Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog

Category: education – Page 175

An old-fashioned AI has won a Starcraft shootout

En Taro AI

The latest results in a long-running contest of video-game-playing AIs reveal how hard it is for machines to master swarming insectoid Zergs or blitzing Protos. They also show that even old-school approaches can still sometimes win out.

The AIIDE Starcraft Contest has been running at Memorial University in Newfoundland, Canada, since 2010. Participating teams submit bots that play an original version of Starcraft, a sprawling sci-fi-themed game, in a series of one-on-one showdowns.

Starcraftiness: Video games are generally useful in AI because they offers a constrained environment and a good way to quantify progress. The popular online strategy game Starcraft has emerged as an important benchmark for AI both because it is extremely complicated and because it’s a game where it’s hard to measure progress. There are a vast number of possible states and a huge number of potential moves at every moment. And it can be hard to tell if a strategy is a good one until much later on in a battle.

By solving a mystery of gene repair, scientists uncover an exception to biology’s rules

👀

About 15 years ago, UNC Lineberger’s Dale Ramsden, Ph.D., was looking through a textbook with one of his students when they stumbled upon a scientific mystery.

A small line in the book indicated that a protein that helps major breaks in our did so by adding DNA, or deoxyribonucleic acid, as expected. However, there were hints that it could also add RNA, or ribonucleic acid, at least in a test tube. It seemed unlikely that this would occur during repair of DNA in living , since RNA is normally used only as a messenger to carry information from the genetic code to make proteins.

“You would think they must only add DNA during repair of our genetic code, because that’s the core of the central dogma of life; genetic information has to be DNA all the time,” said Ramsden, who is a professor in the UNC School of Medicine Department of Biochemistry and Biophysics. “That’s the way it’s supposed to be. That’s what we’re taught in school.”

At Last, Physicists Understand Where Matter’s Mass Comes From

Lattice QCD is not only teaching us how the strong interactions lead to the overwhelming majority of the mass of normal matter in our Universe, but holds the potential to teach us about all sorts of other phenomena, from nuclear reactions to dark matter.

Later today, November 7th, physics professor Phiala Shanahan will be delivering a public lecture from Perimeter Institute, and we&s;ll be live-blogging it right here at 7 PM ET / 4 PM PT. You can watch the talk right here, and follow along with my commentary below. Shanahan is an expert in theoretical nuclear and particle physics and specializes in supercomputer work involving QCD, and I&s;m so curious what else she has to say.

AI Guru Andrew Ng on the Job Market of Tomorrow

…but Our Timelines Are Too Rosy I would actually welcome a correction in public opinion about what AI can and cannot do. This has happened to me multiple times, where I would listen to a CEO on stage make an announcement about what their company is doing with AI, and then 20 minutes later I’d talk to one of their engineers, and they’d say, “No, we’re not doing that, and we have no idea how to do it.” I think it still takes judgment to know what is and what isn’t possible with AI, and when the C-suite does not yet have that judgment it’s possible for companies to make promises very publicly that are just not feasible. Frankly, we see some of this in the self-driving space. Multiple auto [original equipment manufacturer] CEOs have promised self-driving car roadmaps that their own engineers think are unrealistic. I feel [CEOs are] being sincere but just not really understanding what can be done in a certain timeframe.

The co-founder of Google’s deep-learning research team on the promise of a conditional basic income, the need for a skills-based education system and what CEOs don’t understand about artificial intelligence.

190 universities just launched 600 free online courses. Here’s the full list

If you haven’t heard, universities around the world are offering their courses online for free (or at least partially free). These courses are collectively called MOOCs or Massive Open Online Courses.

In the past six years or so, over 800 universities have created more than 10,000 of these MOOCs. And I’ve been keeping track of these MOOCs the entire time over at Class Central, ever since they rose to prominence.

In the past four months alone, 190 universities have announced 600 such free online courses. I’ve compiled a list of them and categorized them according to the following subjects: Computer Science, Mathematics, Programming, Data Science, Humanities, Social Sciences, Education & Teaching, Health & Medicine, Business, Personal Development, Engineering, Art & Design, and finally Science.

Look to Africa to advance artificial intelligence

That will require widening of the locations where AI is done. The vast majority of experts are in North America, Europe and Asia. Africa, in particular, is barely represented. Such lack of diversity can entrench unintended algorithmic biases and build discrimination into AI products. And that’s not the only gap: fewer African AI researchers and engineers means fewer opportunities to use AI to improve the lives of Africans. The research community is also missing out on talented individuals simply because they have not received the right education.

If AI is to improve lives and reduce inequalities, we must build expertise beyond the present-day centres of innovation, says Moustapha Cisse.

New schemes teach the masses to build AI

That is changing. This month fast.ai, an education non-profit based in San Francisco, kicked off the third year of its course in deep learning. Since its inception it has attracted more than 100,000 students, scattered around the globe from India to Nigeria. The course and others like it come with a simple proposition: there is no need to spend years obtaining a phd in order to practise deep learning. Creating software that learns can be taught as a craft, not as a high intellectual pursuit to be undertaken only in an ivory tower. Fast.ai’s course can be completed in just seven weeks.

Treating it like a craft is paying dividends.