Frontier Graph Beta
Explorer

Why Frontier Graph exists.

Frontier Graph came out of a simple personal problem: how to choose a research question when your interests are wider than any one field, paper, or method.

Why I built it

Since PhD application season, I have found it hard to start from one narrow question. My interests have always ranged widely, and I cared less about picking something I could supply quickly than about finding the question where an hour of attention might matter most.

In the early years of my PhD, that was difficult. Later, once I was writing papers and working in teams, I could spread across multiple projects. Over time I developed a specialization in measurement using machine-learning tools and network ideas, but my topic interests stayed eclectic. Frontier Graph comes out of that tension.

Why it matters now

Frontier Graph is meant to help researchers explore questions that still look open but already sit close to enough related work to investigate seriously.

That matters even more as AI increases how much work we can do. If the supply of analysis rises, direction matters more, not less. The scarce input becomes attention and judgment.

What you can do here

Start on Questions if you want possible paper ideas. Open Explorer when you want the nearby topics, paths, and papers behind one question. Use the paper or downloads if you want the method and release files.

Limitations

Frontier Graph is a discovery aid, not a substitute for judgment. It can miss genuinely novel questions, unevenly represent some literatures, or reflect imperfect topic normalization.