When do cheap embeddings beat protein language models? A theoretically-grounded hashing sketch for biological sequence classification
Murmur2Vec is introduced, an alignment-free, training-free embedding that aggregates k-mer counts into a small hash table via the deterministic MurmurHash function, and it is cast as a randomized sketch of the classical $k-mer spectrum kernel.
Sarwan Ali, T. Murad, Imdadullah Khan et al.
· 0 citations