search

waynebhayes / SANA

Simulating Annealing Network Aligner
25 stars 40 forks source link

Development #13

Closed vietly23 closed 6 years ago

vietly23 commented 6 years ago

Add base class for Measures in SANA. Recommend merging and squashing my commits as to not pollute commit history.

vietly23 commented 6 years ago

Resolves #10

waynebhayes commented 6 years ago

Viet,

Not sure which email address you're still reading so I'm responding to your pull request.

Thanks for these additions to SANA 2.0! However where we still really need help is the old multi-pairwise. We're still having incremental evaluation problems (although not as bad as previously). Your WEIGHTED code seems fine, and Pasha and William have been adding required functionality to get multi-pairwise close-to-working. We're really close, except for that damn incremetal eval bug on the SES score. At some point you said you'd fixed it; right now we're mostly only getting small errors (eg., 2nd or 3rd digit of the SES score), but if you think you knew what was causing the problem can you go see if that is still the problem?

I can provide you example networks where it gives the incremental eval warning if you want.

vietly23 commented 6 years ago

Hi Professor Hayes,

I don't know what could be causing the problem, but if you could give me example networks I'd be more than happy to take a look! Feel free to email me at my uci email viettl@uci.edu