September 22, 2017 | Written by: Niklas Lyng Pedersen
Categorized: Cloud | Watson
Share this post:
Everybody is talking about Artificial Intelligence (AI) without really understanding that AI and Machine Learning (ML), like all other areas of new strategic decisions, demand the right foundation.
So what are the considerations and where are the opportunities and pitfalls?
Right now, we are almost at the same level of enthusiasm as with the birth of the car at the end of the 19th. century: “Hurrah it can drive by itself! Let’s make it look like a horse wagon and put a walking man in front to warn the citizens of this beast coming into our towns!”
But nobody was really considering what was underneath the hood, the size of the engine, the speed, the consumption, economy, robustness etc.
The more mature the AI/ML technologies become, the more questions and demands will come from the users and the management regarding response time, availability, functionality, security etc.
In order to address these demands and to be able to adapt to the upcoming demands, which are not even considered right now, it is important to create the right foundation and policies for your AI/ML environment.
Please find some of the areas you should have in mind below:
Public -, Private -, On Premise – or hybrid cloud
Many think they can only choose the major public cloud when choosing AI/ML platform, but like all other IT areas you have the choice between public, private, on premise and hybrid cloud. You need to consider how much ownership you will have on your artificial intelligence and machine learning projects.
If you choose a public cloud solution, you will learn that it is very easy to start a project. However, you need to clarify the following:
- What is the cost model on long term based on the expected needs?
- How standardized is the offering and will that fulfill your needs looking forward?
- How open is the solution and is there any vendor login?
- Will the solution be able to scale?
- Who is the owner of the algorithms, the data and the results?
Most AI/ML projects will start in the public cloud as Proof of Concepts connecting to different data sources internally and externally. But based on success the organization will often demand a more robust, available and tailor-based solution.
At IBM we have both AI foundation solutions in the:
Here, the advantage is that all platforms are open, scalable and hybrid.
If you are interested in learning more about how to build the right foundation for your AI/ML projects, please feel free to contact me for a dialogue: NIKLASP@dk.ibm.com.