Dear RG members,
In this age of Data Science, do you think that in the Academic systems worldwide providing Degrees in Applied Mathematics, A course on Mathematical Foundations, Tools and Methodologies pertaining to Machine Learning be offered, at different levels of depth and scope, is the urgent need of the times?
If you were to design such a course, based on your subject expertise, which set of topics do you think should be included in such an Academic program ?
In orthopedic surgery there has been a recent interest in machine learning and we are seeing a large increase in studies starting to be published. Unfortunately, the majority of specialists in the field aren't familiar with the theory or methods. I would include topics pertaining to how data is handled, where machine learning can be applied appropriately, how to use and interpret different classification methods, and understanding the current limitations of machine learning such as need for external validation and the "black box" functions.
Thank you Prof. Short for your esteemed opinion and feedback.
In this regard i would ask your opinion on whether Matrix Calculus should also be included as a part of the course, or be left as a separate pre-requisite course ?
Hi Debopam Ghosh
Though my research is mainly in the AI branch of Natural Language Processing (NLP), formal models and theories are shared widely across AI sub fields of which Machine Learning is one of them.
For instance models such as state machines, formal rule systems, logic, probabilistic and vector spaces are fundamental in NLP with state machines, formal rules & logic driving what we term as classical NLP & Probabilistic & Vector space models driving contemporary NLP, all these are basal on the mathematical sub fields of Probability, Statistics, Calculus, Algebra especially Linear Algebra and in other sub-fields such as computer vision and image processing Geometry & Trigonometry are additional mathematical fields whose theories are widely used. In addition to the above the knowledge of set theory and number theory are also fundamental.
Suffice it to say modern day Machine Learning is dominated by what can be termed as Statistical Machine Learning which is mainly driven by Probabilistic & Statistical Models & a lot of Calculus.
- Badges
- Science topic