How PayPal uses deep learning and detective work to fight fraud
New crime needs new algorithms
Hui Wang has seen the nature of online fraud change a lot in the 11 years she’s been at PayPal. In fact, a continuous…
Deep learning
Advertisement
IBM acquires deep learning startup AlchemyAPI
Watson looks a lot smarter
So much for AlchemyAPI CEO Elliot Turner’s statement that his company is not for sale. IBM has bought the Denver-based deep learning…
Google, Stanford say big data is key to deep learning for drug discovery
Bigger data, better model
A team of researchers from Stanford University and Google have released a paper highlighting a deep learning approach they say shows promise in the…
AWS suits up more enterprise perks
The week in cloud
More AWS perks for business users Amazon Web Services has beefed up its identity management and access control capabilities so that businesses…
Advertisement
Why deep learning is at least inspired by biology, if not the brain
A neuroscientist explains
As deep learning continues gathering steam among researchers, entrepreneurs and the press, there’s a loud-and-getting-louder debate about whether its algorithms actually operate…
Microsoft says its new computer vision system can outperform humans
Deep learning FTW, again
Microsoft researchers claim in a recently published paper that they have developed the first computer system capable of outperforming humans on a…
Advertisement
No, you don’t need a ton of data to do deep learning
Structure Show
There are a couple of seemingly contradictory memes rolling around the deep learning field. One is that you need a truly epic amount of…
Gigaom’s new AI and deep learning conference: Structure Intelligence
Kicks off September 22
We are thrilled to announce the launch of Gigaom’s newest conference, Structure Intelligence. In the past year we’ve seen massive growth in…
Advertisement
PhotoTime is a deep learning application for the rest of us
Nicely polished neural nets
A Sunnyvale, California, startup called Orbeus has developed what could be the best application yet for letting everyday consumers benefit from advances…
TeraDeep wants to bring deep learning to your dumb devices
Turning webcams into Dropcams
Open the closet of any gadget geek or computer nerd, and you’re likely to find a lot of skeletons. Stacked deep in…
Advertisement
New to deep learning? Here are 4 easy lessons from Google
Do you have data? Get some
Google employs some of the world’s smartest researchers in deep learning and artificial intelligence, so it’s not a bad idea to listen to…
Hilary Mason on taking big data from theory to reality
The Structure Show
If you’re interested in assessing how and when a given data technology — deep learning, machine intelligence, natural language generation — can…
Here are the winners of the 2015 Structure Data Awards
Highlighting the best of 2014
The second-annual Structure Data Awards are here, where Gigaom picks the most-interesting and most-promising data startups that launched in the previous year.…
Facebook open sources tools for bigger, faster deep learning models
Sowing the seeds of better AI
Facebook on Friday open sourced a handful of software libraries that it claims will help users build bigger, faster deep learning models…
Baidu built a supercomputer for deep learning
Mine sees better than yours
Chinese search engine company Baidu says it has built the world’s most-accurate computer vision system, dubbed Deep Image, which runs on a supercomputer…
Artificial intelligence is real now and it’s just getting started
Keep up if you can
Artificial intelligence is already very real. Not conscious machines, omnipotent machines or even reasoning machines (yet), but statistical machines that automate and increasingly…
Machine learning will eventually solve your JPEG problem
You're welcome, Ansel
I take a lot of photos on my smartphone. So many, in fact, that my wife calls me Cellphone Ansel Adams. I can’t…
The 5 stories that defined the big data market in 2014
A banner year for big data
There is no other way to put it: 2014 was a huge year for the big data market. It seems years of talk…
The year in tech: Net neutrality, IoT grows up, Uber turns heads
What we cared about in 2014
As 2014 draws to a close, the tech world seems a little weary. It was a draining year if you were plugged…
Deep learning now tackling autism and matching monkeys’ vision
Maybe start believing the hype
Two studies published this week provide even more evidence that deep learning models are very good at computer vision and might be able…
Big-data-based food startup Hampton Creek raises $90M
Scaling beyond Just Mayo
Hampton Creek, a San Francisco startup that uses advanced data analysis to develop eco-conscious, egg-free food products, has raised a $90 million series…
Baidu claims deep learning breakthrough with Deep Speech
Filtering out the noise
Chinese search engine giant Baidu says it has developed a speech recognition system, called Deep Speech, the likes of which has never…
IBM bringing its skin-cancer computer vision system to hospitals
Machine learning for melanoma
IBM says it has developed a machine learning system that identified images of skin cancer with better than 95 percent accuracy in…
Skype can translate spoken Spanish to English in near real-time
Se Habla Español
Microsoft started rolling out a new feature for Skype on Monday called Skype Translator which will translate communications from users using different languages in near real-time — that is, as you’re chatting on Skype.
Deep learning might help you get an ultrasound at Walgreens
A new startup called Butterfly Network, from genomic-technology pioneer Jonathan Rothberg, hopes to improve the world of medical imaging using advanced chip technologies, tablet devices and deep learning. Rothberg explains how and why deep learning is key to the company’s plans.
On Reddit, Geoff Hinton talks Google and future of deep learning
University of Toronto researcher and part-time Google distinguished researcher Geoff Hinton is responsible for many recent advances in deep learning, and many advances in neural network research over the past few decades. Here are some highlight of a recent Reddit AMA with Hinton.
Applications are driving investment in deep learning startups
New funding for deep learning startups focused on medical imaging suggests investors see more promise in the applications of deep learning than for the models themselves.
It doesn’t matter if deep learning mimics the brain or Watson is cognitive. It matters if they work
Recent comments by machine learning experts have caused a stir, but debate over the novelty or architecture of deep learning might be best left in academia . As AI techniques make their way into developers’ hands, whether they catch on depends on whether they’re useful.
Google buys two Oxford University AI spin-offs to bolster its DeepMind team
The Google-owned artificial intelligence outfit DeepMind has revealed a partnership with Oxford University that involves a “substantial donation” to set up a…
Researchers are using deep learning to predict how we pose. It’s more important than it sounds
The science of identifying the positions of arms, legs and joints within images is becoming a popular pastime among deep learning researchers. The techniques they develop could be particularly important in fields such as human-computer interaction and computer animation.
GraphLab’s machine learning platform now supports deep learning
Machine learning startup GraphLab is boosting the capabilities of its platform, which aims to help users build and then serve their predictive models, has released a new version of its software that includes support for deep learning models.
AlchemyAPI now recognizes famous faces (and can learn yours, too)
AlchemyAPI’s deep-learning-based system can now be used to tag famous people in photos, or to recognize non-celebrities via labeled images in corporate or social networks. It’s the latest — but not the last — service the company has released since getting started analyzing text in 2012.
Here’s what you missed at Gigaom’s ‘Future of AI’ meetup
Gigaom recently hosted a meetup featuring thought leaders and entrepreneurs in the world of artificial intelligence and deep learning. Watch the videos of their presentations here.
Microsoft Azure now uses speech recognition to help users search videos
Microsoft Azure now provides a speech-recognition service for audio-visual content that indexes the files based on what’s actually said in them. This could automate the searchability, categorization and description of content that used to be a mystery if it wasn’t properly labeled.
New research suggests deep learning could improve AI in video games
North Carolina State researchers built a deep learning system that was able to predict players’ goals in an open-ended video game nearly 63 percent of time. It’s not perfect and it’s not proof of a gaming revolution, but it’s a big improvement and a good start.
Nvidia stakes its claim in deep learning by making its GPUs easier to program
Nvidia is fully embracing the effectiveness of GPUs for running deep learning algorithms, releasing over the weekend a new set of libraries designed to let researchers experience the performance boost of GPUs without having to optimize their models for the hardware.
Google explains its latest gains in computer vision
The annual ImageNet computer vision competition took place in August, and once again the winning techniques blew away past winners. John Markoff…
Baidu says its massive deep-learning system is nearly complete
Baidu says its 100-billion-neuron deep learning system will be complete within six months, powering a fast transition away from text as the dominant search input. Thanks to smartphones and its new Baidu Eye technology, the company expects voice and image search to dominate within five years.
Nervana Systems raises $3.3M to build hardware designed for deep learning
A San Diego-based startup called Nervana Systems has raised a series A round for its specialized deep learning computing system. It’s a smart move given the hype and legitimate promise of the space, but the best path for commercializing the tech is an open question.
Researchers hope deep learning algorithms can run on FPGAs and supercomputers
The NSF has funded projects that will investigate how deep learning algorithms run on FPGAs and across systems using the high-performance RDMA interconnect. Another project, led by Andrew Ng and two supercomputing experts, wants to put the models on supercomputers and give them a Python interface.
Gigaom meetup: Talk, learn and hear about the future of AI and deep learning
Join Gigaom for a free meetup on artificial intelligence and machine learning. Speakers from Google, Microsoft Research, Baidu and more will discuss the state of the art in fields such as deep learning, and how things will change when they go mainstream.
How Spotify is working on deep learning to improve playlists
A Spotify intern and Ph.D. student published a blog post laying out his work to improve Spotify’s recommendation algorithms using deep learning to analyze the acoustic properties of songs. He hopes his models can help listeners discover new and relatively unheard music.
Startup claims it’s revolutionizing personalization with deep learning
A Cambridge, Massachusetts, startup called Nara has released a service the company claims use “deep learning artificial intelligence” to improve online personalization. Essentially, the…
Researchers say deep learning can help discover Higgs bosons and other particles
Scientists have researched the effectiveness of deep learning techniques for discovering exotic particles and found some significant improvements over previous methods. They believe deep learning could help analyze data from the Large Hadron Collider.
Microsoft to provide drag-and-drop machine learning on Azure
Microsoft, which recently showed off its machine learning research with Skype Translate, is opening up those capabilities with a new cloud service called Azure Machine Learning.
More deep learning for the masses, courtesy of Ersatz Labs
A startup called Ersatz Labs is promising deep learning, delivered via appliance or the cloud, usable by pretty much anybody already familiar with machine learning. It’s the latest attempt to take the data-processing approach out of the lab and into the mainstream.
A startup called Skymind launches, pushing open source deep learning
Skymind is providing commercial support and services for an open source project called deeplearning4j. It’s a collection of of approaches to deep learning that mimic those developed by leading researchers, but tuned for enterprise adoption.
Skype will soon get real-time speech translation based on deep learning
It’s been many years in the making and it may take a couple more before it’s widely available, but Microsoft’s research into real-time voice translation looks set to pay off.
AlchemyAPI rolls out deep-learning-based computer vision as a service
Deep learning startup AlchemyAPI is expanding its scope from text analysis to computer vision with a new API-based service called AlchemyVision. The company’s neural network system is continuously scanning images on the web and getting smarter by the day.
5 technologies that will help big data cross the chasm
Big data has been a buzzword for years, but it’s a lot more than just buzz. There are now so many tools and technologies for creating, collecting and analyzing data that almost anything is possible if you know where to look.
Democratizing deep learning with an iPhone app and open source SDK
Computer vision startup Jetpac, which specialized in categorizing Instagram images from vacation destinations, has released an iOS software development kit and app based on a popular deep learning architecture.
In the rush to hype AI, it’s easy to forget about web search
Microsoft keeps rolling out new features in Bing that it claims make it superior to, or at least more interesting, than Google’s dominant search engine. They’re not the sexiest applications of artificial intelligence, but they are easy, practical and tied to big money.
Why Nvidia thinks it can power the AI revolution
GPU maker Nvidia is hoping to ride the wave of artificial intelligence. The company is already powering machine learning workloads within data centers of large companies, but now it’s targeting individuals with a cheap-but-powerful development kit targeted at robotics and the internet of things.
When emotionless machines are better than people at reading emotions
New research highlights a computer vision system that’s much better at telling when people are faking expressions of pain than are other humans. It’s the latest in a series of computer vision advances that foretell a brave, new and possibly creepy world.
Artificial intelligence startup Vicarious collects $40 million from tech elites
The company wants to build a system that sees and learns like the human brain. That could be decades out, but Silicon Valley is very interested.
Facebook shows off its deep learning skills with DeepFace
A Facebook research paper details a new method for recognizing the people in images by combining deep learning techniques with a method for recomposing angled images as straight-on ones. It’s the latest in a series of advances web companies have made in this field.
These Structure Data speakers are blazing new trails in the information age
Structure Data has a great lineup of speakers, including a handful that will be talking about how to take advantage of new types of data. Here is a list of sessions anyone interested in sensors, location or artificial intelligence won’t want to miss.
Robot startup gets patent for neural networks on GPUs
A robotics startup called Neurala has received a patent (No. 8,648,867) for a GPU-based system designed to run artificial neural network models. The…
Why video is the next big thing in big data
One of the big themes at our Structure Data conference in March is the advent of new techniques to make sense of new data sources. One of the most-promising is video, which had value well beyond capturing crimes and making us laugh on YouTube.
If Watson is going to be a $10B biz, IBM had better figure out the cloud
IBM has launched a whole new division around Watson, but a slow start in terms of uptake might be a sign of concern. Watson’s best chances for success might lie in the cloud, where its capabilities can really be pushed to the limit.
How Google cracked house number identification in Street View
http://www.technologyreview.com/view/523326/how-google-cracked-house-number-identification-in-street-view/ This post from the MIT Technology Review discusses how Google used deep learning to recognize houses numbers and make Street View…
New vision algorithm will label you ‘goth,’ try to sell you black makeup
A new algorithm out of the University of California, San Diego, is getting better at predicting what subculture, or “urban tribe,” people belong to by learning what visual features usually accompany goths, bikers, surfers and others in online photos.
Facebook hires NYU deep learning expert to run its new AI lab
Facebook has hired deep learning expert Yann Lecun from New York University to head up its new artificial intelligence lab. It’s part of a bigger push along with — and against — companies like Google and Microsoft to advance research while improving their platforms.
How deep learning can teach computers Spanish without a tutor
Google has released another paper showing off the power of its deep learning techniques for text analysis. It shows how models can detect similar usage of words across different languages, meaning it can accurately translate words and concepts from one language to another.
AlchemyAPI says it’s delivering Google-level deep learning as a service
A Denver-based startup called AlchemyAPI is close to rolling out deep-learning-based image recognition via its API service. The company has made something of a name for itself in the text-analysis world, and it says it can do image recognition as well as Google.
Why machine learning might be a wearable camera’s best friend
The glut of research in teaching computers to analyze and understand images could prove very helpful in letting us take full advantage of the countless hours of video we’ll produce as wearable cameras go mainstream.
GigaOM Show: Samsung watch secrets spilled! B&N’s Nook not killed! word2vec has us thrilled!
On this week’s show we give you the scoop on Samsung’s watch, ponder if the Nook is cooked and learn all about deep learning.
We’re on the cusp of deep learning for the masses. You can thank Google later
Google researchers have developed new methods for analyzing language using deep learning techniques. They’ve also open sourced an implementation of their work so any researchers can experiment with it. It could be the first of many deep learning tools designed for mass consumption.
Want cheap high-performance computing at scale like Google? Look to GPUs
New research shows that it’s possible to train Google-style neural networks with GPUs and Infiniband at low cost. The work could help drive further GPU adoption.
