HyperMixer: An MLP-based Low Cost Alternative to Transformers
This work addresses the problem of high resource requirements for NLP models, offering a more efficient alternative for practitioners, though it is incremental as it builds on existing MLP-based architectures.
The paper tackles the high computational and data costs of Transformers for natural language understanding by proposing HyperMixer, an MLP-based variant that uses hypernetworks for dynamic token mixing, achieving performance on par with Transformers at substantially lower costs in processing time, training data, and hyperparameter tuning.
Transformer-based architectures are the model of choice for natural language understanding, but they come at a significant cost, as they have quadratic complexity in the input length, require a lot of training data, and can be difficult to tune. In the pursuit of lower costs, we investigate simple MLP-based architectures. We find that existing architectures such as MLPMixer, which achieves token mixing through a static MLP applied to each feature independently, are too detached from the inductive biases required for natural language understanding. In this paper, we propose a simple variant, HyperMixer, which forms the token mixing MLP dynamically using hypernetworks. Empirically, we demonstrate that our model performs better than alternative MLP-based models, and on par with Transformers. In contrast to Transformers, HyperMixer achieves these results at substantially lower costs in terms of processing time, training data, and hyperparameter tuning.