Machine learning startup Startup.ML today unveiled a new website called Deep Learning News. It’s a place to go to find information online about deep learning, which is an increasingly trendy type of artificial intelligence.
The new site is in the style of Hacker News, a popular online water cooler for programmers from prominent Silicon Valley accelerator Y Combinator. It’s also similar to DataTau, a sort of Hacker News for data science that was formed in 2013.
“We made a hackernews clone dedicated to deep learning because DataTau and Reddit are too general (and often cluttered) when it comes to ‘data science’ topics,” Arshak Navruzyan, one person involved with Startup.ML, told VentureBeat in an email. “DL is a pretty specialized branch of machine learning and deserves a proper community.”
Major tech companies like Facebook, Google, and Microsoft, as well as smaller outfits like Pinterest and Snapchat, have lately been allocating more and more of resources to deep learning. The approach involves training artificial neural networks on large sets of data — like pictures — and then throwing a new piece of data at them in order to get an inference about it in response. Startups like Clarifai, MetaMind, Nervana, and Skymind, among others, are striving to make deep learning available to a wide swath of companies. And now, for those who want a forum to discuss developments in the field, there’s Deep Learning News.
On the front page of the site today, there are links to a radio interview with deep learning luminary Geoff Hinton, a blog post from postdoctoral student Kyunghyun Cho about machine translation using deep learning, and a recent academic paper on visualization tools for artificial neural nets.
The entity behind the new site, San Francisco-based Startup.ML, provides fellowships for aspiring data scientists to build machine-learning systems for companies to put into production.