LGMLNov 1, 2021

Kernel Deformed Exponential Families for Sparse Continuous Attention

arXiv:2111.01222v2
Originality Incremental advance
AI Analysis

This work provides a method for improving attention mechanisms in machine learning, but it is incremental as it builds on prior continuous attention frameworks.

The paper tackles the problem of creating flexible and sparse continuous attention mechanisms by introducing kernel deformed exponential families, which can attend to multiple compact regions of the data domain.

Attention mechanisms take an expectation of a data representation with respect to probability weights. This creates summary statistics that focus on important features. Recently, (Martins et al. 2020, 2021) proposed continuous attention mechanisms, focusing on unimodal attention densities from the exponential and deformed exponential families: the latter has sparse support. (Farinhas et al. 2021) extended this to use Gaussian mixture attention densities, which are a flexible class with dense support. In this paper, we extend this to two general flexible classes: kernel exponential families and our new sparse counterpart kernel deformed exponential families. Theoretically, we show new existence results for both kernel exponential and deformed exponential families, and that the deformed case has similar approximation capabilities to kernel exponential families. Experiments show that kernel deformed exponential families can attend to multiple compact regions of the data domain.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes