Loading
Zero-shot prediction applies a model to a task or class it was never trained on, with no task-specific labels at all.
Zero-shot learning is the setting in which a model must predict classes it did not see in training (Wikipedia). For biology foundation models the zero-shot checks are whether embeddings cluster by tissue or subtype without labels (cluster purity, from cluster analysis) and whether a sequence model scores pathogenic variants above benign ones without being told which is which. A refuted zero-shot claim is still informative: it says the pretraining did not capture that biology.
Shares Genomic and protein language models: Evo 2, Enformer, ESM, Embedding (learned representation), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Transfer learning and the low-label regime, Embedding (learned representation), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Autoregressive (next-token) modelling, Embedding (learned representation), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Transfer learning and the low-label regime, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Embedding (learned representation), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Embedding (learned representation), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Variant effect prediction, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Embedding (learned representation), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.