Preeti
The 2024 Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL) took place in Mexico City, where researchers from USC’s Viterbi School of Engineering presented work in various research tracks. The conference, organized by NAACL, accepted 565 out of 2434 submissions, showcasing the latest in natural language research.
One research spotlight focused on using language models to improve online conversations by generating response recommendations for conversational moderation. Another study explored the use of spaced repetition to help machine learning models retain information learned in different languages. Additionally, researchers introduced a method called span-and-relation parsing to enhance AI’s understanding of legal texts, leading to the creation of a dataset named LegalDiscourse.
Other topics covered at the conference included training models to identify objects based on descriptive language, investigating how models memorize information, analyzing over-specification in language models, and enhancing comparative knowledge processing in AI. Furthermore, researchers developed a method for efficient visual document understanding in infographics without relying on expensive tools like optical character recognition.
USC researchers presented a range of papers at the conference, covering topics such as language model moderation, continual learning, argument quality assessment, and rule-based machine translation for low-resource languages. These studies contribute to the advancement of AI technology and its applications across various domains.
Source link
Source link: https://viterbischool.usc.edu/news/2024/06/usc-at-the-naacl-24-conference/
GIPHY App Key not set. Please check settings