motion capture equipment, or running video  footage through a machine learning segmentation model to remove items rather than painting it over by hand. - THE VERGE

  • Email gray
  • Permalink gray

8. The AI startup Kyndi Inc. raised $20 million in a Series B funding round. Intel Capital led the round, with prior investors UL Ventures and PivotNorth Capital also participating. Silicon Valley-based Kyndi, which developed an AI that can summarize large amounts of text data, plans to use the infusion to hire more employees, particularly in its sales and engineering departments. - SILICONANGLE

  • Email gray
  • Permalink gray

9. Alibaba set a new record in AI reading comprehension. The AI program using the Microsoft Machine Reading Comprehension (MS MARCO) data set was based on questions that Bing users have asked. According to two measures, the program's scores surpassed humans' in some cases, although it is not technically better at reading comprehension than people since it's merely using statistical pattern recognition without comprehending the meaning of the words. - MIT TECH REVIEW

  • Email gray
  • Permalink gray

10. ITProPortal lists the top 5 programming languages right now in AI and machine learning, including the pros and cons of each. The languages, which are supported by machine and deep learning libraries, are Java for AI and ML, Python for AI and ML, C++ for AI and ML, R for AI and ML, JavaScript for AI and ML. - ITPROPORTAL

  • Email gray
  • Permalink gray

Written and curated by Beth Duckett in Orange County. Beth is a former reporter for The Arizona Republic who has written for USA Today, Get Out magazine and other publications. Follow her tweets about breaking news and other topics in southern California here.

Editing team: Kim Lyons (Pittsburgh-based journalist and managing editor at Inside), David Stegon (senior editor at Inside, whose reporting experience includes cryptocurrency and technology), and Bobby Cherry (senior editor at Inside, who’s always on social media).

  • Email gray
  • Permalink gray

Subscribe to Inside AI