Neurally-Guided Structure Inference

Sidi Lu, Jiayuan Mao, Joshua Tenenbaum, Jiajun Wu
Proceedings of the 36th International Conference on Machine Learning, PMLR 97:4144-4153, 2019.

Abstract

Most structure inference methods either rely on exhaustive search or are purely data-driven. Exhaustive search robustly infers the structure of arbitrarily complex data, but it is slow. Data-driven methods allow efficient inference, but do not generalize when test data have more complex structures than training data. In this paper, we propose a hybrid inference algorithm, the Neurally-Guided Structure Inference (NG-SI), keeping the advantages of both search-based and data-driven methods. The key idea of NG-SI is to use a neural network to guide the hierarchical, layer-wise search over the compositional space of structures. We evaluate our algorithm on two representative structure inference tasks: probabilistic matrix decomposition and symbolic program parsing. It outperforms data-driven and search-based alternatives on both tasks.

Cite this Paper


BibTeX
@InProceedings{pmlr-v97-lu19b, title = {Neurally-Guided Structure Inference}, author = {Lu, Sidi and Mao, Jiayuan and Tenenbaum, Joshua and Wu, Jiajun}, booktitle = {Proceedings of the 36th International Conference on Machine Learning}, pages = {4144--4153}, year = {2019}, editor = {Chaudhuri, Kamalika and Salakhutdinov, Ruslan}, volume = {97}, series = {Proceedings of Machine Learning Research}, month = {09--15 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v97/lu19b/lu19b.pdf}, url = {https://proceedings.mlr.press/v97/lu19b.html}, abstract = {Most structure inference methods either rely on exhaustive search or are purely data-driven. Exhaustive search robustly infers the structure of arbitrarily complex data, but it is slow. Data-driven methods allow efficient inference, but do not generalize when test data have more complex structures than training data. In this paper, we propose a hybrid inference algorithm, the Neurally-Guided Structure Inference (NG-SI), keeping the advantages of both search-based and data-driven methods. The key idea of NG-SI is to use a neural network to guide the hierarchical, layer-wise search over the compositional space of structures. We evaluate our algorithm on two representative structure inference tasks: probabilistic matrix decomposition and symbolic program parsing. It outperforms data-driven and search-based alternatives on both tasks.} }
Endnote
%0 Conference Paper %T Neurally-Guided Structure Inference %A Sidi Lu %A Jiayuan Mao %A Joshua Tenenbaum %A Jiajun Wu %B Proceedings of the 36th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2019 %E Kamalika Chaudhuri %E Ruslan Salakhutdinov %F pmlr-v97-lu19b %I PMLR %P 4144--4153 %U https://proceedings.mlr.press/v97/lu19b.html %V 97 %X Most structure inference methods either rely on exhaustive search or are purely data-driven. Exhaustive search robustly infers the structure of arbitrarily complex data, but it is slow. Data-driven methods allow efficient inference, but do not generalize when test data have more complex structures than training data. In this paper, we propose a hybrid inference algorithm, the Neurally-Guided Structure Inference (NG-SI), keeping the advantages of both search-based and data-driven methods. The key idea of NG-SI is to use a neural network to guide the hierarchical, layer-wise search over the compositional space of structures. We evaluate our algorithm on two representative structure inference tasks: probabilistic matrix decomposition and symbolic program parsing. It outperforms data-driven and search-based alternatives on both tasks.
APA
Lu, S., Mao, J., Tenenbaum, J. & Wu, J.. (2019). Neurally-Guided Structure Inference. Proceedings of the 36th International Conference on Machine Learning, in Proceedings of Machine Learning Research 97:4144-4153 Available from https://proceedings.mlr.press/v97/lu19b.html.

Related Material