Research Talk: Enhancing the robustness of massive language models via invariant risk minimization

Microsoft Research - Research Talk: Enhancing the robustness of massive language models via invariant risk minimization