MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202621111349 A) filed by Lnct University; Mr. Anurag Jain; and Dr. Vikas Sakalle on September 16, 2026, for Attentive Multi-View Bidirectional Recurrent Framework For Imbalanced Cloud Session Traffic Attack Detection.
Inventors include Mr. Anurag Jain; and Dr. Vikas Sakalle.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: A cloud session traffic attack detection system is disclosed for identifying malicious and benign network sessions under imbalanced traffic conditions. The system acquires cloud session traffic (100, 102), forms a plurality of complementary feature views (106a, 106b, 106c, 106d) representing temporal, statistical, directional, protocol, and behavioural characteristics, and supplies the feature views to bidirectional recurrent processing branches (108). Each branch generates context-aware representations from forward and reverse session sequences. An attention module (110) assigns relevance weights to features, time steps, or view-specific representations and produces an attention-weighted fused representation (112). An imbalance-aware learning module (114) reduces bias toward majority traffic classes by applying class-sensitive training, resampling, weighted loss, or combinations thereof. A classification module (116) processes the fused representation to generate an attack class or benign-session output. The system may further provide confidence information, logging, and security alerts (118, 122). The architecture enables contextual analysis of heterogeneous cloud-session behaviour while improving detection of infrequent attack categories without relying solely on majority-class traffic patterns. (Accompanied Figure No. 1-2)
Disclaimer: Curated by HT Syndication.
