Machine Learning

Google Assistant will be smarter about names and contexts

Google Assistant will be smarter about names and contexts

Smart assistants sound almost magical in what they're able to do at our command but they are all powered by computers and code behind the scenes. Unfortunately, computers can actually be quite dumb sometimes, especially with things we take for granted because they come so naturally to us humans. That's especially true with the names we give to our friends and acquaintances. Unsurprisingly, that has also been one of Google Assistant's weaknesses but Google will soon let you fix that by teaching it how those names really sound.

Continue Reading

Alphabet’s X moonshot division wants to bring AI to the electric grid

Alphabet’s X moonshot division wants to bring AI to the electric grid

Google parent Alphabet has been working on "a moonshot" for the electric grid, with a secret project in its X R&D division aiming to figure out how to make power use more stable, and more green, than it is today. The research, revealed at the White House Leaders Summit on Climate, has been underway for the past three years.

Continue Reading

Google Play Store app install optimization rolls out

Google Play Store app install optimization rolls out

Google has always been banking on AI and machine learning to help improve its process in lieu of manpower. Machine learning, however, always requires more data to learn from, and there is perhaps no better source of millions of pieces of data than the millions of Android users out there. One such example is in optimizing how apps are downloaded and installed, a federated learning feature that Google has now started rolling out to Android users.

Continue Reading

Google Lens OCR lands on the desktop via Google Photos

Google Lens OCR lands on the desktop via Google Photos

Google Lens has long been a handy tool that demonstrates the strength of Google's various computer vision and AI capabilities. It has, however, mostly been limited to its smartphone incarnation where it admittedly has the most opportunities to be used. Google Lens, however, also has some features that can be useful even when not used in real-time, like when looking at images already stored on your computer. Thankfully, Google is slowly bringing those features to the desktop via Google Photos on a desktop web browser.

Continue Reading

Google Assistant to improve its hearing through federated learning

Google Assistant to improve its hearing through federated learning

Smart assistants like Siri have once been the butt of jokes when it comes to mishearing their activation hotwords. Users eventually realized, however, how those mistakes can cost them, especially when it came to potential privacy problems and embarrassing situations. The companies behind these smart assistants are always trying to improve how they properly detect hotwords and Google is now trying to practically crowdsource the data Google Assistant needs to learn from but without going through the potentially messy and controversial route of the cloud.

Continue Reading

New virtual unfolding tech unlocks antique letters without damage

New virtual unfolding tech unlocks antique letters without damage

When King Tut was removed from his tomb in 1922, his body was effectively ripped to pieces in order to separate his jewelry for transport and/or display. Here in modernity, scientists and archeologists take a much more conservative approach to preserving materials for optimum study. Since the dawn of time, ancient folded or "locked" documents, like those part of a study published this week, have been cut to bits in order to read the writings they contained With what's been developed by researchers and presented here in 2021, antique document destruction-by-default may become a thing of the past.

Continue Reading

High-tech threads sense how and when the wearer moves

High-tech threads sense how and when the wearer moves

Tufts University engineers have developed a flexible thread-based sensor that can measure neck movement and provide data on direction, angle of rotation, and degree of displacement of the head. Researchers say the breakthrough has the potential to be used as a thin tattoo-like patch that could measure athletic performance, monitor worker or driver fatigue, assist with physical therapy, enhance VR games, and improve computer-generated imagery for cinematographers. The thread-based sensors can be woven directly into textiles.

Continue Reading

Bluebot robots demo surreal “Blueswarm” like a school of fish

Bluebot robots demo surreal “Blueswarm” like a school of fish

A swarm robotic fish donned the cover of Science Robotics this week courtesy of researchers from Harvard University. Florian Berlinger, Melvin Gauchi, and Radhika Nagpal delivered a paper on "Implicit coordination for 3D underwater collective behaviors in a fish-inspired robot swarm." These little robots are very, very cute, too.

Continue Reading

DeepMind MuZero AI can master games without knowing the rules

DeepMind MuZero AI can master games without knowing the rules

The holy grail of AI has always been to enable computers to learn the way humans do. The most powerful AIs today, however, still rely on having certain known rules, like rules for a game of chess or Go. Human learning, however, is often messy in inferential, learning the rules of life as we go. DeepMind has long been trying to create such AIs using games as their environment and test suite. Google's sister company focusing on AI research has just revealed its latest achievement in MuZero, an AI that can master a game without learning the rules beforehand.

Continue Reading

DeepMind AI MuZero can learn AND master Chess, Go, Atari

DeepMind AI MuZero can learn AND master Chess, Go, Atari

Today the machine learning algorithm MuZero was detailed in a feature research paper in Nature. MuZero expands on the abilities of systems like AlphaGo, AlphaGo Zero, and AlphaZero. Each new algorithm allowed a smart machine to become better at mastering games, starting with Go, then Chess and Shogi, now Atari!

Continue Reading

Google’s Blob Opera experiment lets anyone create dramatic music

Google’s Blob Opera experiment lets anyone create dramatic music

If you've dreamed about creating your own opera masterpiece, Google has your back. The company has introduced a new experiment under its Arts & Culture non-profit initiative called 'Blob Opera.' The name sounds like nonsense until you learn exactly what Blob Opera is -- a tool featuring colorful blobs that, you know, sing opera music.

Continue Reading

NVIDIA Research AI can learn new things with fewer training data

NVIDIA Research AI can learn new things with fewer training data

AI and machine learning have become hot topics in tech and even mainstream news these days but most take for granted the work these computers do behind the scenes. Their amazing feats require poring through tens of thousands of related data at a speed only computers are capable of. That, in turn, implies that there are tens of thousands of data for the AI to learn, which may not always be the case. NVIDIA's Research arm is now boasting of a milestone in its Generative Adversarial Network or GAN that allows an AI to learn even when presented with a significantly smaller data set.

Continue Reading

1 2 3 4 Next