Reliable and Efficient Machine Learning (Final Technical Report)
Du, Qiang · 2025
Read Reliable and Efficient Machine Learning (Final Technical Report) by Du, Qiang (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/doe-osti-3007930-reliable-and-efficient-machine-learning.txt.
Description
Machine learning that scientists can trust is the aim of Qiang Du of Columbia in a November 2025 final report for DOE grant DE-SC0022317. The work builds math tools for learning systems whose answers must be exact, clear, and in line with the laws of physics.
Three fronts were covered: pulling features from fast-changing data, the theory of finding dynamics with neural nets, and rigorous optimization. One product is a two-stage algorithm for stretched non-negative matrix factorization. It beat the existing NENMF method on real diffraction data in both precision and speed. Theory results sharpen optimal transport bounds used in uncertainty estimates.
Catalog record
| CREATOR | Du, Qiang |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 21 KB |
| LICENSE | Public Domain |
| RIGHTS | PUBLIC DOMAIN |
| SOURCE | DOE OSTI |