Aidong Zhang

Learn about spurious correlations and their impact on machine learning systems.

New approaches improve understanding and transferability in neural networks.

A new method improves language models' adaptability to unseen tasks.

A new benchmark highlights the risks of spurious bias in multimodal language models.

A new method combining concept learning and disentangled representations for better model understanding.

i-MedRAG improves medical question answering through iterative follow-up queries.

KG-CoI improves the accuracy of hypotheses generated by large language models.