MUMBAI, India, June 22 -- Intellectual Property India has published a patent application (202641050059 A) filed by Dr. S. Subasree; Mr. Vishwanatha Sriram M; Mr. Jeyaprabakaran S; Mr. Albin Tenny; and Mr. Sasidharan S on April 20, 2026, for Artificial Intelligence Powered Password Security:an Ensemble Deep Learning And Statistical Similarity Framework.

Inventors include Dr. S. Subasree; Mr. Vishwanatha Sriram M; Mr. Jeyaprabakaran S; Mr. Albin Tenny; and Mr. Sasidharan S.

The application for the patent was published on June 12, 2026, under issue no. 24/2026.

Abstract: ABSTRACT OF THE INVENTION The " Artificial Intelligence-Powered Password Security: An Ensemble Deep Learning and Statistical Similarity Framework " is a comprehensive cybersecurity system designed to evaluate password integrity through a multi-dimensional analysis framework. Unlike traditional meters that rely solely on character complexity, this invention integrates three distinct assessment engines to provide a holistic risk profile. The system first utilizes TF-IDF (Term Frequency- Inverse Document Frequency) Vectorization and Cosine Similarity to compare user input against a high-volume database of over 14 million known leaked passwords, identifying patterns that are structurally similar to previous breaches. Secondly, it employs an Ensemble Deep Learning Architecture-consisting of Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM), and Probability-based Character Scoring (PPSM)-to predict how susceptible a password is to neural- network-based cracking attempts. Finally, the system incorporates real-time k-Anonymity API integration to check for direct exposure in public breach databases and a heuristic analysis engine to detect localized character weaknesses and keyboard patterns. The invention further includes an automated remedi_ation module that generates secure, Al-enhanced password alternatives based on the user's original input, effectively bridging the gap between user memorability and cryptographic strength. This multi-layered approach provides a more robust defense against modern bruteforce and credential-stuffing attacks than standard rule-based validation methods.

Disclaimer: Curated by HT Syndication.