Zero Shot Causal

Zero-shot causal learning aims to predict the effects of interventions, particularly novel ones, without explicit training data for those specific interventions. Current research focuses on leveraging large language models (LLMs), particularly transformer architectures, to infer causal relationships from text or to directly perform causal inference through mechanisms like self-attention. This approach shows promise for applications in diverse fields like medicine and policy by enabling causal reasoning from existing knowledge and limited data, potentially accelerating scientific discovery and improving decision-making. The ability to generalize causal knowledge to unseen interventions is a key challenge and area of active investigation.

Papers

December 10, 2024

Causal World Representation in the GPT Model
Raanan Y. Rohekar, Yaniv Gurwicz, Sungduk Yu, Vasudev Lal
Causal Structure Generative Pre Trained Transformer Causal Representation Zero Shot Causal Causal World Model

December 22, 2023

Zero-shot Causal Graph Extrapolation from Text via LLMs
Alessandro Antonucci, Gregorio Piqué, Marco Zaffalon
Large Language Model Text Modality Causal Inference Causal Graph Causal Relation Causal View Ground Truth Causal Zero Shot Causal

November 13, 2023

Do large language models and humans have similar behaviors in causal inference with script knowledge?
Xudong Hong, Margarita Ryzhova, Daniel Adrian Biondi, Vera Demberg
Large Language Model Causal Inference Real Human Script Knowledge Behavioral Similarity Zero Shot Causal

October 31, 2023

Causal Interpretation of Self-Attention in Pre-Trained Transformers
Raanan Y. Rohekar, Yaniv Gurwicz, Shami Nisimov
Self Attention Causal Structure Pre Trained Transformer Causal Explanation Causal Interpretation Zero Shot Causal

October 1, 2023

Towards Causal Foundation Model: on Duality between Causal Inference and Attention
Jiaqi Zhang, Joel Jennings, Agrin Hilmkil, Nick Pawlowski, Cheng Zhang, Chao Ma
Foundation Model Human Attention Causal Inference Causal Learning Strong Duality Zero Shot Causal

January 28, 2023

Zero-shot causal learning
Hamed Nilforoshan, Michael Moor, Yusuf Roohani, Yining Chen, Anja Šurina, Michihiro Yasunaga, Sara Oblak, Jure Leskovec
Zero Shot Personalized Medicine Personalized Prediction Zero Shot Causal