“Serving AI with Data” – Recap of the Microsoft AI Immersion Workshop

Artificial Intelligence is enjoying its day in the sun. Articles about AI get regularly featured as front page news, and in a wide variety of applications, including conversational bots, autonomous vehicles, connected machines, medical diagnosis, and even better-than-human speech/image recognition, it seems. AI is increasing interest to millions of developers worldwide, as well as companies of all stripes, from garage startups to the largest enterprises.

Microsoft has a long-term commitment to the field of AI. We are partnering with developers and customers as we work towards our dream, which is to democratize AI and its benefits to every person and every organization. AI was a central theme at Microsoft Build earlier this month, where, based on feedback received from attendees and in media reports such as this one, it appears that the community appreciates our AI vision and roadmap.

But it’s still very early days for AI.

A majority of developers are still ramping up, wrapping their heads around this new slew of technologies and accompanying TLAs. Realizing this to be the case, we decided to host a first-of-its-kind Microsoft AI Immersion Workshop earlier this month – a free pre-event to Build 2017, with the goal of giving as many developers as we could accommodate an opportunity to get their hands dirty, write code and build apps using Microsoft’s AI and intelligent data platform stack.

This post is a short summary of the Workshop, and includes key resources from it.


To kick things off, Joseph Sirosh, Corporate VP for the Microsoft Data Group, delivered a keynote talk on Serving AI with Data to a packed house. In it, he describes how the confluence of the cloud, data, and intelligence are driving a new AI revolution that will rival the industrial revolution itself, in terms of its transformative power. Joseph’s keynote included many demos of AI scenarios and customer success stories, and was followed by an open Q&A session with him, which proved super popular as well. Catch his talk, demos and Q&A using the video link below – it’s a great way to get an overview of Microsoft’s comprehensive approach to this space and especially how it will benefit you as a developer:


Following Joseph’s keynote, we ran five parallel workshops, on the topics listed below. Each workshop lasted six hours and featured hands-on technical learning delivered by seasoned Microsoft experts. Materials pertaining to these workshops are now posted to this GitHub site.

WORKSHOP TITLE ABSTRACT
Applied Machine Learning for Developers Workshop focused on Microsoft R Server and Azure Machine Learning. Brief introduction to the R language and IDEs for R, followed by R in the cloud and deployment option use cases. Basics of the data science process, from identifying the question that you would like to see answered, to determining which algorithm to apply, to evaluating how good your model is. How to use Azure ML to apply the data science process and build an ML model, and deploy it as an API on Azure.
Big AI – Applying Artificial Intelligence at Scale Workshop focused on applying AI at scale using a variety of technologies. Topics covered include planet-scale AI on structured and unstructured data, deep learning at scale using Azure, and intelligence at scale using Azure Data Lake.
Building Intelligent SaaS Applications Workshop covering the end-to-end flow of building intelligent Software as a Service (SaaS) apps in Azure. Combining presentations and hands-on experiences, this covers the data tier, including different SaaS design patterns to consider when building, managing and monitoring a scalable and cost-efficient multi-tenant data tier; the app tier, featuring the powerful AI technologies available for you to build intelligent apps; followed by the analytics tier, showcasing how to power interactive analysis on your app telemetry data, backing your SaaS app with a highly scalable data warehousing solution.
Deep Learning and the Microsoft Cognitive Toolkit Workshop focused on deep learning. Introduces deep learning and common use cases, including a hands-on tutorial on how to use the Microsoft Cognitive Toolkit, followed by the different Azure services that you can use to build an end-to-end deep learning solution. The session helps attendees familiarize themselves with the different areas of deep learning, gain hands-on experience in practical applications of deep learning, and recognize the different Azure services to use, when building deep learning solutions.
Weaving Cognitive Services, Azure Services, and Bots to Provide Next-Generation Intelligence This workshop is about how to use the power of deep learning models that have been pre-trained to solve common ML problems such as face/emotion detection, sentiment analysis, and more. It starts with a guided tour of Microsoft Cognitive Services and the Azure service offerings that pair well with them. It then walks through multiple application scenarios involving integrating Cognitive Services together, as well as with Azure Search, Cosmos DB (NoSQL), and other Microsoft technologies to provide cohesive end-to-end solutions. Attendees use their learning to create a bot that uses these services to find specific vacation photos (e.g. “find photos with kids outdoors looking happy”).

Both workshop presenters and participants showed tremendous enthusiasm – indeed, a couple of sessions were still going strong well past our event closing time of 5PM. Attendees had exceedingly positive to say about the workshop and they got tremendous value from it. Their favorite aspect of the event was the opportunity to not just see but also use our technologies, and with seasoned experts close at hand, as guides. Attendees appreciated the access to top-notch Microsoft engineers and data scientists, and enjoyed their opportunity to engage deeply with them.aiiw-2 

These quotes from our attendees nicely sum up the day’s proceedings:

  • “Well-planned, well-organized, very heavily attended, and provided a lot of value. Probably the most productive software engineering workshop I have attended in the past ten years or so. There was no fluff – it was all hard information and guidance.”
  • “Presenters were very highly qualified, world-class ML professionals, but all of them were down to earth, easy communicators that spoke at my level (beginner / intermediate ML) where I could take away ideas I can use in my real day job.”
  • “Keynote Q&A was excellent, very direct answers to a wide variety of broad questions.”
  • “Seeing the momentum of investments and the opportunity to connect to Microsoft folks that work on the technology.”

Thanks to all attendees for spending a day with us, learning about AI. Workshop materials are posted to this GitHub site. For a comprehensive overview of Microsoft’s AI and intelligent data platform vision and direction, tune in to the keynote talk, available here on Channel 9.

We hope you join us on our mission to democratize AI !

CIML Blog Team

from Cortana Intelligence and Machine Learning Blog https://blogs.technet.microsoft.com/machinelearning/2017/05/23/serving-ai-with-data-recap-of-the-microsoft-ai-immersion-workshop/

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s