Comments/thoughts from May 25

[mgelbart]

• Speed up the synthetic experiments by vectorizing the evaluation of the synthetic functions (already done for Rosenbrock)
• Compare with paper for AdaLIPO (hopefully now we're confident the evaluation scheme is OK, so if there are bugs, focus on the AdaLIPO implementation)
  • may need to speed it up
  • talk about profiling next time?

[mgelbart]

Look for one other optimizer (one with a pre-existing package) and one other function (e.g. housing or breast cancer should be already in sklearn). Maybe we can try dealing with AdaLIPO together as that's harder.

[mgelbart]

@bradleypick I made a new branch with the changes we made together together. I'll let you merge them in as appropriate. Some you probably don't want, like various print/debug stuff.