Unverified50% confidenceFactExact time
Scaffold split groups molecules by their Murcko scaffolds, ensuring that training-set and test-set molecules have different scaffold types, and is a rigorous generalization evaluation protocol in molecular machine learning.
1
Sources
50%
Confidence
Long-term
Relevance
10/22/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedScaffold Split groups molecules according to their Bemis-Murcko scaffolds, ensuring that all molecules sharing the same scaffold appear only on the training set or test set side, in order to simulate the ability to predict unfamiliar chemical scaffolds.80% similarUnverifiedWhen processing molecular tasks, Scaffold Split should be used instead of random splitting; otherwise, the model's performance will be inflated due to data leakage.69% similarUnverifiedDesignability is a commonly used evaluation criterion in generative protein design: the generated backbone is fed into an independent sequence design model (such as ProteinMPNN), and the predicted sequence is then refolded using a structure prediction tool (such as ESMFold or AlphaFold); if the predicted structure is highly consistent with the original backbone (usually judged by an scRMSD threshold), the backbone is considered designable.63% similarUnverifiedJoint modeling allows the multimodal associations between sequences and structures to be captured end-to-end within the model, rather than being split across different training stages.61% similarUnverifiedMolCLR constructs positive and negative sample pairs by applying different data augmentations to molecular graphs, while Grover and GraphMAE adopt masked pre-training strategies to learn molecular representations.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1084249API
curl https://kongchang.com/api/v1/knowledge/claims/1084249MCP
get_claim(id=1084249)