Luke Zettlemoyer

TL;DR: This paper presents a new dataset, collected to describe Xbox avatars, as well as models for learning the relationships between these avatars and their literal and sentimental descriptions, and demonstrates that sentimental language provides a concise (though noisy) means of specifying low-level visual properties.

TL;DR: The FaVIQ dataset as discussed by the authors uses information-seeking questions posed by real users who do not know how to answer, which enables automatically constructing true and false claims that reflect confusions arisen from users (e.g., the year of the movie being filmed vs. being released).

TL;DR: This paper proposed FEWS (Few-shot Examples of Word Senses), a new low-shot WSD dataset automatically extracted from example sentences in Wiktionary, which provides high sense coverage across different natural language domains and provides a large training set that covers many more senses than previous datasets and a comprehensive evaluation set containing few- and zero-shot examples of a wide variety of senses.

TL;DR: The DESCGEN dataset as discussed by the authors consists of 37k entity descriptions from Wikipedia and Fandom, each paired with nine evidence documents on average, and the documents were collected using a combination of entity linking and hyperlinks to the Wikipedia entity pages, which together provided high-quality distant supervision.

TL;DR: The authors combine unsupervised bitext mining and word alignment to improve the quality of bilingual lexicons, achieving the state-of-the-art performance on the BUCC 2020 shared task.