DrupalCon Baltimore 2017: Artificial Intelligence Gives Your Site Superpowers
Drupal's core feature as a CMS is the collection and organization of content. It's been said that Data is the "fuel" for the latest machine learning (ML) techniques called "Deep Learning" and "Neural Networks". It turns out that Drupal's structured content is exactly the type of fuel these techniques need to learn useful patterns and provide state of the art functionality that wasn't even possible 5 years ago.
The audience for this talk should be interested in machine learning and it's applications, but no coding background is necessary. You will learn how to look at your data from the perspective of an ML engineer. You will also discover how you can leverage Deep Learning and your own site content to give Drupal superpowers by using the same technology that's exploding at Facebook, Google, and Amazon.
We will cover:
What is Deep Learning?
What types of problems is Deep Learning good at and how reliable is it?
How to mine your Drupal content as the fuel for Deep Learning.
When should you use existing ML models or services and when should you create your own?
How do you deploy these ML models and use them in production?
Alternatives to training your own. Writing your own classifier isn't that hard, but it can be prohibitive for those without an ML background or compute resources. Luckily, there are solutions for implementing existing ML applications to analyze your site's data, even if you don't have any ML experience. We will demonstate some alternatives using free pre-built models and paid services from Amazon, IBM, Microsoft, Google, and others.
Use a Drag and Drop interface to create, train, and deploy a simple ML model to the cloud using the Microsoft Azure ML API
Use Google Speech API to turn spoken audio into text for use with chatbots and virtual assitants.
Sentiment Analysis (i.e. are reviews positive or negative) using the Watson REST API
Uploaded images can add Face, Logo, and Object detection using the Google Vision API module
Automatically creating summaries from node content using Microsoft's ML API
The audience for this talk should be interested in machine learning and it's applications, but no coding background is necessary. You will learn how to look at your data from the perspective of an ML engineer. You will also discover how you can leverage Deep Learning and your own site content to give Drupal superpowers by using the same technology that's exploding at Facebook, Google, and Amazon.
We will cover:
What is Deep Learning?
What types of problems is Deep Learning good at and how reliable is it?
How to mine your Drupal content as the fuel for Deep Learning.
When should you use existing ML models or services and when should you create your own?
How do you deploy these ML models and use them in production?
Alternatives to training your own. Writing your own classifier isn't that hard, but it can be prohibitive for those without an ML background or compute resources. Luckily, there are solutions for implementing existing ML applications to analyze your site's data, even if you don't have any ML experience. We will demonstate some alternatives using free pre-built models and paid services from Amazon, IBM, Microsoft, Google, and others.
Use a Drag and Drop interface to create, train, and deploy a simple ML model to the cloud using the Microsoft Azure ML API
Use Google Speech API to turn spoken audio into text for use with chatbots and virtual assitants.
Sentiment Analysis (i.e. are reviews positive or negative) using the Watson REST API
Uploaded images can add Face, Logo, and Object detection using the Google Vision API module
Automatically creating summaries from node content using Microsoft's ML API