In case you missed DeepMind’s PathNet from last month, we are sharing with you a helpful summary of the key take-home points. PathNet uses Transfer Learning to play games (like Pong) which it was not trained on. This is an exciting advancement since it is on the road to general AI (btw, don’t forget about the General AI Challenge, which is offering a total of $5M in prizes).
We wanted to ask you: do you think that DeepMind will be the first to create General AI? Click ‘Yes’ or ‘No’ below to express your opinion and please comment as well.
Last month, DeepMind published an important paper
describing how general AI potentially will look like. Since the paper is a fairly-difficult read for the uninitiated, we thought it would be useful to share a summary of Théo Szymkowiak, president of McGill University’s AI Society. Therein, he describes PathNet’s ability to generalize via Transfer Learning to play games that it was not trained on.
Apple has bought Israeli startup RealFace, a cybersecurity and machine learning firm specializing in facial recognition technology. The Times of Israel reported on Sunday that the Tel Aviv-based company, founded in 2014, was snapped up by Apple for an estimated $2 million. RealFace’s website is currently offline, but according to promotional material, the startup had developed a unique facial recognition technology that integrates AI and “brings back human perception to digital processes”.
RealFace is the fourth Israel-based firm Apple is known to have acquired. In 2011 it bought flash memory maker Anobit for a reported $400 million, then in November 2013 it acquired 3D sensor company PrimeSense for an estimated $345 million. Most recently in 2015, Apple bought LinX for around $20 million.
An AI watchdog should be set up to make sure people are not discriminated against by the automated computer systems making important decisions about their lives, say experts.
In a new report, Sandra Wachter, Brent Mittelstadt, and Luciano Floridi, a research team at the Alan Turing Institute in London and the University of Oxford, call for a trusted third party body that can investigate AI decisions for people who believe they have been discriminated against. “What we’d like to see is a trusted third party, perhaps a regulatory or supervisory body, that would have the power to scrutinise and audit algorithms, so they could go in and see whether the system is actually transparent and fair,” said Wachter.
Researchers have developed an algorithm to accurately predict which patients diagnosed with acute myelogenous leukemia (AML), a cancer of the blood and bone marrow, will go into remission following treatment and which ones will relapse.
Using bone marrow data and medical histories of AML patients, as well as blood data from healthy individuals, researchers were able to teach a standard 64-bit computer workstation running Windows to predict remission with 100 percent accuracy, while relapse was
correctly predicted in 90 percent of relevant cases.
“It’s pretty straightforward to teach a computer to recognize AML, once you develop a robust algorithm, and in previous work we did it with almost 100 percent accuracy,” said Murat Dundar, associate professor of computer science in the School of Science at Indiana University-Purdue University Indianapolis.
Waya.ai recently open sourced the core components of its skin cancer diagnosis software and made the data sets it collected for this task easily available. The goal is to release a free & open source product in early May that diagnoses skin cancer with (validated) dermatologist-level accuracy (or better).
This tutorial dives into Waya.ai’s simple code that they have open-sourced that serves as a general starting point for image classification and other tasks in computer vision, and is used to train their production skin-cancer diagnosis models. It will allow you to get up-and-running with this code, training a deep convolutional neural network to diagnose skin-cancer w/ dermatologist-level accuracy!
Predicting colour is easy: Shine a light with a wavelength of 510 nanometers, and most people will say it looks green. Yet figuring out exactly how a particular molecule will smell is much tougher. Now, 22 teams of computer scientists have unveiled a set of algorithms able to predict the odor of different molecules based on their chemical structure. It remains to be seen how broadly useful such programs will be, but one hope is that such algorithms may help fragrance makers and food producers design new odorants with precisely tailored scents.
This latest smell prediction effort began with a recent study by olfactory researcher Leslie Vosshall and colleagues at The Rockefeller University in New York City, in which 49 volunteers rated the smell of 476 vials of pure odorants. For each one, the volunteers labeled the smell with one of 19 descriptors, including “fish,” “garlic,” “sweet,” or “burnt.” They also rated each odor’s pleasantness and intensity, creating a massive database of more than 1 million data points for all the odorant molecules in their study.
General Motors Co plans to deploy thousands of self-driving electric cars in test fleets in partnership with ride-sharing affiliate Lyft Inc, beginning in 2018, two sources familiar with the automaker’s plans said last week.
It is expected to be the largest such test of fully autonomous vehicles by any major automaker before 2020, when several companies have said they plan to begin building and deploying such vehicles in higher volumes. Alphabet Inc’s Waymo subsidiary, in comparison, is currently testing about 60 self-driving prototypes in four states.
Most of the specially equipped versions of the Chevrolet Bolt electric vehicle will be used by San Francisco-based Lyft, which will test them in its ride-sharing fleet in several states, one of the sources said. GM has no immediate plans to sell the Bolt AV to individual customers, according to the source.
Seattle startup ReplyYes has sold more than 100,000 vinyl records purely through text messaging. And now the company, which spun out of Madrona Venture Labs in December 2015, has just inked a deal for Universal Music Group to use ReplyYes’ technology to sell its own albums and merchandise.
ReplyYes is an e-commerce company with an AI focus. It runs a chat bot that suggests products to customers via text and Facebook Messenger and learns about customer preferences over time. The startup has been running its own two stores called The Edit and Origin Bound, which sell vinyl records and comic books, respectively, since last spring.
The Edit, for instance, sends a text message to customers each day suggesting a record to buy. If the customer wants it, they can reply “yes.” They can also
indicate “like” and “dislike” to help the technology learn what types of music they like.