News 

Neural Magic gets $15M seed to run machine learning models on commodity CPUs

“Neural Magic, a startup founded by a couple of MIT professors, who figured out a way to run machine learning models on commodity CPUs, announced a $15 million seed investment today” writes Ron Miller for yahoo.com. That means that it could greatly reduce the cost associated with machine learning projects by allowing data scientists to use commodity hardware.Gil Beyda, managing director at lead investor Comcast Ventures, sees a huge market opportunity with an approach that lets people use commodity hardware.”I discovered that with the right algorithms we could run these…

Read More
News 

How private companies work to gain access to our brains

“But recent advances in brain-computer interface (BCI) technology, which integrates cognitive activity with a computer, might challenge this” writes Oscar Schwartz for theguardian.com. State-of-the-art BCIs can only decode the neural signals associated with attempted speech, or the physical act of articulation, Slutzky told me.While the management of this collateral data is heavily regulated in research institutes Haynes told me that no such regulations are in place for technology companies.Because neural signals in the brain are often noisy, decoding is extremely difficult.Using only data collected from neural activity, the algorithmic systems…

Read More
News 

Apple Says A13 Bionic Chip Was Designed With Performance-Per-Watt Focus

“But the reality is, we view it as performance per watt” writes Am Pdt for macrumors.com. “We talk about performance a lot publicly.”There wasn’t machine learning running ten years ago.”For applications that don’t need the additional performance, you can run at the performance of last year’s and just do it at a much lower power,” Shimpi said.Machine learning also plays a big role in the A13 chip, helping to manage battery life and optimize performance, according to Schiller. Source: macrumors.com

Read More
News 

How an engineer’s accident at Google changed the art industry

“The vibrant neural network art world arose in the past few years, in part, from developments in computer science” writes Associated Press for marketwatch.com. But, like previous groundbreaking art movements, neural network art raises difficult questions: How do we think of authorship and ownership when these artworks come from the contributions of so many different creative individuals and algorithms?. It has led to the creation of major software tools, like Linux and major neural network software, that could not have been developed otherwise.There’s an active creative community of neural network…

Read More
News 

OpenAI and Google detail activation atlases, a technique for visualizing AI decision-making

“One portion of the InceptionV1 atlas contains what looks to be animal heads; zooming in reveals eyes, fur, and noses” writes Kyle Wiggers for venturebeat.com. By creating what the researchers call a “class activation atlas,” which involves further dimensionality reduction, it’s possible to isolate and analyze the object detectors within each layer, they said. Source: venturebeat.com

Read More
News 

Beyond Fake News, Artificial Intelligence Finally Understands the World

“OpenAI decided that the GPT-2 neural network was dangerous and chose to make only a simplified version of the model available to the public” writes Tim O’hear for impactia.org. But GPT-2 is much more than that, as it is a “Transformer” type neural network.Working towards AI for humanity, the OpenAI Foundation has published the results of its latest neural network, elegantly named “GPT-2”, which is capable of generating prose of amazing consistency.This model is as much a grammatical model of language as it is a model of the nature of…

Read More
News 

AI detects potentially damaging ice on wind turbines

“Over time, ice shedding from blades can damage other blades or overstress internal components, necessitating costly repairs” writes Kyle Wiggers for venturebeat.com. In a set of simulations conducted on data from Goldwind, one of the largest wind turbine manufacturers in China, WaveletFCNN had a prediction accuracy of 81.82 percent, compared with the original FCNN classifier’s 65.91 percent. Source: venturebeat.com

Read More