Will we run out of data? Limits of LLM scaling based on human-generated data
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2022
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916274003509248 |
|---|---|
| author | Villalobos, Pablo Ho, Anson Sevilla, Jaime Besiroglu, Tamay Heim, Lennart Hobbhahn, Marius |
| author_facet | Villalobos, Pablo Ho, Anson Sevilla, Jaime Besiroglu, Tamay Heim, Lennart Hobbhahn, Marius |
| contents | We investigate the potential constraints on LLM scaling posed by the availability of public human-generated text data. We forecast the growing demand for training data based on current trends and estimate the total stock of public human text data. Our findings indicate that if current LLM development trends continue, models will be trained on datasets roughly equal in size to the available stock of public human text data between 2026 and 2032, or slightly earlier if models are overtrained. We explore how progress in language modeling can continue when human-generated text datasets cannot be scaled any further. We argue that synthetic data generation, transfer learning from data-rich domains, and data efficiency improvements might support further progress. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2211_04325 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | Will we run out of data? Limits of LLM scaling based on human-generated data Villalobos, Pablo Ho, Anson Sevilla, Jaime Besiroglu, Tamay Heim, Lennart Hobbhahn, Marius Machine Learning Artificial Intelligence Computation and Language Computer Vision and Pattern Recognition Computers and Society We investigate the potential constraints on LLM scaling posed by the availability of public human-generated text data. We forecast the growing demand for training data based on current trends and estimate the total stock of public human text data. Our findings indicate that if current LLM development trends continue, models will be trained on datasets roughly equal in size to the available stock of public human text data between 2026 and 2032, or slightly earlier if models are overtrained. We explore how progress in language modeling can continue when human-generated text datasets cannot be scaled any further. We argue that synthetic data generation, transfer learning from data-rich domains, and data efficiency improvements might support further progress. |
| title | Will we run out of data? Limits of LLM scaling based on human-generated data |
| topic | Machine Learning Artificial Intelligence Computation and Language Computer Vision and Pattern Recognition Computers and Society |
| url | https://arxiv.org/abs/2211.04325 |