Course Project for Introduction to Machine Learning, Supervisor: Prof. Yujiao Shi

May. 2025 - June. 2025

• Fine-tuned MoLFormer, a large chemical language model, on the GS-LF olfactory dataset to adapt its generalized representations for specialized odor prediction tasks.
• Evaluated the model against the state-of-the-art OpenPOM on the Keller-2016 dataset, performing odor label classification and pleasantness rating prediction.
• Achieved state-of-the-art performance, with the fine-tuned model matching or surpassing OpenPOM on key metrics, demonstrating the efficacy of fine-tuning for machine olfaction.