ChordMixer: A Scalable Neural Attention Model for Sequences with Different Lengths
This addresses a scalability issue in neural attention for domains with long or variable-length sequences, offering a novel method that could benefit fields like natural language processing and bioinformatics.
The authors tackled the problem of modeling attention for sequences of varying lengths, proposing ChordMixer, a neural network building block that outperforms existing neural attention models on tasks like long document classification and DNA sequence-based taxonomy classification.
Sequential data naturally have different lengths in many domains, with some very long sequences. As an important modeling tool, neural attention should capture long-range interaction in such sequences. However, most existing neural attention models admit only short sequences, or they have to employ chunking or padding to enforce a constant input length. Here we propose a simple neural network building block called ChordMixer which can model the attention for long sequences with variable lengths. Each ChordMixer block consists of a position-wise rotation layer without learnable parameters and an element-wise MLP layer. Repeatedly applying such blocks forms an effective network backbone that mixes the input signals towards the learning targets. We have tested ChordMixer on the synthetic adding problem, long document classification, and DNA sequence-based taxonomy classification. The experiment results show that our method substantially outperforms other neural attention models.