Text abstract generation model training method, related equipment and readable storage medium

• Main Authors: LIU QING, LI XIAOYU, ZHANG ZEQUN, KANG HANKUN, DONG PENGCHENG, JIN LI, SUN XIAN
• Format: Patent
• Language: Chinese; English
• Published: 15.07.2022
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention discloses a text abstract generation model training method, related equipment and a readable storage medium. When the method is adopted to train the text abstract generation model, the abstract label, the text category label and the text style label of the training text are determined by adopting a remote supervision thought, so that manual data labeling is not needed; a text abstract generation model composed of a coding layer, an interaction layer, a text abstract automatic generation layer, a text category self-adaption layer and a text style self-adaption layer is utilized to solve the problems of polysemy and gradient disappearance of a sequence-to-sequence model based on the long and short term memory network. In addition, the influence of the text category and the text style on the abstract content is considered during training. Therefore, the text abstract is automatically generated by adopting the text abstract generation model obtained by training, and the accuracy of the generated tex