Profile cover photo
Profile photo
Wayne Radinsky
18,612 followers -
Software Design Engineer
Software Design Engineer

18,612 followers
About
Posts

Post has attachment
"Julian Gough's novel Connect is a science fiction thriller set in a world on the verge of a technological breakthrough that could change the world forever." Asked, "We're moving into an age of ubiquitous surveillance and artificial intelligence. Do you think we've got a chance of escaping the kinds of dangers hinted at in Connect?" he said, "If you live in China: no. The war has already been lost there. They are well on the way to a total surveillance state that will persist until the dysfunctional Chinese economy at some point collapses, and discredits the Communist Party. But the thing is, a lot of people welcome such surveillance."

"The problems in Europe and America are slightly different. Jaron Lanier's new book, Ten Arguments for Deleting Your Social Media Accounts Right Now, is the best guide to the problems we face in the West. The problem isn't simple surveillance any more; it's that that surveillance is then being used for behaviour manipulation. We are being algorithmically sorted and controlled, and it's often not even by consciously evil manipulators, which would at least give it dignity; it's by dumb, corporate-owned AIs, tweaking algorithms to make us more anxious, so we will click on stuff that makes us more anxious, ..."
Add a comment...

Post has attachment
"How to use trained Keras and TensorFlow machine learning models within Ruby on Rails." I've never used Ruby on Rails, so I'm just passing this along for those of you who do. Basically you export the model and put it on Google Cloud. From Ruby, since Google's API client library for Ruby is still in alpha, you invoke a Python script.
Add a comment...

Post has attachment
"List of machine learning / deep learning conferences in 2018 / 2019." There's so many someone had to make a list.
Add a comment...

Post has attachment
"Deploying Keras deep Learning models with Java." "One approach that's often used is converting Keras models to TensorFlow graphs, and then using these graphs in other runtines that support TensorFlow. I recently discovered the Deeplearning4J (DL4J) project, which natively supports Keras models, making it easy to get up and running with deep learning in Java."
Add a comment...

Post has attachment
"Eager execution will be a central feature of 2.0. It aligns users' expectations about the programming model better with TensorFlow practice and should make TensorFlow easier to learn and apply. Support for more platforms and languages, and improved compatibility and parity between these components via standardization on exchange formats and alignment of APIs. We will remove deprecated APIs and reduce the amount of duplication, which has caused confusion for users."
Add a comment...

Post has attachment
DeepMind's Neural Arithmetic Logic Units (NALU) in TensorFlow.
Add a comment...

Post has attachment
7 planets in the TRAPPIST-1 system are rocky, and two of them are in the habitable zone. The TRAPPIST-1 system is "only" 40 light-years from here.
Add a comment...

Post has attachment
"The data is 3D scans of patients' eyes made using a technique known as optical coherence tomography, or OCT. Creating these scans takes around 10 minutes and involves bouncing near-infrared light off of the interior surfaces of the eye."

"The software was trained on nearly 15,000 OCT scans from some 7,500 patients. These individuals had all been treated at sites operated by Moorfields, which is the largest eye hospital in Europe and North America. The system was fed their scans alongside diagnoses by human doctors. From this, it learned how to first identify the different anatomical elements of the eye (a process known as segmentation) and then recommend clinical action based on the various signs of diseases that the scans show."
Add a comment...

Post has attachment
From here to human-level artificial general intelligence in four (not all that) simple steps, according to Ben Goertzel. "Just as we have many different workable approaches to manned flight -- airplanes, helicopters, blimps, rockets, etc. -- there may be many viable paths to AGI, some of which are more biologically inspired than others. And, somewhat like the Wright brothers, today's AGI pioneers are proceeding largely via experiment and intuition, in part because we don't yet know enough useful theoretical laws of general intelligence to proceed with AGI engineering in a mainly theory-guided way; the theory of AGI is evolving organically alongside the practice."

Make cognitive synergy practical (use AI systems together "in sufficiently sophisticated ways"), bridge symbolic and subsymbolic AI (bridging algorithms used for low-level intelligence, such as perception and movement with algorithms used for high-level abstract reasoning), whole-organism architecture (AI systems in physical systems capable of interacting with the everyday human world in nuanced ways), and scalable meta-learning (AGI needs not just learning but also learning how to learn).
Add a comment...

Post has attachment
"In one hand, he held a bug-eyed toy robot, its lines voiced by the program. With the other, he pretended to grip a steering wheel. The robot was playing Piotr Mirowski's partner; they were taking a scenic drive together."

"'I am not trying to be angry,' the robot said, suddenly breaking the mood."

"'I don't want you to be angry -- this is our quality time,' Mr. Mirowski replied."

"'I'm sure that you will find love,' the robot said after an awkward pause, drawing a firm end to the couple's relationship, and prompting laughter from the audience. 'I'm so tired,' it added. Mr. Mirowski made a final attempt to save things, but the robot would not listen. 'You are not me. You're my friend,' it said, emotionless."
Add a comment...
Wait while more posts are being loaded