Publication Date

Spring 2026

Degree Type

Master's Project

Degree Name

Master of Science in Computer Science (MSCS)

Department

Computer Science

First Advisor

William Andreopoulos

Second Advisor

Robert Chun

Third Advisor

Rashmi Vishwanath Bhat

Keywords

generative engine optimization, knowledge graphs, intent inference, natural language processing, content visibility

Abstract

The way that people access information has changed vastly with the advent of Generative Search Engines (GSEs). Previously, they used to click on the list of websites returned by the Search Engines. Now, they get answers to the questions they have directly from the GSEs in the form of natural language responses. So, from the point of view of content creators, visibility no longer just means to be ranking higher in search results. It now means to write the content in a way that is preferred by the GSEs. But there are different types of GSEs that use different algorithms to retrieve content, generate answers and cite sources. Unfortunately, this is not made public to content creators. We propose Query Implied Generative Engine Optimization (QIO) to address this challenge. This is a method that identifies the intent of targeted readers, looks for information gaps they might be interested in and rewrites the content by filling the gaps. We achieved an increase in visibility compared to the unchanged content by using QIO. It also works consistently well across different phrasings of the same query highlighting its robustness in real-world search scenarios.

Download

Available for download on Saturday, May 22, 2027

Share

COinS