Attentive Pooling Networks

TL;DR

This paper introduces Attentive Pooling, a two-way attention mechanism that enhances neural network models for pair-wise tasks by allowing mutual influence of input representations, leading to improved performance across multiple benchmarks.

Contribution

It presents a novel two-way attention pooling method applicable to CNNs and RNNs, significantly improving performance in question answering and answer selection tasks.

Findings

Outperforms strong baselines on three benchmark tasks

Achieves state-of-the-art results in all tested benchmarks

Demonstrates effectiveness of two-way attention in neural models

Abstract

In this work, we propose Attentive Pooling (AP), a two-way attention mechanism for discriminative model training. In the context of pair-wise ranking or classification with neural networks, AP enables the pooling layer to be aware of the current input pair, in a way that information from the two input items can directly influence the computation of each other's representations. Along with such representations of the paired inputs, AP jointly learns a similarity measure over projected segments (e.g. trigrams) of the pair, and subsequently, derives the corresponding attention vector for each input to guide the pooling. Our two-way attention mechanism is a general framework independent of the underlying representation learning, and it has been applied to both convolutional neural networks (CNNs) and recurrent neural networks (RNNs) in our studies. The empirical results, from three very different benchmark tasks of question answering/answer selection, demonstrate that our proposed models outperform a variety of strong baselines and achieve state-of-the-art performance in all the benchmarks.