Posts Tagged ‘deep neural networks’

Deep Neural Network Learns Language from Wikipedia

Wikipedia

San Francisco, CA (PRWEB) December 19, 2013 — Big news for big data: the makers of Ersatz (http://ersatz1.com/), a platform for building “deep neural networks” in the cloud, have fed their algorithm over 4 million Wikipedia articles, and this word cloud is what it learned: http://wordcloud.ersatz1.com.

 

The 3D word cloud combines over 25,000 words into a highly ordered network of associations. To accomplish this task, Ersatz used the patterns of the English language to learn what words go where and how they are most often presented. Read more

AlchemyAPI Improves Service Using Deep Neural Networks

While Google, Microsoft and Apple have been making significant product breakthroughs using deep neural networks (DNN), the question remains: Is using DNN a practical approach for companies that are not web giants? For some, the answer is a resounding “yes.” AlchemyAPI, a natural language processing platform, has developed the ability to train very large and deep neural networks efficiently while remaining cost effective.

“In the area of language modeling, for example, we are able to train on huge amounts of data (a hundred billion words or more) in less than 24 hours whereas in the research community, many groups are spending 4-6 weeks to train a system on fewer than a billion words,” according to Elliot Turner, AlchemyAPI Founder and CEO. Read more