EMNLP 2022main12 citations

Meta-Learning Fast Weight Language Models

Kevin Clark, Kelvin Guu, Ming-Wei Chang, Panupong Pasupat, Geoffrey Hinton, Mohammad Norouzi

Abstract

Dynamic evaluation of language models (LMs) adapts model parameters at test time using gradient information from previous tokens and substantially improves LM performance. However, it requires over 3x more compute than standard inference. We present Fast Weight Layers (FWLs), a neural component that provides the benefits of dynamic evaluation much more efficiently by expressing gradient updates as linear attention. A key improvement over dynamic evaluation is that FWLs can also be applied at training time, so the model learns to make good use of gradient updates. FWLs can easily be added on top of existing transformer models, require relatively little extra compute or memory to run, and significantly improve language modeling perplexity.

BibTeX
@inproceedings{clark-etal-2022-meta,
    title = "Meta-Learning Fast Weight Language Models",
    author = "Clark, Kevin  and
      Guu, Kelvin  and
      Chang, Ming-Wei  and
      Pasupat, Panupong  and
      Hinton, Geoffrey  and
      Norouzi, Mohammad",
    editor = "Goldberg, Yoav  and
      Kozareva, Zornitsa  and
      Zhang, Yue",
    booktitle = "Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing",
    month = dec,
    year = "2022",
    address = "Abu Dhabi, United Arab Emirates",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2022.emnlp-main.661/",
    doi = "10.18653/v1/2022.emnlp-main.661",
    pages = "9751--9757"
}
Meta-Learning Fast Weight Language Models · EMNLP 2022