Inside | Real news, curated by real humans
Inside AI

Inside AI (Sep 11th, 2019)

1. U.S. federal government agencies are requesting nearly $1 billion in nondefense AI research spending for the next year. The amount, unveiled Tuesday, has drawn a negative response from companies such as Intel Corp. and Nvidia Corp., which argue that the U.S. should invest more into AI. Jackie Medecki, director and managing attorney of Intel’s U.S. AI and health-care policy division, said the U.S. is "nowhere near where we should be” when it comes to government funding for AI research. Trump administration officials responded by saying that there is no alleged disparity between U.S. government spending on AI R&D and China's, for example. In the U.S., " you won’t find aspirational expenditures or cryptic funding mechanisms," said U.S. CTO Michael Kratsios. Tuesday's announcement of $973.5 million toward AI research is the first time the federal government tallied agency-specific requests for that type of spending. - WSJ

  • Email gray
  • Permalink gray

2. A new study published on ArXiv describes an AI that can turn a person's selfie into a personalized video game character. The tool, from developers at NetEase, maps the person's features and generates them into a character based on the Chinese game “Justice Online." According to Futurism, the AI more closely matches players' faces because it was trained to reconstruct them based on bone structure, rather than skin. - FUTURISM

  • Email gray
  • Permalink gray

3. The Pentagon is seeking an AI ethicist for its Joint Artificial Intelligence Center. The AI ethical adviser would help shape the way the center incorporates AI capabilities in defense work. "We are going to bring in someone who will have a deep background in ethics, and then the lawyers within the department will be looking at how we actually bake this into the Department of Defense," said Air Force Lt. Gen. Jack Shanahan, the JAIC's director, who previously led Google's controversial Project Maven for the department. - THE GUARDIAN

  • Email gray
  • Permalink gray

4. The startup Explorium, which developed a data discovery platform for ML models, has raised a total of $19.1 million over the last year. The startup, which made two funding rounds public as of Tuesday, raised a $3.6 million seed round last September led by Emerge with participation from F2 Capital. It was followed by a $15.5 million Series A round in March, led by Zeev Ventures with participation from the seed investors. The company, which has 63 employees in three offices, built a data discovery platform for scientists to locate the best data for their models. - ZDNET

  • Email gray
  • Permalink gray

5. Several factories around the world have started using AI and IoT robots on their shop floors, leaving the labor force to wonder what the lasting impacts will look like. Currently, there are 2.25 million robots used worldwide and trends suggest this number will increase to 20 million by 2030. While some industry experts say the future looks more human than you might expect, a newly installed robot displaces 1.6 factory workers. Robots can work around the clock without tiring, eating or needing a break, and some factories in Japan are completely robot-run. - LIVE MINT

A version of this story first appeared in Inside IoT.

  • Email gray
  • Permalink gray

6. Denmark organizations launched a project that would use AI to deliver information about sea-ice conditions to ships and other maritime causes. The Automated Sea Ice Products (ASIP), initiated by the Danish Meteorological Institute (DMI) and Technical University of Denmark, aims to provide more timely and detailed maps about Arctic sea ice automatically. It uses a convolutional neural network trained on datasets of ice charts. - TECHXPLORE

  • Email gray
  • Permalink gray

7. A new AI that combines NL and 3D pose models can translate words into human-like movement. The Joint Language-to-Pose, or JL2P, model was developed by AI researchers at Carnegie Mellon University. According to Venture Beat, JL2P can walk or run, play musical instruments, control speed, and follow directional instructions. CMU Language Technology Institute researcher Chaitanya Ahuja plans to present the research at the International Conference on 3D Vision in Quebec on Sept. 19. - VENTURE BEAT

  • Email gray
  • Permalink gray

8. LVMH selected AI handbag counterfeit app Entrupy as one of 26 companies that will participate in its La Maison des Startups incubator program. Entrupy can distinguish among 15 different brands of luxury handbags, using around 500 data points per bag, to determine if if is counterfeit or not. The system has reached an accuracy of 99.1 percent. - HYPEBEAST

  • Email gray
  • Permalink gray

9. Authors David A. Bray and R “Ray” Wang make a case for embracing people-centered principles for AI as organizations move into third wave "perception AI," which focuses more on deep learning. The principles aim to benefit all individuals and communities, versus benefitting only a few at the expense of others. The authors suggest three key principles - transparency, explainability, and reversibility - that organizations should consider for any AI implementation. Bray is the executive director of the People-Centered Internet coalition and senior fellow at the Florida Institute for Human & Machine Cognition, and Wang is the CEO of Constellation Research in Palo Alto, California. - MIT SLOAN MANAGEMENT REVIEW

  • Email gray
  • Permalink gray

10. AB InBev, which produces Budweiser, is using machine learning to make better quality beer. The company has several initiatives that incorporate AI and big data, including the use of the SenseAI system to measure data from the brewing process and predict a beer's end quality. The company, which brews Stella Artois and Corona, also uses a risk analysis model to determine the amount of stock a distributor needs and how much credit to offer them. AB InBev has a "Beer Garage" based in Silicon Valley, where it develops and tests its AI-powered initiatives. - FORBES

  • Email gray
  • Permalink gray

CORRECTION: Yesterday's newsletter incorrectly identified Mayo Clinic's headquarters as being in Rochester, New York. They are located in Rochester, Minnesota. Thanks to a reader for pointing it out!

  • 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.

Editor: Kim Lyons (Pittsburgh-based journalist and managing editor at Inside).

  • Email gray
  • Permalink gray

Subscribe to Inside AI