Sparse and structured neural attention mechanisms
Switch branches/tags
Clone or download
Latest commit dc6e0e9 Nov 6, 2017

README.md

Sparse and structured attention mechanisms

Build Status PyPI version


Efficient implementation of structured sparsity inducing attention mechanisms: fusedmax, oscarmax and sparsemax.

Currently available for pytorch v0.2. Requires python (3.6, 3.5, or 2.7), cython, numpy, scipy, scikit-learn, and lightning.

For details, check out our paper:

Vlad Niculae and Mathieu Blondel A Regularized Framework for Sparse and Structured Neural Attention In: Proceedings of NIPS, 2017. https://arxiv.org/abs/1705.07704

See also:

André F. T. Martins and Ramón Fernandez Astudillo From Softmax to Sparsemax: A Sparse Model of Attention and Multi-Label Classification In: Proceedings of ICML, 2016 https://arxiv.org/abs/1602.02068

X. Zeng and M. Figueiredo, The ordered weighted L1 norm: Atomic formulation, dual norm, and projections. eprint http://arxiv.org/abs/1409.4271