Today I started thinking about actively working on two different papers/presentations. One of them I started some time ago revolves around calculating machine learning ROI with examples. The new paper that is going to exist in my stack of stuff will be a paper on very publicly failed machine learning launches in production. Both of those papers will probably try to focus on 10 examples with extensive annotations, but that number could vary widely.
I ended up watching a couple videos on YouTube:
“Daniel Shank – Neural Turing Machines: Perils and Promise – MLconf SF 2016”
“RNN Symposium 2016: Alex Graves – Differentiable Neural Computer”