How to Explore a New Research Field Using ScholarAssist

Scenario: Understanding the Vector Search Landscape

In this example, we follow a researcher who wants to explore the rapidly evolving yet unfamiliar field of vector search. Instead of manually searching, skimming dozens of papers, and stitching information together, the researcher interacts with AgenticScholar through a sequence of natural language queries.

Walkthrough: From Research Trends to New Research Ideas

Step 1 — Explore Research Trends

Let’s begin with an open question: How has research on vector search evolved recently? Instead of manually searching and reading papers, we explore the topic directly through the taxonomy.

Step 2 — Identify Influential Papers

I'm interested in the topic of dynamic updates for approximate nearest neighbor search. Now I want to know which papers are the most influential.

Step 3 — Drill Down into a Specific Paper

I’m curious about FreshDiskANN. How exactly does it work?

Step 4 — Compare Against Milestone Papers

Let’s compare it with other milestone papers. How does its performance stand relative to others?

Step 5 — Discover New Research Ideas

It seems strong, but is there still room for new research in this area?