Unverified50% confidenceFactExact time
Deep learning models such as NetMHCpan are used to predict peptide–HLA binding affinity, with prediction accuracy already exceeding 90%.
1
Sources
50%
Confidence
Long-term
Relevance
9/7/2026
First Seen
Sources
Related Entities
Related Claims
Unverified许多深度学习模型在权重量化到4位后仍能保持95%以上的推理精度68% similarUnverifiedDeep learning models combining embedding representations and attention mechanisms have already surpassed traditional XGBoost in team composition win-rate prediction.63% similarVerifiedDeepMind公开了超过2亿个蛋白质的预测结构数据库，覆盖了几乎所有已知蛋白质63% similarUnverifiedDeepMind的AlphaFold2借助Transformer的多序列比对注意力模块，将蛋白质三维结构预测精度提升至接近实验测量水平63% similarUnverifiedDeepMind launched the GraphCast model in 2023, which is based on a graph neural network architecture and surpassed the then state-of-the-art HRES system on more than 90% of metrics in 10-day forecast accuracy on ECMWF's WeatherBench 2 benchmark.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/873266API
curl https://kongchang.com/api/v1/knowledge/claims/873266MCP
get_claim(id=873266)