We present a novel counterfactual regularization algorithm to train partner-aware collaborator LLM agents that can discriminate helpful vs. unhelpful information in collaborator utterances

Learning spectral representation from diverse hyperspectral data to improve plant trait prediction (multi-regression)