Model Comparison for Semantic Grouping

Francisco Vargas, Kamen Brestnichki, Nils Hammerla
Proceedings of the 36th International Conference on Machine Learning, PMLR 97:6410-6417, 2019.

Abstract

We introduce a probabilistic framework for quantifying the semantic similarity between two groups of embeddings. We formulate the task of semantic similarity as a model comparison task in which we contrast a generative model which jointly models two sentences versus one that does not. We illustrate how this framework can be used for the Semantic Textual Similarity tasks using clear assumptions about how the embeddings of words are generated. We apply model comparison that utilises information criteria to address some of the shortcomings of Bayesian model comparison, whilst still penalising model complexity. We achieve competitive results by applying the proposed framework with an appropriate choice of likelihood on the STS datasets.

Cite this Paper

BibTeX
@InProceedings{pmlr-v97-vargas19a, title = {Model Comparison for Semantic Grouping}, author = {Vargas, Francisco and Brestnichki, Kamen and Hammerla, Nils}, booktitle = {Proceedings of the 36th International Conference on Machine Learning}, pages = {6410--6417}, 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/vargas19a/vargas19a.pdf}, url = {https://proceedings.mlr.press/v97/vargas19a.html}, abstract = {We introduce a probabilistic framework for quantifying the semantic similarity between two groups of embeddings. We formulate the task of semantic similarity as a model comparison task in which we contrast a generative model which jointly models two sentences versus one that does not. We illustrate how this framework can be used for the Semantic Textual Similarity tasks using clear assumptions about how the embeddings of words are generated. We apply model comparison that utilises information criteria to address some of the shortcomings of Bayesian model comparison, whilst still penalising model complexity. We achieve competitive results by applying the proposed framework with an appropriate choice of likelihood on the STS datasets.} }
Endnote
%0 Conference Paper %T Model Comparison for Semantic Grouping %A Francisco Vargas %A Kamen Brestnichki %A Nils Hammerla %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-vargas19a %I PMLR %P 6410--6417 %U https://proceedings.mlr.press/v97/vargas19a.html %V 97 %X We introduce a probabilistic framework for quantifying the semantic similarity between two groups of embeddings. We formulate the task of semantic similarity as a model comparison task in which we contrast a generative model which jointly models two sentences versus one that does not. We illustrate how this framework can be used for the Semantic Textual Similarity tasks using clear assumptions about how the embeddings of words are generated. We apply model comparison that utilises information criteria to address some of the shortcomings of Bayesian model comparison, whilst still penalising model complexity. We achieve competitive results by applying the proposed framework with an appropriate choice of likelihood on the STS datasets.
APA
Vargas, F., Brestnichki, K. & Hammerla, N.. (2019). Model Comparison for Semantic Grouping. Proceedings of the 36th International Conference on Machine Learning, in Proceedings of Machine Learning Research 97:6410-6417 Available from https://proceedings.mlr.press/v97/vargas19a.html.

Related Material