Work

SPERT: Reinforcement Learning-Enhanced Transformer Model for Agile Story Point Estimation

FWCI: 2.66
Cites: 14
Cited by: 7
Related to: 10