# arXiv paper: Solving an Open Problem in Theoretical Physics using AI-Assisted Discovery submitted

_Friday, June 26, 2026 at 6:39 PM EDT · science · Latest · Tier 2 — Notable_

![arXiv paper: Solving an Open Problem in Theoretical Physics using AI-Assisted Discovery submitted — Primary](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png)

The arXiv paper titled Solving an Open Problem in Theoretical Physics using AI-Assisted Discovery was submitted in the Computer Science Artificial Intelligence category. It describes a neuro-symbolic system that pairs the Gemini Deep Think large language model with a Tree Search framework and automated numerical feedback.

The system autonomously solved an open problem by deriving novel exact analytical solutions for the power spectrum of gravitational radiation from cosmic strings. The agent evaluated the core integral I(N, alpha) for arbitrary loop geometries and improved on prior AI-assisted work that produced only partial asymptotic solutions.

To support claims of AI-accelerated discovery the paper details system prompts, search constraints and intermittent feedback loops. The agent identified six different analytical methods. The most elegant approach expands the kernel in Gegenbauer polynomials C sub l to the power 3/2 to absorb the integrand's singularities.

The methods produce an asymptotic result for I(N, alpha) at large N. This result agrees with numerical findings and connects to the continuous Feynman parameterization of Quantum Field Theory. The paper outlines both the algorithmic methodology and the mathematical derivations.

## Sources

- [arXiv](https://arxiv.org/abs/2603.04735)

---
Canonical: https://techandbusiness.org/newswire/WMYow9Ig064KslncDO0Y3i
Retrieved: 2026-06-27T03:04:46.764Z
Publisher: Tech & Business (techandbusiness.org)