This week - Microsoft gets exclusive licensing access to GPT-3; a robot beats world-class curling teams; a long-discussed legal question involving self-driving cars is answered; and more!
More Than A Human
This article describes how this amazing brain-controlled dress was made. The dress is the result of a collaboration between researchers at Johannes Kepler University Linz (JKU), developers at medical engineering company G.tec, and fashiontech designer Anouk Wipprecht. Known as the Pangolin dress, it has 1,024 individual head-mounted electrodes in 64 groups of 16. These detect electrical signals coming from the brain. The data from these sensors is combined, analyzed, and converted into colors displayed by 32 Neopixel LEDs and 32 servo-driven scales, creating a whole-body visualization of neural activity.
Scientists have developed a backpack that tracks and stimulates brain activity as people go about their daily lives. The advance could allow researchers to get a sense of how the brain works outside of a laboratory—and how to monitor diseases such as Parkinson’s and post-traumatic stress disorder in real-world settings.
I wonder if Microsoft is going to buy OpenAI to its own DeepMind. They already invested $1 billion to build ‘Supercomputer AI’ hosted on Azure for OpenAI last year. Later, OpenAI used Microsoft’s cloud infrastructure to develop GPT-3. What are your thoughts?
AI researchers see International Mathematical Olympiad as a great proving ground for machines designed to think like humans. Together, they launched IMO Grand Challenge, where the goal is to build an AI that can win a gold medal in the competition. “Maybe in Go the goal is to find the best move, whereas in math the goal is to find the best game and then to find the best move in that game,” said Daniel Selsam, a member or the challenge committee, to describe what the AI needs to do.
Chinese streaming platform iQIYI has announced that it’s launching the country’s first-ever “virtual idol variety show,” in a move that’s somehow both pretty out-there and oddly predictable. Titled Dimension Nova, the new show will feature “more than 30 virtual competitors [navigating] a series of competitions and challenges to select the final winner.”
This article lists some recent examples where racial and gender bias leads to removing black people on Zoom when a virtual background is applied or Twitter removing images of women in previews. Then it shows what are the consequences of such bias - from silly things like soap dispenser not able to recognise darker skin colours to serious like being wrongly accused of crime - and what the AI community can do to fix this problem.
Dagogo Altraide from YouTube channel ColdFusion (I enjoy what he's doing) explains what Generative Adversarial Networks (GANs) are. In this video, he shows how GANs are the closest we have right now to creativity in AIs with multiple examples, like toonification or upscaling old movies to 4K 60fps.
When GPT-3 came out, OpenAI was boasting that their text generator uses 175 billion machine learning parameters. Last week, AI researchers from the Ludwig Maximilian University (LMU) of Munich have shared their GPT-3-level text generator which uses only 223 million parameters. If you are interested in more details, their paper is available on arXiv.
Another sport where robots are now better than humans. A robot named Curly has beaten top-ranked human opponents from South Korean curling teams.
For years, people were debating the legal implications involving self-driving cars. The main question was if a self-driving car is involved in a fatal crash, who should be responsible - the driver, the car manufacturer or the AI? Well, we have an answer now. Prosecutors in Arizona charged the safety behind the wheel of a self-driving Uber that struck and killed a woman in 2018 has been charged with a crime. Uber, the safety driver employer and the company that built the automated system involved in the fatal collision, won’t face charges.
Well-intentioned “citizen scientists” developing homemade COVID-19 vaccines may believe they’re inoculating themselves against the ongoing pandemic, but the practice of self-experimentation with do-it-yourself medical innovations is fraught with important legal, ethical and public health issues, according to a new paper in the journal Science co-written by a University of Illinois, Urbana-Champaign legal expert who studies the policy implications of advanced biotechnologies.