Bucket Masking is introduced, a structure-aware masking strategy that selects groups of residues based on their proximity in three-dimensional space, preferentially masking structurally coupled regions during training, and enables up to a 14% improvement over standard random masking in downstream protein fitness predic...
Thomas Walton, Ayan Goel, Amirali Aghazadeh· arXiv.org· 1 citation
An evolution-aware spectral framework is developed that incorporates the evolutionary distribution of amino acids at each sequence position, inducing an orthogonal decomposition under the evolutionary measure while preserving efficient spectral algorithms for scalable analysis and enables more accurate recovery of fitn...
This work introduces function-aware masking, a family of pretraining algorithms that align mask placement with specific functional priors (e.g., from IMGT annotations or structure predictions) to shape the learned representation space, and demonstrates that informed mask placement provides a parameter-free mechanism fo...
Ayan Goel, Thomas Walton, Amirali Aghazadeh· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.