Σ-GPTs: A New Approach to Autoregressive Models
Abstract
Autoregressive models, such as the GPT family, use a fixed order, usually left-to-right, to generate sequences. However, this is not a necessity. In this paper, we challenge this assumption and show that by simply adding a positional encoding for the output, this order can be modulated on-the-fly per-sample which offers key advantageous properties. It allows for the sampling of and conditioning on arbitrary subsets of tokens, and it also allows sampling in one shot multiple tokens dynamically according to a rejection strategy, leading to a sub-linear number of model evaluations. We evaluate our method across various domains, including language modeling, path-solving, and aircraft vertical rate prediction, decreasing the number of steps required for generation by an order of magnitude (The code of this work is available at https://github.com/idiap/sigma-gpt ).
Cite
Text
Pannatier et al. "Σ-GPTs: A New Approach to Autoregressive Models." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2024. doi:10.1007/978-3-031-70368-3_9Markdown
[Pannatier et al. "Σ-GPTs: A New Approach to Autoregressive Models." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2024.](https://mlanthology.org/ecmlpkdd/2024/pannatier2024ecmlpkdd-gpts/) doi:10.1007/978-3-031-70368-3_9BibTeX
@inproceedings{pannatier2024ecmlpkdd-gpts,
title = {{Σ-GPTs: A New Approach to Autoregressive Models}},
author = {Pannatier, Arnaud and Courdier, Evann and Fleuret, François},
booktitle = {European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases},
year = {2024},
pages = {143-159},
doi = {10.1007/978-3-031-70368-3_9},
url = {https://mlanthology.org/ecmlpkdd/2024/pannatier2024ecmlpkdd-gpts/}
}