Both scientists and children make important structural discoveries, yet their computational underpinnings are not well understood. Structure discovery has previously been formalized as probabilistic inference about the right structural form—where form could be a tree, ring, chain, grid, etc. (Kemp & Tenenbaum, 2008). Although this approach can learn intuitive organizations, including a tree for animals and a ring for the color circle, it assumes a strong inductive bias that considers only these particular forms, and each form is explicitly provided as initial knowledge. Here we introduce a new computational model of how organizing structure can be discovered, utilizing a broad hypothesis space with a preference for sparse connectivity. Given that the inductive bias is more general, the model’s initial knowledge shows little qualitative resemblance to some of the discoveries it supports. As a consequence, the model can also learn complex structures for domains that lack intuitive description, as well as predict human property induction judgments without explicit structural forms. By allowing form to emerge from sparsity, our approach clarifies how both the richness and flexibility of human conceptual organization can coexist.

@Article{efficient-inference-for-sparse-latent-variable-models-of-transcriptional-regulation,
title = {The Emergence of Organizing Efficient inference for sparse latent variable models of transcriptional regulation},
journal = {Bioinformatics},
author = {Zhenwen Dai and Mudassar Iqbal and Neil D. Lawrence and Magnus Rattray},
year = {2017},
month = {00},
edit = {https://github.com/lawrennd//publications/edit/gh-pages/_posts/2017-08-26-efficient-inference-for-sparse-latent-variable-models-of-transcriptional-regulation.md},
url = {http://inverseprobability.com/publications/efficient-inference-for-sparse-latent-variable-models-of-transcriptional-regulation.html},
abstract = {Both scientists and children make important structural discoveries, yet their computational underpinnings are not well understood. Structure discovery has previously been formalized as probabilistic inference about the right structural form—where form could be a tree, ring, chain, grid, etc. (Kemp & Tenenbaum, 2008). Although this approach can learn intuitive organizations, including a tree for animals and a ring for the color circle, it assumes a strong inductive bias that considers only these particular forms, and each form is explicitly provided as initial knowledge. Here we introduce a new computational model of how organizing structure can be discovered, utilizing a broad hypothesis space with a preference for sparse connectivity. Given that the inductive bias is more general, the model's initial knowledge shows little qualitative resemblance to some of the discoveries it supports. As a consequence, the model can also learn complex structures for domains that lack intuitive description, as well as predict human property induction judgments without explicit structural forms. By allowing form to emerge from sparsity, our approach clarifies how both the richness and flexibility of human conceptual organization can coexist.},
key = {Dai:sparse18},
doi = {10.1093/bioinformatics/btx508},
linkpdf = {https://academic.oup.com/bioinformatics/article-pdf/33/23/3776/22031627/btx508.pdf},
OPTgroup = {}
}

%T The Emergence of Organizing Efficient inference for sparse latent variable models of transcriptional regulation
%A Zhenwen Dai and Mudassar Iqbal and Neil D. Lawrence and Magnus Rattray
%B
%C Bioinformatics
%D
%F efficient-inference-for-sparse-latent-variable-models-of-transcriptional-regulation
%J Bioinformatics
%P --
%R 10.1093/bioinformatics/btx508
%U http://inverseprobability.com/publications/efficient-inference-for-sparse-latent-variable-models-of-transcriptional-regulation.html
%X Both scientists and children make important structural discoveries, yet their computational underpinnings are not well understood. Structure discovery has previously been formalized as probabilistic inference about the right structural form—where form could be a tree, ring, chain, grid, etc. (Kemp & Tenenbaum, 2008). Although this approach can learn intuitive organizations, including a tree for animals and a ring for the color circle, it assumes a strong inductive bias that considers only these particular forms, and each form is explicitly provided as initial knowledge. Here we introduce a new computational model of how organizing structure can be discovered, utilizing a broad hypothesis space with a preference for sparse connectivity. Given that the inductive bias is more general, the model's initial knowledge shows little qualitative resemblance to some of the discoveries it supports. As a consequence, the model can also learn complex structures for domains that lack intuitive description, as well as predict human property induction judgments without explicit structural forms. By allowing form to emerge from sparsity, our approach clarifies how both the richness and flexibility of human conceptual organization can coexist.

TY - CPAPER
TI - The Emergence of Organizing Efficient inference for sparse latent variable models of transcriptional regulation
AU - Zhenwen Dai
AU - Mudassar Iqbal
AU - Neil D. Lawrence
AU - Magnus Rattray
PY - 2017/08/26
DA - 2017/08/26
ID - efficient-inference-for-sparse-latent-variable-models-of-transcriptional-regulation
SP -
EP -
DO - 10.1093/bioinformatics/btx508
L1 - https://academic.oup.com/bioinformatics/article-pdf/33/23/3776/22031627/btx508.pdf
UR - http://inverseprobability.com/publications/efficient-inference-for-sparse-latent-variable-models-of-transcriptional-regulation.html
AB - Both scientists and children make important structural discoveries, yet their computational underpinnings are not well understood. Structure discovery has previously been formalized as probabilistic inference about the right structural form—where form could be a tree, ring, chain, grid, etc. (Kemp & Tenenbaum, 2008). Although this approach can learn intuitive organizations, including a tree for animals and a ring for the color circle, it assumes a strong inductive bias that considers only these particular forms, and each form is explicitly provided as initial knowledge. Here we introduce a new computational model of how organizing structure can be discovered, utilizing a broad hypothesis space with a preference for sparse connectivity. Given that the inductive bias is more general, the model's initial knowledge shows little qualitative resemblance to some of the discoveries it supports. As a consequence, the model can also learn complex structures for domains that lack intuitive description, as well as predict human property induction judgments without explicit structural forms. By allowing form to emerge from sparsity, our approach clarifies how both the richness and flexibility of human conceptual organization can coexist.
ER -

Dai, Z., Iqbal, M., Lawrence, N.D. & Rattray, M.. (2017). The Emergence of Organizing Efficient inference for sparse latent variable models of transcriptional regulation. Bioinformatics :-