Full talk · edited transcript

Data Labeling: Experts, Crowdsourcing and Large Language Models

Maria Molchanova25 May 2024

Maria Molchanova, researcher and analyst, Laboratory of Neural Systems and Deep Learning, MIPT.

Three approaches to annotation (experts, crowdsourcers, large language models) are demonstrated on one task from the laboratory. Article on the research: Ostyakova, Smilga, Petukhova, Molchanova, Kornev, «ChatGPT vs. Crowdsourcing vs. Experts: Annotating Open-Domain Conversations with Speech Functions», SIGDIAL 2023 (24th Meeting of the Special Interest Group on Discourse and Dialogue), pp. 242–254, https://aclanthology.org/2023.sigdial-1.23. The main focus of the article is large language models; it contains prompts and a link to the annotation code: https://github.com/deeppavlov/sf_corpus/.

Continue reading — it's free

Google or a one-time email code. No password.