andy99

Location: Canada (EST) Remote: Remote or hybrid, travel a few days per week is fine Willing to relocate: Open to discussing it, prefer to stay where I am and travel for in-person stuff Technologies: P…

I'd call all of these things specific cases of some of the general problems we've faced with using neural networks for years. There's a big gap between demo and product. One one hand Op…

Skills and experience testing for AI models Explanation: http://marble.onl/managing_ml.html Code: https://github.com/rbitr/pytkml I didn't explain it well; …

I've been playing with it a bit to understand if/how I can use it for manufacturing. It feels like out of the box it's not great for really specific stuff that might be useful industria…

I worked on (wouldn't say I completed) a testing framework for ML models, where you can specify a series of tests for a model to pass. The unique part is emphasizing support from the training dat…

I've thought about something similar before in the context of explainable machine learning [0]. The author alludes to it - explicit knowledge is like an expert system. And I would argue that taci…

Hi! This is a very interesting area, and I agree with you that the future of ML depends on this kind of division of work between what people are good at and ML is good at. I work in a very similar spa…

Well I'd argue that if you take this approach to hiring, you can only hire people who share your area of expertise, that thinks the way you do. This is not generally the case with how ML models w…

> overstates the benefits and undercounts the drawbacks of requiring black-box algorithms to be explainable. I wrote the essay linked below [0] a few months ago. It is very relevant here. I argue t…

Related to understanding how ML models are making their decisions is the concept of influence [1] (or broadly, instance based explanations) [2] where the training data that is most influential in a pr…

Oh cool, I misunderstood, that actually seems like it could work well then