A State-Vector Framework for Dataset Effects
Sahak, Esmat, Zhu, Zining, Rudzicz, Frank
–arXiv.org Artificial Intelligence
The impressive success of recent deep neural network (DNN)-based systems is significantly influenced by the high-quality datasets used in training. However, the effects of the datasets, especially how they interact with each other, remain underexplored. We propose a state-vector framework to enable rigorous studies in this direction. This framework uses idealized probing test results as the bases of a vector space. This framework allows us to quantify the effects of both standalone and interacting datasets. We show that the significant effects of some commonly-used language understanding datasets are characteristic and are concentrated on a few linguistic dimensions. Additionally, we observe some ``spill-over'' effects: the datasets could impact the models along dimensions that may seem unrelated to the intended tasks. Our state-vector framework paves the way for a systematic understanding of the dataset effects, a crucial component in responsible and robust model development.
arXiv.org Artificial Intelligence
Oct-16-2023
- Country:
- North America
- Dominican Republic (0.04)
- United States
- Texas (0.04)
- Washington > King County
- Seattle (0.04)
- Minnesota > Hennepin County
- Minneapolis (0.14)
- Canada > Ontario
- Toronto (0.14)
- Europe
- Asia
- Middle East > Jordan (0.04)
- China > Hong Kong (0.04)
- Japan > Kyūshū & Okinawa
- Kyūshū > Miyazaki Prefecture > Miyazaki (0.04)
- North America
- Genre:
- Research Report > Experimental Study (0.48)
- Technology: