Secure Computation Protocol of Text Similarity against Malicious Attacks for Text Classification in Deep-Learning Technology

With the development of deep learning, demand for similarity matching between texts in text classification is becoming increasingly high. How to match quickly under premise keeping private information secure has become a research hotspot. However, most existing protocols currently have full set limitations, and applicability these methods limited when data size large scattered. Therefore, this paper applies vector calculation method case without any complete constraints, it designs computation protocol (SCTS) based on semi-honest model. At same time, elliptic-curve cryptography technology used greatly improve execution efficiency protocol. In addition, we also analyzed possibility malicious behavior participants semi-honest-model protocol, further designed an SCTS suitable model using cut-and-choose zero-knowledge-proof methods. By proposing security mechanism, aims provide reliable computing solution that can effectively prevent attacks interference. Finally, through analysis efficiencies protocols, are verified, practical value learning demonstrated.