Summary: weekly mtg 20170302 (Byron, Matt, me) - mobeets/nullSpaceControl GitHub Wiki

should we go to NCM conference for more feedback from motor people?

minimal firing: if you don't add noise, does it really predict the same exact point (e.g., without correcting for firing bounds)

how to add minimal firing noise: add poisson noise with the rate we observe

minimal deviation: search through firing space; that way we can add noise in the same way

• compare to baseline as well; we might add baseline back in depending on how this turns out

mention baseline in the text?

• mention that if the baseline rate were the best, minimal-deviation could find it

SSS: output-potent angle vs. cursor velocity?

• send email to better explain the issue
• basically need to choose between: simpler vs. consistent with other figs
• Matt: can assume v_t = v_{t-1}: (I-A)^{-1}(Bu_t + c)