BigQuery and Cloud Machine Learning: advancing neural network predictions (Google Cloud Next '17)

Опубликовано 9 марта 2017, 0:46

The real value of BigQuery is not its speed. It's the power of "democratizing enterprise data." Because of BigQuery's scalability, you can isolate any workload on BigQuery from others. That means you can let non-engineers, such as sales, marketing, support and others, execute arbitrary quick-and-dirty SQL on BigQuery directly. Any employees in your enterprise can access its big data and quickly do data analytics without affecting performance to the production system.

Now, imagine what would happen if you could use BigQuery for deep learning as well. After having data scientists training the cutting edge intelligent neural network model with TensorFlow or Google Cloud Machine Learning, you can move the model to BigQuery and execute predictions with the model inside BigQuery. This means you can let any employee in your company use the power of BigQuery for their daily data analytics tasks, including image analytics and business data analytics on terabytes of data, processed in tens of seconds, solely on BigQuery without any engineering knowledge.

In this video, Kaz Sato looks at how you can combine Cloud Machine Learning and BigQuery to realize this vision. By watching a demo, you'll see how BigQuery's power of ""democratizing enterprise data"" can be enhanced with a deep neural network model trained with Cloud Machine Learning.

Missed the conference? Watch all the talks here: goo.gl/c1Vs3h
Watch more talks about Big Data & Machine Learning here: goo.gl/OcqI9k