J
João Sedoc
Researcher at New York University
Publications - 59
Citations - 874
João Sedoc is an academic researcher from New York University. The author has contributed to research in topics: Sentence & Empathy. The author has an hindex of 13, co-authored 59 publications receiving 540 citations. Previous affiliations of João Sedoc include Johns Hopkins University & University of Pennsylvania.
Papers
More filters
Proceedings ArticleDOI
Comparison of Diverse Decoding Methods from Conditional Language Models
TL;DR: This article presented a survey of decoding-time strategies for generating diverse outputs from a conditional language model, where over-sample candidates, then use clustering to remove similar sequences, thus achieving high diversity without sacrificing quality.
Proceedings ArticleDOI
ChatEval: A Tool for Chatbot Evaluation.
TL;DR: A unified framework for human evaluation of chatbots that augments existing tools and provides a web-based hub for researchers to share and compare their dialog systems and open-source baseline models and evaluation datasets are introduced.
Proceedings ArticleDOI
Complexity-Weighted Loss and Diverse Reranking for Sentence Simplification
Reno Kriz,João Sedoc,Marianna Apidianaki,Carolina Zheng,Gaurav Kumar,Eleni Miltsakaki,Chris Callison-Burch +6 more
TL;DR: The authors incorporate content word complexities, as predicted with a leveled word complexity model, into the loss function during training and generate a large set of diverse candidate simplifications at test time, and rerank these to promote fluency, adequacy, and simplicity.
Proceedings ArticleDOI
Modeling Empathy and Distress in Reaction to News Stories.
TL;DR: The authors presented the first publicly available gold standard for text-based empathy prediction, which is constructed using a novel annotation methodology which reliably captures empathy assessments by the writer of a statement using multi-item scales.
Proceedings ArticleDOI
Continual Learning for Sentence Representations Using Conceptors
TL;DR: The authors propose to initialize sentence encoders with the help of corpus-independent features, and then sequentially update sentence encoder using Boolean operations of conceptor matrices to learn corpus-dependent features.