“I’d quite like to explore neural nets that are a thousand times bigger than that,” Hinton says. “When you get to [...]]]> Focus of the recent Wired article was again deep learning researches being done at the Google X-labs. In that same article G. Hinton mentions that he would like to train Neural Nets with trillions of parameters:

“I’d quite like to explore neural nets that are a thousand times bigger than that,” Hinton says. “When you get to a trillion [parameters], you’re getting to something that’s got a chance of really understanding some stuff.”

Source: Wired Article

As part of his new job, G. Hinton is going [...]]]> Google acquired the new startup, DNNResearch, established by Geoffrey Hinton and his two graduate students Ilya Sutskever and Alex Krishevsky. The same team won the Imagenet Challenge in 2012. Hinton and his team is going to focus on improving the Deep Learning applications already being used by Google.

As part of his new job, G. Hinton is going to stay with U. Toronto, splitting his time between Google and his duties at the University of Toronto, while Krizhevsky and Sutskever fly south to work at Google’s Mountain View, California campus.

Sources:

http://thenextweb.com/google/2013/03/12/google-acquires-canadian-neural-networks-startup-dnnresearch-aims-to-improve-image-and-voice-search/

http://www.wired.com/wiredenterprise/2013/03/google_hinton/

https://plus.google.com/u/0/102889418997957626067/posts/GWe4AscQdS7

Deep learning techniques have enjoyed enormous success in the speech and language processing community over the past few years, beating previous state-of-the-art approaches [...]]]> A workshop on Deep Learning for Audio, Speech and Language Processing will be held June 16th, 2013 in Atlanta, Georgia. This is right after HLT-NAACL and before ICML, both of which are in Atlanta.

Deep learning techniques have enjoyed enormous success in the speech and language processing community over the past few years, beating previous state-of-the-art approaches to acoustic modeling, language modeling, and natural language processing. The focus of this workshop will be on deep learning approaches to problems in audio, speech, and language. Talks and papers on new models and learning algorithms that can address some of the challenges of these tasks, such as their inherent temporal structure or the need to handle very large data sets, but that have not yet been applied to audio, speech, or language, are encouraged. The goal of this workshop is to provide a uniquely focused forum for the discussion of the intersection of fields of deep learning and audio, speech, and language, bringing together researchers to investigate some of these novel deep learning techniques, and discuss how they can be incorporated into audio, speech, and language processing. This one-day workshop will include a mixture of invited talks, contributed talks, and poster sessions. One goal in selecting both invited and contributed talks will be to cover a broad range of subjects pertinent to the workshop theme, because we believe that an important role of these workshops is the promotion of cross-pollination between fields.

Please visit the following website for more information: https://sites.google.com/site/deeplearningicml2013/

Organizers: Brian Kingsbury, IBM, Tara Sainath, IBM, Li Deng, Microsoft, Andrew Senior, Google
Submission deadline: March 30, 2013
Acceptance notification: April 30, 2103
Final paper submission: May 15, 2013
Workshop date: June 16, 2013