M
Melissa Dell
Researcher at Harvard University
Publications - 41
Citations - 5790
Melissa Dell is an academic researcher from Harvard University. The author has contributed to research in topics: Productivity & Computer science. The author has an hindex of 18, co-authored 36 publications receiving 4518 citations. Previous affiliations of Melissa Dell include National Bureau of Economic Research & Massachusetts Institute of Technology.
Papers
More filters
Posted Content
LayoutParser: A Unified Toolkit for Deep Learning Based Document Image Analysis.
Zejiang Shen,Ruochen Zhang,Melissa Dell,Benjamin Charles Germain Lee,Jacob C. Carlson,Weining Li +5 more
TL;DR: The layoutparser library as mentioned in this paper provides a set of simple and intuitive interfaces for applying and customizing deep learning models for layout detection, character recognition, and many other document processing tasks.
Journal ArticleDOI
Efficient OCR for Building a Diverse Digital History
Jacob Carlson,Melissa Dell +1 more
TL;DR: In this paper , the authors model OCR as a character level image retrieval problem, using a contrastively trained vision encoder, which is more sample efficient and extensible than existing architectures, enabling accurate OCR in settings where existing solutions fail.
Journal ArticleDOI
Linking Representations with Multimodal Contrastive Learning
TL;DR: Li et al. as mentioned in this paper proposed a multimodal framework for record linkage, which employs end-to-end training of symmetric vision and language bi-encoders, aligned through contrastive language-image pre-training, to learn a metric space where the pooled imagetext representation for a given instance is close to representations in the same class and distant from representations in different classes.
Posted Content
The Historical State, Local Collective Action, and Economic Development in Vietnam
TL;DR: This paper examined how the historical state conditions long-run development, using Vietnam as a laboratory and found that areas historically under a strong state have higher living standards today and better economic outcomes over the past 150 years.
Posted Content
OLALA: Object-Level Active Learning Based Layout Annotation.
TL;DR: This work introduces an Object-Level Active Learning based Layout Annotation framework, OLALA, which includes an object scoring method and a prediction correction algorithm that selects only the most ambiguous object prediction regions within an image for annotators to label, optimizing the use of the annotation budget.