MUMBAI, India, Feb. 13 -- Intellectual Property India has published a patent application (202611000729 A) filed by Sherish Johri; Dr. Dhowmya Bhatt; and Dr. Amit Singhal, Ghaziabad, Uttar Pradesh, on Jan. 4, for 'system and method for improving text summarization quality using attention-augmented long short-term memory networks.'
Inventor(s) include Sherish Johri; Dr. Dhowmya Bhatt; and Dr. Amit Singhal.
The application for the patent was published on Feb. 13, under issue no. 07/2026.
According to the abstract released by the Intellectual Property India: "An attention augmented long short term memory (LSTM) based text summarization system is disclosed, comprising a text preprocessing module, an embedding and encoder module, an attention computation module, a decoder module, a post processing and summary generation module, an evaluation and optimization module, and input/output interfaces. The text preprocessing module normalizes and tokenizes an input document, which is then encoded by an LSTM based encoder into a sequence of hidden state vectors; an attention module computes context vectors over these states for each decoding step, and an LSTM based decoder generates summary tokens conditioned on both the context vectors and previously generated tokens. A post processing unit converts the token sequence into a human readable summary, while an evaluation and optimization unit updates encoder, attention, and decoder parameters using quality metrics such as ROUGE or loss values, thereby improving summary coherence and informativeness over time. The system provides an integrated pipeline that enhances summarization quality by explicitly modeling salient content in the source document through attention augmented LSTM components."
Disclaimer: Curated by HT Syndication.
