[2102.11107] Towards Causal Representation Learning

The two fields of machine learning and graphical causality arose and developed separately. However, there is now cross-pollination and increasing interest in both fields to benefit from the advances of the other. In the present paper, we review fundamental concepts of causal inference and relate them to crucial open problems of machine learning, including transfer and generalization, thereby assaying how causality can contribute to modern machine learning research. This also applies in the opposite direction: we note that most work in causality starts from the premise that the causal variables are given. A central problem for AI and causality is, thus, causal representation learning, the discovery of high-level causal variables from low-level observations. Finally, we delineate some implications of causality for machine learning and propose key research areas at the intersection of both communities.

13 mentions: @arankomatsuzaki@anirudhg9119@s_scardapane@Paimaamu@arxivabs@arxivabs@arxivabs@arxivabs
Date: 2021/02/23 03:51

Referring Tweets

@arankomatsuzaki Towards Causal Representation Learning Reviews the bidirecitonal relationship of ML and graphical causality and delineate some implications of causality for ML and propose key research areas at their intersection. t.co/mqvWmnDTJN t.co/Z52BDwLnjw
@Paimaamu The title should be towards the holy grail of AI. t.co/EJNreijsAu

Related Entries

Read more [2102.05182] A Deep Learning Approach for Characterizing Major Galaxy Mergers
0 users, 2 mentions 2021/02/13 00:53
Read more [1912.12033] Deep Learning for 3D Point Clouds: A Survey
0 users, 2 mentions 2021/02/25 02:22
Read more [2102.10484] CheXseg: Combining Expert Annotations with DNN-generated Saliency Maps for X-ray Segmen...
0 users, 4 mentions 2021/02/25 09:51
Read more [2102.07432] Fast and accurate optimization on the orthogonal manifold without retraction
0 users, 1 mentions 2021/02/27 11:22
Read more [2102.13243] Swift for TensorFlow: A portable, flexible platform for deep learning
0 users, 2 mentions 2021/03/01 06:54