Machine Learning Times
Machine Learning Times

2 years ago
Twelve Hot Deep Learning Applications Featured at Deep Learning World


For today’s leading deep learning methods and technology, attend the conference and training workshops at Deep Learning World Las Vegas, June 3-7, 2018.  

Deep learning is white hot – both in buzz and in actual value. This recently-proven collection of advanced neural network methods scales machine learning to a new level of capabilities – namely, achieving high performance for large-signal input problems, such as for the classification of images (self-driving cars, medical images), sound (speech recognition, speaker recognition), text (document classification), and even “standard” business problems, e.g., by processing high dimensional clickstreams.

To help catalyze deep learning’s commercial deployment, we’re introducing the all-new Deep Learning World conference, June 3-7 in Las Vegas. DLW is part of Mega-PAW Vegas, which also features four (4) concurrent Predictive Analytics World conferences: PAW Business, PAW Financial, PAW Healthcare, and PAW Manufacturing.

The inaugural Deep Learning World takes place June 3-7, 2018 in Las Vegas

Lead by Founding Chair Luba Gloukhova, DLW’s mission is to foster breakthroughs in the value-driven operationalization of established deep learning methods. The event program case study presentations by speakers from: Capital One, Cisco, John Hancock, Lyft, Microsoft, Monsanto, Northwestern Mutual, PayPal, Seagate, Uber, and Vevo.

Companies presenting case studies at Deep Learning World, June 3-7, 2018 in Las Vegas


Applications of Deep Learning

By way of these case studies, Deep Learning World Vegas will cover a good portion of the wide range of deep learning application areas, including:

  • Self-driving cars
  • Computer vision
  • Speech recognition
  • Entertainment (Vevo)
  • Fake news detection
  • Malware detection
  • Fraud detection
  • Insurance risk and underwriting
  • Agriculture
  • Manufacturing
  • Image resolution enhancement
  • Detecting developmental delay in children

For more on these, check out the detailed conference agenda here.

Neural networks were already exciting 21 years ago, when I first taught grad students about them in my machine learning course at Columbia University. They weren’t “deep” – they were shallow, typically with only one intermediate layer between the input and output layers. And yet, for many domains, they were the leading option. Back then, shallow neural networks were already driving cars around Carnegie Mellon and doing face recognition. Now that we can scale them to harness the power of many intermediate layers, the potential has multiplied many times over.

Microsoft and Deep Learning

Microsoft is adopting deep learning aggressively, and their staff is coming to DLW to speak and to lead the training workshop. Microsoft Research follows a CEO-mandated initiative to transfer deep learning intelligence into all products, services, and supporting systems across the enterprise. Central to that effort is James McCaffrey, who will present on long short-term memory during the DLW conference, and will co-instruct the full-day pre-conference training workshop, “Deep Learning in Practice: A Hands-On Introduction.”

Google and Deep Learning

No discussion of deep learning would be complete without mentioning Google’s unmatched contributions. Like Microsoft, their work begins internally. With deep learning, Google has made significant improvements to most of our main products, including Android, Apps, Maps, Speech, Search, Translation, and YouTube. For example, Gmail now intercepts 99.9% of spam. And your unlabeled photos kept in Google Photos are searchable by ad hoc terms such as, “hug.”

Externally, Google has fostered a great range of deep learning applications, including 1) discovering new planets, 2) assessing the cardiovascular risk of a heart attack and stroke, 3) predicting flight routes through crowded airspace, and 4) detecting logging trucks and other illegal activities in rainforests via auditory data.

Google gave the world the open-source TensorFlow, the leading software solution for deep learning. When comparing to other deep learning libraries by popularity, “TensorFlow is at least two standard deviations above the mean on all calculated metrics.” Most of the DLW case studies mentioned above were built at least in part with TensorFlow.

If you’re climbing the learning curve of deep learning, check out Google’s Distill, which creatively illustrate the technical concepts, and Google’s TensorFlow playground interactive demo lets you play, visualize, and learn. See also Google Brain’s Big Picture Group, which “explores how information visualization can make complex data accessible, useful, and even fun.” Also, I recommend these three videos of Google talks.

Finally, I must mention how much I love the visuals of Google’s Deep Dream; the supermarket video is particularly cool. This video from Nottingham University researcher Mike Pound is a clear, concrete explanation of how Deep Dream works.

Welcome to the New Wave of neural networks! We hope to see you at Deep Learning World.

About the Author

Eric Siegel, Ph.D., founder of the Predictive Analytics World and Deep Learning World conference series, and executive editor of The Predictive Analytics Times, makes the how and why of predictive analytics (aka machine learning) understandable and captivating. He is the author of the award-winning Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die, a former Columbia University professor who used to sing to his students, and a renowned speakereducator, and leader in the field.

Leave a Reply

Pin It on Pinterest

Share This