Episodic learning for genomic sequences

A new episodic learning capability lets Synthetic Cognition process DNA as sequences rather than static features. Benchmarked across 57 genomic classification tasks against DNABERT-2, NT-v2, and HyenaDNA, it won outright in 52% of them — using a fraction of a percent of the pretraining data those foundation models require.

Rather than fine-tuning a pretrained language model on genomic k-mers, the engine treats each sequence episode as its own learning problem, extracting the regulatory signal directly from the data it is given — no massive pretraining corpus required.