Papers by Daniel Spokoyny
Masked Measurement Prediction: Learning to Jointly Predict Quantities and Units from Textual Context (2022.findings-naacl)
Copied to clipboard
| Challenge: | Current benchmarks do not evaluate numeracy of pretraining language models on measurements. |
| Approach: | They propose a new task where a model learns to reconstruct a number with its associated unit given masked text. |
| Outcome: | The proposed model significantly underperforms pre-trained model with baselines and ablations. |
An Empirical Investigation of Contextualized Number Prediction (2020.emnlp-main)
Copied to clipboard
| Challenge: | a large scale empirical investigation of contextualized number prediction in running text is needed. |
| Approach: | They propose a suite of output distribution parameterizations that incorporate latent variables to add expressivity and better fit the natural distribution of numeric values in running text. |
| Outcome: | The proposed models outperform flow-based models on two numeric datasets in the financial and scientific domain. |