Advanced Maths Training for Open-Source Contributors.
Dear friends,
I'm the Co-Founder & Head of Mathematical Sciences at Zaiku Group. Are you a software engineer or data scientist who may be interested in making impactful open source contributions to emerging tech topics such as Geometric Deep Learning & Quantum Computing? Is mathematics a major stumbling block for you? Well, if yes, then we may be able to help! More contextual information is highlighted below.
1. Geometric Deep Learning
In 2019, Professor Michael Bronstein sold his startup (Fabula AI) to Twitter, where he is now leading their Graph ML team. Professor Bronstein recently wrote an influential post entitled 'Graph Neural Networks through the lens of Differential Geometry and Algebraic Topology'. In this post, he makes solid arguments to why the two advanced branches of pure mathematics referenced in the article can indeed help bring a new perspective on Graph Neural Networks (GNNs). For more insights on the applications of GNNs, please see this paper. Bronstein also recently co-authored ‘What does 2022 hold for Geometric & Graph ML?’ article that you find insightful.
Now, knowledge of Algebraic Topology & Differential Geometry is not common, even for some of the brightest AI/Data Scientists, whether in academia or industry! Hence, we have decided to help remove the mathematical friction by offering free online courses (live) on these advanced mathematical topics for people from software engineering backgrounds. We've already delivered a crash course on Topology & Differential Geometry that is now available on this YouTube playlist. This crash course covers the basics of mathematical objects called 'Manifolds', which play an essential role in applied topics such as Geometric Deep Learning.
The referenced crash course above was initially produced for our quantumformalism.com community. However, we are now looking to do another version more tailored to those interested in Geometric Deep Learning & Topological Data Analysis. Please visit our newly created AI/ML community initiative kahler.ai and subscribe for insights and updates about our upcoming activities relevant to AI/ML.
2. Quantum Computing
Quantum computing (QC) is in a more infancy stage than AI despite the current media hype! Even in the Noisy intermediate scale (NISQ) era that we are in, there are some interesting results for deep-neural network use cases. For example, a research team from a well-known Asset Management firm recently published a paper claiming that they had a classical model that needed thousands of iterations during the training to reach an acceptable performance level. However, they claim they managed to build a hybrid quantum deep-neural net that could reach a comparable performance level, using just a few hundred iterations!
Our quantumformalism.com initiative aims to expose advanced mathematical topics relevant to quantum computing to a diverse group of STEM professionals, like developers, data scientists and researchers. These are often professionals looking to switch to quantum computing, and they have gone through the introductory courses to quantum computing. But they want to move to the next level, to the professional quantum algorithms developer level, in the industry or academia. But to do that, they need a whole level of very abstract mathematical knowledge & skills to stay in touch with the professional algorithms developers in the field. This is because the QC algorithm development is still done at a very low level, i.e. at the gate level! So the underlying mathematics of the quantum gates becomes very important.
If you would like to take the upcoming (Feb 25 start) crash course on Group theory, please visit https://www.quantumformalism.com/group-theory and subscribe. Group theory applies to both quantum computing, Geometry & TDA. The only prerequisite is understanding the basics of set theory. However, if you have had an introduction to Linear Algebra, that will be an excellent addition for the advanced sections of the course when we cover the General Linear groups GL(C, n) & GL(R,n) and their subgroups such as; U(n), SU(n), O(n) and SO(n) etc. If you take the course, you'll understand what all this means!:)
For more info about the upcoming course above, please check out my recent post:
3. An Open Call to Open-Source Contributors
We are very keen to connect with those interested in collaborating with us to potentially contribute to open-source projects on the pipeline or contributing to existing open-source projects where having an advanced mathematical training could make a massive difference in terms of progress.
We’ll provide a dedicated mathematical upskilling program for open source contributors, and we’ll offer graduate-level mathematicians to help and guide contributors on the mathematical side of things. If you would like to learn more about our mathematical training scheme for open-source, please fill out this Google form so we can follow up with you https://forms.gle/yE5tKx4xj5bq5VYQA.
Finally, please feel free to connect with me on LinkedIn and Twitter.
Many thanks,
Bambordé