Neural Networks Assisted Vigenère Cipher Encryption of Text

With the recent advent of the digital interface in our day-to-day life, internet usage for a variety of purposes such as bank transactions, social media, communication and cloud storage has risen exponentially. Since the internet is surfed worldwide, the odds of the data being exposed to cyber criminals are significantly high. This establishes a strong rationale for devising an automated framework for end-to-end encryption for securing the data. In the present study, we utilized the computational efficiency of long short-term memory (LSTM) networks to predict the accurate vigenère cipher. The synthetic dataset with 10000 samples consisting of the alphabetical plain text and corresponding Vigenère cipher based on the fixed key “KEY” is utilized to train, test and validate the deep learning model. The validation of the LSTM model depicted exceptional accuracy, which is further reflected while performing different experiments of encrypting the plain text in the final stage of this study.