olBlend: Language-Guided Molecular Blend Retrieval with Mixture Odor Prediction
This work proposes olBlend, a language-guided framework for molecular mixture retrieval that uses MixPOM to predict odor probabilities for candidate molecular pairs and stores these predictions for retrieval, and uses a language model to translate each odor description into three types of semantic requirements.
Dian Jin, Zhe-Chang Chen, Qi-Lin Liang et al.
· 0 citations