Mini Briefs

Scientists develop AI that can turn brain activity into text

This article begins by stating “reading minds has just come a step closer to reality,” but the work it covers is much more grounded. Researchers collected electrode implant data of 4 participants while they were reading a set of 50 short sentences out loud. They then trained a neural network model to predict sequence of words from brain activity data while a participant is reading the words out loud.

While the model achieved low error rates for some participants, it performs much worse on sentences outside of the original 50. As it also requires brain implants and the person to actually speak the words, the research team stresses that this is not a way to read minds:

This is not translation of thought [but of brain activity involved in speech].

Using AI responsibly to fight the coronavirus pandemic

Many governments are turning to AI-enabled surveillance technology, such as facial recognition, to track the movement of people amid the Covid-19 outbreak. While the growing use of surveillance helps to curb viral spread, there are “very real and serious human rights concerns that can be damaging and undermine the trust placed in government by communities.”

This article lists several ways that AI can be deployed in ways that protect privacy and non-discrimination rights. In particular, data collected from individuals should be anonymized, the new data collection schemes for surveillance should have well-defined, limited scopes and time-horizons, and that AI use cases should be made transparent to public and civil society organizations.

It is essential that, even in times of great crisis, we remain conscience of the duality of AI and strive to advance AI for good.

Podcast

Check out our weekly podcast covering these stories! Website | RSS | iTunes | Spotify | YouTube

News

Advances & Business

• DeepMind’s Agent57 AI agent can best human players across a suite of 57 Atari games - Alphabet company’s most recent feat is Agent57, a learning agent that can beat the average human on each of 57 Atari games with a wide range of difficulty, characteristics and gameplay styles.

• Towards an ImageNet Moment for Speech-to-Text - Speech-to-text (STT), also known as automated-speech-recognition (ASR), has a long history and has made amazing progress over the past decade.

• UC Berkeley scientists spin up a robotic COVID-19 testing lab - As doctors around the country scramble to diagnose cases of COVID-19, scientists at the University of California, Berkeley’s Innovative Genomics Institute (IGI) are creating from scratch a diagnostic lab with the capability to process more than 1,000 patient samples per day.

• Physical Distancing Boosts AI-Powered Online Education in China - To help control COVID-19 transmission and spread, people across China are spending more time at home and generally reducing outdoor activities.

• Governments turn to AI in fight against coronavirus - Countries around the world - including the US, South Korea and Taiwan - are using artificial intelligence (AI) to help slow the spread of COVID-19.

• STAT’s guide to how hospitals are using AI to fight Covid-19 - The coronavirus outbreak has rapidly accelerated the nation’s slow-moving effort to incorporate artificial intelligence into medical care, as hospitals grasp onto experimental technologies to relieve an unprecedented strain on their resources.

• AI tool finds 3 coronavirus signs that often lead to a severe case - A new study shows that AI has been able to highlight three COVID-19 symptoms that are indicative of severe COVID-19 complications.

• This AI startup rents out robots that can do human jobs amid the coronavirus pandemic - As demand for warehouse work grows, perhaps robots can bear some of the brunt.

• Google taught this robotic dog to learn new tricks by imitating a real one - Google researchers are using imitation learning to teach autonomous robots how to pace, spin, and move in more agile ways.

• In artificial intelligence, enterprises still not minding their data - Data is the raw material that fuels artificial intelligence and machine learning initiatives, but it actually can’t be that raw. It needs to be as accurate, timely and well-vetted as possible – or else AI will deliver erroneous or biased results.

Concerns & Hype

• Detecting COVID-19 with 97% accuracy: beware of the AI hype! - Deep learning is not a silver bullet solution. Many unprepared companies who tried to internalize it into data squads went nuts after they saw their cost rising while little or nothing was going into production.

• Debate flares over using AI to detect Covid-19 in lung scans - A series of studies, starting as a steady drip and quickening to a deluge, has reported the same core finding amid the global spread of Covid-19: Artificial intelligence could analyze chest images to accurately detect the disease in legions of untested patients.

• I asked eight chatbots whether I had Covid-19. The answers ranged from “low” risk to “start home isolation” - U.S. hospitals, public health authorities, and digital health companies have quickly deployed online symptom checkers to screen patients for signs of Covid-19.

• AI can’t predict how a child’s life will turn out even with a ton of data - Hundreds of researchers attempted to predict children’s and families’ outcomes, using 15 years of data. None were able to do so with meaningful accuracy.

Analysis & Policy

• How AI Will Impact The Future Of Patient Care - The human body is a complex system, with an estimated 30 to 40 trillion cells and dozens of organs that all interact with each other in a network.

• AI can help with the COVID-19 crisis - but the right human input is key - Artificial intelligence (AI) has the potential to help us tackle the pressing issues raised by the COVID-19 pandemic. It is not the technology itself, though, that will make the difference but rather the knowledge and creativity of the humans who use it.

Expert Opinions & Discussion within the field

• Women in AI | Chelsea Finn: “I Certainly Feel Like a Minority” - As part of the Women in AI special project, Synced spoke with Chelsea Finn, an assistant professor in Computer Science and Electrical Engineering at Stanford University.

Explainers

• Unstoppable AI Flywheels and the Making of the New Goliaths - AI creates engines of relentless optimization at all levels. Read the article to figure out how and its consequences to what you’re doing.

Awesome Videos

That’s all for this week! If you are not subscribed and liked this, feel free to subscribe below!