back to main landing Tweets ⌂

Tweets

2022-10-11T00:16:06+00:00 | 🔗

Hmm some correlation I guess, MCC = 0.1, I wonder how the methods that use the structure do. This'll be a good baseline though!

2022-10-10T21:24:05+00:00 | 🔗

I put in a submission to SAMPL9, with (I think) a relatively interesting method. We hyperparameter searched for a LIGAND-BASED similarity data using Firefly Luciferase data, and then applied that method to NanoLuc. Dunno how this will go, but at the very least it’s interesting! https://t.co/Jl5c1cvJMD

Tweet image for 1579583571626725376

2022-10-08T18:36:48+00:00 | 🔗

RT @tunguz: This past week I came across another paper that purports to get the SOTA for NNs for tabular data. Due to the extreme penchant…

2022-10-07T16:56:35+00:00 | 🔗

More info at: https://t.co/0IUHAjDLVk

2022-10-07T16:56:15+00:00 | 🔗

View our database of molecular property predictor benchmark results, and copy-paste OCE Model Parameter string for complete reproducibility, and to use the best models in your workflows. (In process of migrating all results, we are rerunning everything!) https://t.co/McwVzs8N0M

2022-10-07T03:35:56+00:00 | 🔗

to 50%, AUROC reliably increases for each dataset and model. Note that the increases in performance level off once the number of inhibitor samples reaches 33% of the train set, at which point further improvements are small. More at https://t.co/dbD3lZ2ZfJ (author, Andrew)

2022-10-07T03:35:55+00:00 | 🔗

Modeling imbalanced datasets: We held constant the number of data points (molecule + active/inactive annotation) and varied the % active. More diverse and balanced train sets universally lead to improved performance. As the number of inhibitors present in the train set rises... https://t.co/98XP22ZGOR

Tweet image for 1578227595673575427 Tweet image for 1578227595673575427 Tweet image for 1578227595673575427

2022-10-06T14:27:34+00:00 | 🔗

We got new MBTA trains! Glad that Boston has one the best public transit systems in America, it makes life a lot easier :) https://t.co/Tz0e1ztOuE

Tweet image for 1578029201953001472

2022-10-05T16:22:30+00:00 | 🔗

#cheminformatics #qsar #therapeutics #AI #ML #AI4Science

2022-10-05T16:04:16+00:00 | 🔗

@marinkazitnik @KexinHuang5 @cwcoley @WenhaoGao1 Congrats also on your Nature Chem Bio paper for TDC https://t.co/caTtfRgAxK, and nice seeing ya'll's ACS presentation!

Where else should I get information from? contact me somehow if I'm missing something, I really appreciate new interesting sources of information.