AI as a Service?
Digital Infrastructure and Everything as a Service
AaaS — is of course not the most fortunate acronym. It would at this time be fitting to say when you ASSUME means that you make an ass out of you and me. What else is an algorithm if not an assumption? A mathematical assumption, no doubt, it can be right and wrong as human assumptions. There are now a large variety of ‘as a service’ abbreviations, some have even proposed AI as a Service or AIaaS.
AaaS stands for algorithm as a service and is one of many acronyms I came across within the ‘as a service’ cloud of words.
An algorithm is: a process or set of rules to be followed in calculations or other problem-solving operations, especially by a computer. It is a set of instructions, typically to solve a class of problems or perform a computation. Algorithms are unambiguous specifications for performing calculation, data processing, automated reasoning, and other tasks.
In this manner it is a set of assumptions. An assumption is: a thing that is accepted as true or as certain to happen, without proof. the action of taking on power or responsibility. Narrow AI refers to AI which is able to handle just one particular task. As such we could ask what is the difference? An article by Kaya Ismail in CMS Wire the 26th of October 2019 attempts to explain the distinction between an algorithm and AI:
An algorithm is a set of instructions — a preset, rigid, coded recipe that gets executed when it encounters a trigger. AI on the other hand — which is an extremely broad term covering a myriad of AI specializations and subsets — is a group of algorithms that can modify its algorithms and create new algorithms in response to learned inputs and data as opposed to relying solely on the inputs it was designed to recognize as triggers.
Is not AaaS the same as SaaS the same as AIaaS?
You could argue that ass, I mean AaaS, is similar to SaaS (welcome to acronym heaven btw). Software as a service (SaaS) is a software distribution model in which a third-party provider hosts applications and makes them available to customers over the Internet. SaaS is one of three main categories of cloud computing, alongside infrastructure as a service (IaaS) and platform as a service (PaaS). Acronym overload?
How is AIaaS or MLaaS (Machine Learning as a Service) different? Oleksii Kharkovyna, another writer who writes for Towards Data Science, said it well in his article Machine Learning vs. Traditional Programming:
In traditional programming you hard code the behavior of the program. In machine learning, you leave a lot of that to the machine to learn from data.
As such we can somewhat try to distinguish between these acronyms of Saas against AIaas or MLaaS. Within the field of artificial intelligence (AI) machine learning is the most common technique. So the use of AIaas or MLaas may be a useful distinction, but may be too close to each other for comfort.
Google and Amazon seems to talk of AI and machine learning products almost interchangeably (as I have often seen elsewhere). They currently have the largest platforms for AIaas or MLaas so let us look on the surface of these without a comprehensive review). What follows is slightly adjusted from their respective websites.
Google AI Hub and TensorFlow
TensorFlow is a free and open-source software library for dataflow and differentiable programming across a range of tasks. It is a symbolic math library, and is also used for machine learning applications such as neural networks. It is used for both research and production at Google.
Google additionally have several AI and machine learning products.
AI Hub, a hosted repository of plug-and-play AI components.
AI building blocks with sight, language, conversation, and structured data to applications.
AI Platform, code-based data science development environment, for ML developers and data scientists. Part of this is the Cloud Machine Learning Engine, a managed service that lets developers and data scientists build and run machine learning models in production.
Amazon SageMaker
Amazon SageMaker enables developers and data scientists to quickly and easily build, train, and deploy machine learning models at any scale. It removes the complexity that gets in the way of successfully implementing machine learning across use cases and industries—from running models for real-time fraud detection, to virtually analyzing biological impacts of potential drugs, to predicting stolen-base success in baseball.
Choose from TensorFlow, PyTorch, Apache MXNet, and other popular frameworks to experiment with and customize machine learning algorithms. You can use the framework of your choice as a managed experience in Amazon SageMaker or use the AWS Deep Learning AMIs (Amazon machine images), which are fully configured with the latest versions of the most popular deep learning frameworks and tools. According to Amazon:
- 81% of deep learning projects in the cloud run on AWS
- 85% of TensorFlow projects in the cloud run on AWS
In this sense Amazon seems to stake a claim to be the platform of platforms in MLaaS or AIaaS.
Can smaller actors compete?
Of course, there are always niche markets for specialised products in this sphere and we may find actors who distinguish themselves through one innovation (a new method, idea, product, etc.) or techniques.
The process of writing this article began when I read about Algorithm as a Service in the well-written and concise article by Craig E Ryder on Five AI Startups You Need to Know About. In this article I found NextQuestion and their algorithm as a service (AaaS) which caught my attention (the magic of buzzwords). Check out that AaaS if you like.
Anyhow, thank you for tuning in to #500daysofAI, and again: stay upright.
This is day 36 of #500daysofAI, follow me for daily updates on AI.
What is #500daysofAI?
I am challenging myself to write and think about the topic of artificial intelligence for the next 500 days with the #500daysofAI. This is a challenge I invented to keep thinking of this topic and share my updates with you. Learning together is the greatest joy.
