Heads up: Our Ideas Factory has been refreshed, levelled up, and grown-up into Alphero Intelligence. Some of our old posts are pretty cool tho'. Check this one out.
- We use all sorts of third party AI and machine learning platforms and frameworks, both on client projects and when prototyping things for the Ideas Factory.
- We figured it would be handy to build our own reference repository of these that we can refer to for projects or innovation work.
- The more smarts our team have in their kit bags, the better the digital services we can produce for the world.
It’s getting easier for smart developers to integrate AI and machine learning into solutions
Democratisation of technology is leading to a load of out-of-the-box frameworks, tools, and platform-as-a-services where developers (without PhDs) can leverage the smarts of underlying machine learning or scientific engines, and build their capabilities into intelligent new products and services.
Our team has integrated a number of these into new digital products for clients and prototyped and tested many more during Ideas Factory R&D #ideasfests. We’ve used everything from computer vision frameworks, Javascript libraries to identify blocks of trees (in Google Earth images), to natural language processing and machine learning for automated decision making.
Ultimately we want to help our clients to do game-changing things - and to make the world a better place. The more smarts our team have in their kit bags, the better the digital services we can produce for the world.
There’s so much stuff out there that you can leverage… we now have an epic Google workbook tracking and evaluating platforms, tools and frameworks
In the interests of putting some order around what we do in this space (and ensuring we share what we learn), we asked ourselves two questions:
- What Machine Learning, AI or data science capability out of the box can a clever developer use without requiring a PhD in AI or advanced mathematics?
- If we did a market sweep of the ecosystem, what else is out there that we should try? And what are the features and pros and cons of each? And who should or could we partner with?
We now have a living catalogue of over 40 (and counting) interesting and super cool libraries and services that provide advanced machine intelligence services without the need to get into crazy models and statistics. We’ve also done an extensive review of the many services provided by Amazon, Google and Microsoft to understand how the different parts of their ecosystem play together.
Dropping a few names - here are just some of the things we’ve tried out:
- Machine learning frameworks including MLKit, COreML and Tensorflow Lite
- Data science platforms such as KNIME, RapidMiner, Domino, Anaconda and DataBricks
- Platforms that aim to help with asset management such as C3
- Analytics and business intelligence platforms such as Ai-one, Alteryx and Teradata
- Video metadata creation tool such as Valossa
- Analysis of big offerings from companies like Amazon, Google, Microsoft, Oracle and IBM
Here’s some examples of things we have prototyped that leverage freely available frameworks and platforms
