ProtBERT Embeddings
Cross-source consensus on ProtBERT Embeddings from 1 sources and 5 claims.
1 sources · 5 claims
Benefits
Preparation
Risks & contraindications
Evidence quality
Highlighted claims
- ProtBERT embeddings were extracted with Rostlab/prot_bert in inference mode without fine-tuning. — Evaluating the Limitations of Protein Sequence Representations for Parkinson's Disease Classification
- Mean-pooled last hidden layer vectors produced 1024-dimensional fixed-length ProtBERT representations. — Evaluating the Limitations of Protein Sequence Representations for Parkinson's Disease Classification
- ProtBERT embeddings produced the strongest and most balanced supervised performance. — Evaluating the Limitations of Protein Sequence Representations for Parkinson's Disease Classification
- The best ProtBERT shallow MLP configuration reached about 0.704 F1 and 0.748 ROC-AUC, indicating moderate rather than strong discrimination. — Evaluating the Limitations of Protein Sequence Representations for Parkinson's Disease Classification
- Frozen ProtBERT limits conclusions about what fine-tuned protein language models might achieve. — Evaluating the Limitations of Protein Sequence Representations for Parkinson's Disease Classification