Grassmannian Geometry Meets Dynamic Mode Decomposition in DMD-GEN: A New Metric for Mode Collapse in Time Series Generative Models

Yassine Abbahaddou, Amine Mohamed Aboussalah · 2024

Read Grassmannian Geometry Meets Dynamic Mode Decomposition in DMD-GEN: A New Metric for Mode Collapse in Time Series Generative Models by Yassine Abbahaddou, Amine Mohamed Aboussalah (2024) 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/arxiv-2412-11292.txt.

Description

Two researchers tackled a common flaw of AI models in 2024: they copy only part of what they were trained on. This flaw is called mode collapse. It is well studied for images but less so for data that change over time. The authors give a fresh meaning of the flaw for time series data.

Their tool, named DMD-GEN, splits both real and made-up data into moving patterns. It then checks how far the two pattern sets drift apart. Tests on fake and real data sets matched older scores for still data. The tool also shows which patterns were lost, so builders can fix their models.

Catalog record

CREATORYassine Abbahaddou, Amine Mohamed Aboussalah
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE58 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-11292

Browse the full library · About the Tower of Babel