BiEvo: From autoregression to bidirectional genomic representations with MNTP and architectural symmetrization
Ongoing work on bidirectional genomic representation learning. The project is being actively advanced after the KDD submission cycle and is presented here as a manuscript in …
