Yunhua Zhong

Pretrained scientific foundation models can overcome massive chemical and instrumentation shifts entirely at test-time by using reference-guided retrieval to recalibrate latent predictions on the fly without query supervision.

Achieve strong spatio-temporal video grounding with only 10M trainable parameters by smartly adapting pre-trained 2D visual-language models.

Stop struggling with inconsistent benchmarks for mass spec prediction: FlexMS offers a flexible framework to build and evaluate diverse deep learning architectures, revealing key factors that drive performance.