Application Development Blog Posts
Learn and share on deeper, cross technology development topics such as integration and connectivity, automation, cloud extensibility, developing at scale, and security.
cancel
Showing results for 
Search instead for 
Did you mean: 
Vitaliy-R
Developer Advocate
Developer Advocate

This week...

...I joined the SAP HANA Cloud Machine Learning Challenge open to our whole SAP Community.

Have you joined it too? It is running through December 16th, 2022: https://groups.community.sap.com/t5/application-development/community-challenge-alert-the-sap-hana-c...

Joining this challenge made me take a stroll down memory lane and share my story with you as a part of...

... "24 Days of SAP Community" calendar.

It is a very personal story and is based on my own experiences and observations with Artificial Intelligence.

Have you experienced or seen it differently? Please share in the comments!

My AI journey...

...started in the mid-1990s. As a student at the Technical University of Wrocław, I learned for the first time about terms, like Artificial Intelligence (then mostly about developing human-coded rule-based so-called "expert systems") and Artificial Neural Networks. I was fascinated. I wrote my master's thesis about parallelization in Expert Systems computations.

I still have some of the books from which I studied back then.

My Artificial Neural Networks study books from 1990sMy Artificial Neural Networks study books from 1990s

I was eager to apply my knowledge upon joining my first employer, but... Even though I joined a global Decision Support Systems department in IT of one of the biggest consumer goods companies, there was nothing related to Data Mining -- as the discipline used to be called at the beginning of the 2000s. Statisticians in some other departments just kept using computer-aided statistics calculations known since the 1970s, but nothing like that in our department at that time.

Gradually I started forgetting about this area of Computer Science. I focused on regular Business Intelligence and Data Warehousing.

Things changed in 2012.

That year a deep convolutional neural net called AlexNet achieved a 16% error rate in an image classification challenge ImageNet. Before that, a good classification top-5 error rate was about 25%. Almost 10 percentage points improvement was a huge deal! All of the sudden everyone in the data industry started talking about "convolutional neural networks" and "deep learning".

A broader IT industry joined the hype train when Thomas Davenport and DJ Patil published their famous article in Harvard Business Review saying that data scientist is going to be the best job of the 21st century. I do not know how many people read a complete article, but its title was cited by presenters at every IT conference for a long time.

Btw, never heard the name Thomas Davenport before? Well, you should, because earlier in the 1990s "Davenport was also instrumental ... in the success of “business process re-engineering” and “Enterprise Resource Planning (ERP),” so his ability to call a major trend and his timing are indisputable." accordingly to Forbes.

And so, ten years ago the boom for Data Science started! You can see it even in the Google Trends:

DS_trend.jpg

So, regardless of the name -- be these Statistics, Expert Systems, Data Mining, Predictive Analytics, Machine Learning, Data Science, or Artificial Intelligence...

...what is different today,
compared to the 1990s,
when I first programmed it?

Technology. The change is mind-blowing!

In the 1990s we could only program a neural network that would fit into 4MB(!) of RAM and process it sequentially in a single thread on a processor like i386 CPU on data that would fit into 1.44MB floppy diskette using a PC. [Ah, beautiful times, weren't they, @qmacro?]

Nowadays deep neural networks process gigabytes or terabytes of data residing in fast RAM and use massively parallel processing on scale-up or scale-out architectures of powerful servers. And all of that at the most affordable price possible thanks to the cloud.

Technology made it possible. Technology made it affordable.

So, no surprise that we at SAP invest a lot today to help every enterprise to become the Intelligent Enterprise: from embedding AI into our SAP applications (ExpenseIt in Concur mobile app being my favorite example) to making a broad portfolio of technical capabilities available to customers and partners. My teammate @noravonthenen described it very well in Meet SAP’s AI Portfolio and What It Can Do For You

I am excited to be a part of this journey.

And I am excited to gain new experiences by participating in the SAP HANA Machine Learning Challenge.

I hope you are too 🤓

ML-Challenge.png