Using Vector Semantics to Generate Human Readable Text Summaries and Keywords
Abstract (300 words maximum)
In recent times, much attention has been drawn to large language models that can generate text and respond to input in a human-like manner. This is impressive, but simpler applications of natural language processing such as text summarization or keyword extraction can aid in many tasks such as skimming through papers and articles to find relevant information, or querying documents that might have needed information, even if the specific search terms do not show up anywhere. This research will involve analyzing excerpts from various prelabeled articles on Wikipedia in order to train separate recurrent neural networks to output keyword relevance and generative text summaries. An output vector would show how relevant the given text is to certain common topics (Geography, Science, Mathematics, etc.), and a generative model would create text summaries that have comparable parameters to the source text.
Academic department under which the project should be listed
CCSE - Computer Science
Primary Investigator (PI) Name
Abdullah Al Hafiz Khan
Using Vector Semantics to Generate Human Readable Text Summaries and Keywords
In recent times, much attention has been drawn to large language models that can generate text and respond to input in a human-like manner. This is impressive, but simpler applications of natural language processing such as text summarization or keyword extraction can aid in many tasks such as skimming through papers and articles to find relevant information, or querying documents that might have needed information, even if the specific search terms do not show up anywhere. This research will involve analyzing excerpts from various prelabeled articles on Wikipedia in order to train separate recurrent neural networks to output keyword relevance and generative text summaries. An output vector would show how relevant the given text is to certain common topics (Geography, Science, Mathematics, etc.), and a generative model would create text summaries that have comparable parameters to the source text.