Loading
An ablation removes one component or modality at a time and re-measures performance, which is the only way to know what each part contributes; multi-task heads let one shared backbone serve several outputs.
In machine learning, ablation is the removal of a component to determine its contribution to the system (Wikipedia). For a multimodal cancer model the ablation is per modality: drop mutations, copy number, protein or the pretrained encoder and see whether survival or drug-response accuracy moves. Multi-task learning solves several tasks at once through a shared representation with task-specific heads (Wikipedia), which is efficient but can let one easy task (cancer type) dominate the backbone.
Shares Multimodal fusion (early, late, modality dropout), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Accuracy, macro-F1 and confusion matrices, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Mechanism-of-action recovery and known-biology probes, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Accuracy, macro-F1 and confusion matrices, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Multimodal fusion (early, late, modality dropout), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Mechanism-of-action recovery and known-biology probes, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Multimodal fusion (early, late, modality dropout), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.