Discovery of Unstable Singularities · HackerPapers discussions

Discovery of Unstable Singularities — HN submission

Sat, 13 Dec 2025 00:05:04 GMT

Discovery of Unstable Singularities (In 3D Navier-Stokes Équations)

3 points; 1 discussion reply on HN. Paper and source history. A submission is not an endorsement.

Discovery of Unstable Singularities — HN submission

Mon, 20 Oct 2025 16:04:03 GMT

Show HN: Unstable Singularity Detector update 1.4.1

Hi HN, I'm the author of this project. Unstable Singularity Detector is an open-source implementation of a Physics-Informed Neural Network (PINN) system that detects blow-up singularities in PDEs — like those found in fluid dynamics. It’s inspired by the methodology in a recent DeepMind paper, but fully independently built from scratch. The system predicts instability thresholds (lambda values), trains PINNs in multiple stages, and reaches machine-level residuals (10⁻¹³) using a high-precision Gauss–Newton optimizer. This new v1.4.1 release adds: - Multi-stage PINN orchestration with FSDP and…

1 point; 2 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Discovery of Unstable Singularities — HN submission

Mon, 06 Oct 2025 20:43:05 GMT

Discovery of Unstable Singularities in Navier Stokes (DeepMind)

2 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Discovery of Unstable Singularities — HN submission

Sat, 04 Oct 2025 12:15:15 GMT

Show HN: Unstable Singularity Detector

### Why I built - Earlier this year, DeepMind published groundbreaking work on detecting unstable singularities in fluid dynamics—one of the deepest open problems in mathematical physics (arXiv:2509.14185). - The challenge? No public code. - Extreme numerical precision requirements. - Complex multi-stage training. This makes independent verification nearly impossible. - So I built an open-source re-implementation to bridge that gap. --- ### What It Does Detects finite-time blow-ups in PDEs using Physics-Informed Neural Networks: - Lambda prediction formulas (<1% error vs. published…

1 point; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Discovery of Unstable Singularities — HN submission

Mon, 29 Sep 2025 10:20:45 GMT

Discovery of Unstable Singularities

1 point; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Discovery of Unstable Singularities — HN submission

Thu, 18 Sep 2025 19:24:47 GMT

Discovery of Unstable Singularities

4 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.