1 – Computer System Passes “Visual Turing test”
A new AI system out of MIT has successfully proven its ability to make inferences and produce handwritten characters that are mostly indistinguishable from that of a human’s. The system is based on a computational structure known as a “probabilistic program”, and is the thesis work of former PhD student Brenden Lake. Unlike more traditional computer programs that break a complicated system down to its most basic parts, a probabilistic program makes inferences to fill in gaps by analyzing a large array of examples. The system was subjected to a variety of tests, all of which asked the machine to produce characters in a writing system that were different by a matter of degrees than the character to which it had been first exposed. Humans took the same three tests. Judges of all three tasks, in both machine and human, identified the machine product 50% of the time, equatable to chance.
(Read the full article on MIT News)
2 – OpenAI – Musk, Altman, Others Make Plans for New AI Center
Elon Musk and Sam Altman are co-chairing the founding of a new nonprofit research institution along with a number of prominent research engineers and scientists, advisors and investors. OpenAI aims to focus on cultivating positive outcomes for the unknown but vastly advancing potential future of human-level AI. Greg Brockton, the former CTO of Stripe, will serve as the organization’s CTO, with Ilya Sutskever, a research scientist at Google, serving as director of research. The organization pledges to make all results of its research outcomes public and keep all patents royalty-free, putting the possibilities of AI into as many hands as possible.
(Read the full release on OpenAI)
3 – Why Google’s New Quantum Computer Could Launch an Artificial Intelligence Arms Race
In collaboration with NASA, Google has announced its successful creation of the first working quantum computer – the D-Wave 2x – which is supposedly 100 million times faster than any machines known today. A computer of this power magnitude has potentially profound implications for the field of AI; machine learning problems too difficult to solve on today’s computers could, in theory, be solved in instants on a quantum computer. Google’s quantum computer is not without its potential obstacles i.e. taking into account the weird and unpredictable world of quantum particles, for example. There’s also the competition from other “unconventional” computers, such as IBM’s Watson, or the memcomputer, which mimics the human brain.
4 – Facebook Open Sources Its AI Hardware as It Races Google
On Thursday, Facebook announced its plans to open source the hardware designs used to build the computer server, nicknamed “Big Sur”, on which it runs its AI algorithms. The news comes a month after Google released the open-sourced software engine behind its deep-learning capabilities. The motivation behind the release taps into the wider pool of talent and resources of the public, which could help speed the advance of the technology. Open-source software now seems to be a landmark of today’s technological landscape.
(Read the full article on Wired)
5 – Robot Revolution: Scientists Launch AI Advocacy Group to Tackle Urgent Ethical Issues
A new advocacy group for ethical robot use was announced today in London. The Foundation for Ethical Robotics, which consists of 22 scientists, engineers, and law scholars, is inviting people from all fields who have an interest in robotics to join the organization and become an active participant and thinker in this relatively new field of interest. Founder Dr. Aimee van Wynsberghe, an assistant professor in Ethics of Technology at University of Twente’s Department of Philosophy, and Robotics Professor Noel Sharkey from the University of Sheffield founded the organization in part due to the seeming rush towards robotics in the last few years and the apparent gaps in ethical considerations. In a statement to IBT, van Wynsberghe said,
“We urgently need to promote responsibility for the robots embedded in our society. Robots are only as responsible as the humans who build and use them. We must ensure that the future practice of robotics is for the benefit of mankind rather than for short term gains”.
(Read the full article on International Business Times UK)