• Publications
  • Influence
SCROLLS: Standardized CompaRison Over Long Language Sequences
This work introduces SCROLLS, a suite of tasks that require reasoning over long texts, and examines existing long-text datasets, and handpick ones where the text is naturally long, while prioritizing tasks that involve synthesizing information across the input.
Scene Graph tO Image Generation with Contextualized Object Layout Refinement
This work proposes a method that alleviates generated images with high inter-object overlap, empty areas, blurry objects, and overall compromised quality by generating all object layouts together and reducing the reliance on supervised learning.
Achieving Model Robustness through Discrete Adversarial Training
Surprisingly, it is found that random sampling leads to impressive gains in robustness, outperforming the commonly-used offline augmentation, while leading to a speedup at training time of ~10x.
Scaling Laws Under the Microscope: Predicting Transformer Performance from Small Scale Experiments
This work finds that scaling laws emerge at finetuning time in some NLP tasks, and that they can also be exploited for debugging convergence when training large models.
Beyond Importance Scores: Interpreting Tabular ML by Visualizing Feature Semantics
This work introduces Feature Vectors, a new global interpretability method designed for tabular datasets that discovers the inherent semantic relationship among features via an intuitive feature visualization technique.