Hi Tyrel,
I would be personally interested in tackling the problem of materials web scrapping. At the time of the last SS Gordon Conference it was my understanding that if one wanted to develop ML training sets for predicting say...thermoelectric performance, then the feature data had to be obtained by having a poor set of undergrads, grads, postdocs, etc read 1000+ thermoelectric papers, physically record that data, blah, blah, blah. It would obviously be easier if one could web scrap this data. But perhaps I am out of touch as that was a year or two ago, or I am off base. If not, I believe this tool could reduce the time needed to create training sets to develop ML models.
Adam
Hi Tyrel,
I would be personally interested in tackling the problem of materials web scrapping. At the time of the last SS Gordon Conference it was my understanding that if one wanted to develop ML training sets for predicting say...thermoelectric performance, then the feature data had to be obtained by having a poor set of undergrads, grads, postdocs, etc read 1000+ thermoelectric papers, physically record that data, blah, blah, blah. It would obviously be easier if one could web scrap this data. But perhaps I am out of touch as that was a year or two ago, or I am off base. If not, I believe this tool could reduce the time needed to create training sets to develop ML models.
Adam