23 |
Higher-order Derivatives of Weighted Finite-state Machines ...
|
|
|
|
BASE
|
|
Show details
|
|
25 |
A surprisal--duration trade-off across and within the world's languages ...
|
|
|
|
BASE
|
|
Show details
|
|
31 |
What About the Precedent: An Information-Theoretic Analysis of Common Law ...
|
|
|
|
BASE
|
|
Show details
|
|
35 |
Examining the Inductive Bias of Neural Language Models with Artificial Languages ...
|
|
|
|
BASE
|
|
Show details
|
|
36 |
Finding Concept-specific Biases in Form–Meaning Associations ...
|
|
|
|
BASE
|
|
Show details
|
|
39 |
Efficient computation of expectations under spanning tree distributions ...
|
|
|
|
BASE
|
|
Show details
|
|
40 |
Multimodal pretraining unmasked: A meta-analysis and a unified framework of vision-and-language berts ...
|
|
|
|
Abstract:
Large-scale pretraining and task-specific fine-tuning is now the standard methodology for many tasks in computer vision and natural language processing. Recently, a multitude of methods have been proposed for pretraining vision and language BERTs to tackle challenges at the intersection of these two key areas of AI. These models can be categorized into either single-stream or dual-stream encoders. We study the differences between these two categories, and show how they can be unified under a single theoretical framework. We then conduct controlled experiments to discern the empirical differences between five vision and language BERTs. Our experiments show that training data and hyperparameters are responsible for most of the differences between the reported results, but they also reveal that the embedding layer plays a crucial role in these massive models. ... : Transactions of the Association for Computational Linguistics, 9 ...
|
|
URL: http://hdl.handle.net/20.500.11850/517633 https://dx.doi.org/10.3929/ethz-b-000517633
|
|
BASE
|
|
Hide details
|
|
|
|