The key information extraction technology of electronic documents using deep learning and image recognition technology

With the rapid development of Internet information technology, the application of electronic bidding has provided convenience for enterprises, and at the same time has higher requirements for the diversity and timeliness of the channels for enterprises to obtain bidding information. This paper uses web crawler technology to obtain Internet public bidding data, obtains samples through manual annotation, builds a multi-task learning model based on the pre-training model ALBERT to complete Fine-tuning, and achieves project classification and key information extraction at the same time. Experiments show that the program has a certain improvement in various evaluation indicators, and has achieved good experimental results.