MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202641094082 A) filed by Vellore Institute Of Technology on August 03, 2026, for System And Method For Domain-Adaptive Explainable Bug Report Summarization.
Inventors include R. Kannadasan; and Prabakaran. C.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: A Domain-Adaptive Explainable Bug Report Summarization System (104) is provided. The system (104) includes at least one processor (106) and a memory (108) operatively coupled to the at least one processor (106). The memory (108) stores instructions which, when executed by the at least one processor (106), cause the processor (106) to receive one or more software bug reports, preprocess the received software bug reports to generate standardized textual data, generate software- domain contextual representations of the standardized textual data using a domain-adaptive transformer-based machine learning model (116) trained on a corpus of software bug reports, generate an abstractive summary from the software-domain contextual representations, determine token-level attribution values corresponding to the generated abstractive summary using an explainability algorithm applied to the domain-adaptive transformer-based machine learning model (116), and output the generated abstractive summary together with the token-level attribution values. FIG. 1
Disclaimer: Curated by HT Syndication.
