A dual-graph deep learning network that hard-codes petroleum physics alongside learned statistics cuts oil production prediction error to about 20 percent, far outperforming conventional models.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results