I'm in CL and NLP and a lot of the jobs I'm hearing back from tend to be doing stuff for clinical text. I have no background in the life sciences, and I have a math background from undergrad but didn't do much relevant to ML beyond the required analysis and calc and topology.
Part of my question is just what are some good quick ways to get enough of an understanding of life science problems in order to have a vaguely competent view of clinical data. But my other thought was that a lot of the /sci/ textbook guides for other subjects have some helpful recommendations, but I haven't seen anything especially targeted towards building a real foundational understanding of most machine learning stuff (courses always glossed over the more interesting mathematically motivated stuff in favor of staring at diagrams of transformers). I'm aware of things like https://arxiv.org/abs/2104.13478 (a foundational approach to deep learning from the perspective of Geometry) but not much else.
TLDR: Post foundational/theoretic things relevant to ML that you wish people doing work in its applied areas knew.
Part of my question is just what are some good quick ways to get enough of an understanding of life science problems in order to have a vaguely competent view of clinical data. But my other thought was that a lot of the /sci/ textbook guides for other subjects have some helpful recommendations, but I haven't seen anything especially targeted towards building a real foundational understanding of most machine learning stuff (courses always glossed over the more interesting mathematically motivated stuff in favor of staring at diagrams of transformers). I'm aware of things like https://arxiv.org/abs/2104.13478 (a foundational approach to deep learning from the perspective of Geometry) but not much else.
TLDR: Post foundational/theoretic things relevant to ML that you wish people doing work in its applied areas knew.
