Scholar
Yuning Mao
Google Scholar ID: steJe6IAAAAJ
Meta Superintelligence Labs
Citations & Impact
Citations
33,726
H-index
27
i10-index
30
Publications
20
Co-authors
14
list available
Publications
Resume (English only)
Academic Achievements
- Released Llama 3, Llama Guard 2, and Meta AI assistant powered by Llama 3; Released Llama Guard for LLM Input-Output Safetyguard; Released Meta AI assistant in Instagram, WhatsApp, and Messenger; Released Llama 2, a collection of pretrained and fine-tuned LLMs; Released LIMA, a LLaMA-based model fine-tuned on 1,000 curated prompts + responses; Released CiteSum, a new summarization dataset based on citation texts; Published a method to get free gains for openQA by reranking retrieved passages without any training; Published a method to use constrained generation to preserve factual consistency in abstractive summarization; Achieved SOTA on Natural Questions and TriviaQA using BM25; Two papers on multi-document summarization and knowledge graph reasoning accepted to EMNLP 2020; Two papers on self-supervised taxonomy enrichment and knowledge collection of product knowledge graph accepted to KDD 2020; Introduced FAR, a new metric for summarization, along with a detailed analysis of CNN/Daily Mail, accepted to ACL 2020; Released NewSHead, the first dataset on multi-document headline generation, along with the NHNet model in official TensorFlow repo.
Research Experience
- Research Scientist at Meta Superintelligence Labs, working on developing large language models (Llama series). Founding member of Meta GenAI org, recently focused on LLM for code generation, specifically generating websites, animations, and games. Previously, a core contributor to Llama 2 and Llama 3, working on post-training.
Education
- Ph.D. in Computer Science from the University of Illinois at Urbana–Champaign (UIUC), under the supervision of Prof. Jiawei Han; Bachelor's degree from the IEEE Class, Shanghai Jiao Tong University (SJTU).
Background
- Research Interests: Helping humans acquire information and knowledge more effectively and efficiently. Professional Field: Natural Language Processing and Generative AI.
Miscellany
- Enjoys coding to better his (and others') life, loves both engineering work and research-oriented projects. Loves music, enjoys singing privately and publicly, plays piano and guitar, and composes music.
