GPUs promise a revolution in unsupervised learning

SAN FRANCISCO - March 16, 2015 - PRLog -- Today’s bleeding-edge unsupervised learning systems rely on huge amounts of unstructured data to learn complex, nonlinear models with millions of free parameters. While the earliest of these systems used hundreds or even thousands of CPU cores by connecting multi-core CPU machines together, most companies do not have hundreds or thousands of machines at their disposal. An alternate and more affordable solution is to migrate these parallel processing tasks onto GPU processors, which can support this magnitude of processing power on even a single card.

A prime venue for learning about today’s most advanced GPU-based systems is NVIDIA’s GPU Technology Conference (GTC) in Silicon Valley, which brings together GPU developers, game developers, data scientists, and machine learning experts to talk shop and show off their wares. This year the conference highlights the growing revolution in deep learning, and the pivotal role in new advances played by GPU systems.

For this year’s GTC, NVIDIA chose Loop AI Labs to participate in the Emerging Company Summit, where a few startups from around the world will feature their disruptive innovations that will fuel the coming revolution in machine intelligence.

Loop AI Labs' machine intelligence technology is revolutionary because it can autonomously read any unstructured data set and learn from scratch the concepts and relationships within -- much like humans do -- without human supervision or programming.

At GTC, Loop AI Labs will be demonstrating this unsupervised semantic understanding technology, powered by deep learning on GPUs.

“I am often asked if deep learning really has anything useful to offer the field of natural language understanding,” remarks Patrick Ehlen, Head of Deep Learning at Loop AI Labs. “And it certainly does. But with a few exceptions, most of what you see out there is not very ambitious, and was probably done on somebody’s MacBook. To get good results on the really hard problems, like creating semantic understanding systems that will understand our complicated world, we need to pull out the big guns, and GPUs are a great way to do that.”

Loop AI Labs is working to radically change how machines can autonomously learn and understand the human world.

“Organizations today analyze their data using machine learning on big data, but require human brains to synthesize those results into decisions about how the data should impact future decisions. What’s wrong with that picture? Humans still need to do a great deal of work while their computers sit idle, squandering the full potential of humans and machines alike,” says Gianmauro Calafiore, CEO and Founder of Loop AI Labs.

“Whether we want to find cures for diseases or simply figure out government budgeting, we see a pervasive problem where data is underutilized, it’s laying there, there’s value there, but it’s not getting used like it could be. AI as a Service (AIaaS) platforms like ours offer the solution to this problem, and to the dearth of skilled data scientists, enabling organizations of any size to leverage advanced artificial intelligence at a fraction of the cost using their own data or the huge number of data sets that are publicly available.”

Loop AI will be participating in the NVIDIA GPU Technology Conference at the Emerging Companies section, Booth 1234, and demoing the latest generation of its machine intelligence platform.

http://www.gputechconf.com/

http://loop.ai

About Loop AI Labs

Loop AI Labs is working to radically change how machines can autonomously learn and understand the human world.

The company’s machine intelligence platform can autonomously read any unstructured data set and learn from scratch the concepts and relationships within -- much like humans do -- without human supervision or programming. The platform can understand concepts in any domain and any language, without prior knowledge about the syntactic or semantic structure of the language. The company’s AI-as-a-Service platform, which combines unsupervised deep learning and classical symbolic reasoning, will benefit organizations that are struggling to analyze and understand data sets of their own making or from public sources. Loop AI Labs intends to help major sectors of the economy -- such as healthcare, finance, insurance, telecommunications, and retail -- to benefit from the efficiencies of a new era of machine intelligence, and to make people’s lives easier, safer, and more productive.