I was invited to join Grey Matter Ltd , Microsoft and Microsoft for Startups Europe to deliver to sessions at their ISV event at the Madejski Stadium a few weeks ago. There were almost 200 attendees in total at the event which was based around three main tracks:
My sessions fell within the Developer track. The first session was centred around Microsoft Location Services, specifically some of the Bing technologies that sit within the Cognitive Services ecosystem.
Bringing the power of Microsoft Location Services to your applications
During this 45-minute session I introduced:
I ran through some of the key features of each of these APIs, some of the use cases they can be applied to and some demos of the APIs in action using Postman and Visual Studio.
Using AI to transform your apps with Azure Cognitive Services
My next session was focused on artificial intelligence services that are available in Azure, specifically Cognitive Services.
Some of the topics I covered included:
- How you can use AI to surfacing actionable insights in Twitter and Instagram data
- How the Text Analytics API can help you perform sentiment analysis in just a few lines of code
- How Part of Speech (POS) Tagging can help you identify patterns in text
- Common challenges associated when processing human text
- How LUIS, the Language Understanding and NLP service powered by machine learning, lets you easily build solutions that can process and understand natural language
- Live C# demos of Cognitive Services in action (the demo gods were kind!)
I shared how these technologies can also be used to build conversational AI solutions such as chatbots. I discussed common patterns such as using LUIS to identify the underlying intent which chatbots can then use to invoke specific branches of conversational logic.
I introduced my work and experience with the Twitter API, sentiment analysis and text analytics. I shared how I was able to extend my solutions intelligence by using LUIS to help surface users on Twitter that were expressing commercial intent about products, brands or services.
I explained the pain I had to go through in order to build a custom classifier built on Bayesian Theorem to perform sentiment analysis, a process that involved:
- Sourcing training data
- Cleaning training data
- Removing stop words
- Building a data model
- Testing and iterating to refine the classifier accuracy
I went onto explain how the Text Analytics API ships with a sentiment analysis feature out of the box that can be consumed in one easy to use, scalable REST endpoint!
Each session was over subscribed which was great and there were loads of fantastic questions from the attendees throughout both sessions.
Massive thanks to Grey Matter for asking me to speak at the event and to all involved in helping make it possible!