Leveraging State Space Models in Long Range Genomics
Matvei Popov, Aymen Kallala, Anirudha Ramesh, Narimane Hennouni, Rick Gentry, Alain-Sam Cohen · 2025
Read Leveraging State Space Models in Long Range Genomics by Matvei Popov, Aymen Kallala, Anirudha Ramesh, Narimane Hennouni, Rick Gentry, Alain-Sam Cohen (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/arxiv-2504-06304.txt.
Description
Matvei Popov and co-authors test a new model type for gene data in a 2025 paper. Links across long stretches of DNA matter for how genes work, but common models handle them poorly. The standard model type grows costly very fast as sequences lengthen and cannot stretch past its training length.
Two models based on state space ideas were matched against a standard model of the same size on long range gene tasks. The new models tied the standard one on trained lengths. They also held up on longer sequences they had never seen. The results suggest a better fit for genome work where long context is the norm.
Catalog record
| CREATOR | Matvei Popov, Aymen Kallala, Anirudha Ramesh, Narimane Hennouni, Rick Gentry, Alain-Sam Cohen |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 29 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-06304