An Improved Math Word Problem (MWP) Model Using Unified Pretrained Language Model (UniLM) for Pretraining

Table 1

Comparison for math solving task.

Method	Macro precision	Macro recall	F1 score
Sedq2Seq	77.40	76.99	0.77
GTS	72.20	74.30	0.73
Graph2Tree	77.89	75.88	0.76
Our model	79.57	77.69	0.78