Yequan Wang (王业全)

Researcher, Team Leader

Beijing Academy of Artificial Intelligence (BAAI)

Biography

Dr. Wang Yequan is a researcher at Beijing Academy of Artificial Intelligence (BAAI). His research interests mainly include large models. He leads the Recognition Team, aiming to build lower-cost but more powerful large models and conduct relevant large model research, eventually achieving the goal of AGI.

Dr. Wang received Ph.D. in Computer Science from Tsinghua University, supervised by Prof. Xiaoyan Zhu and Minlie Huang. From Sep. 2017 to Sep. 2018, Dr. Wang studied at Nanyang Technological University as a Joint Ph.D. Candidate, supervised by Associate Prof. Aixin Sun, who is also the Assistant Chair (Academic).

Dr. Wang has been recognized as the 2022 AI 2000 Most Influential Scholar Honorable Mention in Natural Language Processing.

From November 2022 to November 2025, serving as the PI for the ‘Next-Generation Artificial Intelligence’ National Key R&D Program.

Team Philosophy for Large Models：Both system capabilities and research capabilities are essential. Without system capabilities, it is not possible to develop large models. Without research capabilities, one can only follow in the footsteps of others. In the case where the leader in large model development chooses to close the source code, it becomes impossible to make further breakthroughs.

Google Scholar Citations: 3,500+

ORCID: 0000-0001-7530-6125

Interests

Large Model
Emboddied AI
Natural Language Processing

Education

PhD in Artificial Intelligence, 2014-2019

Tsinghua University, Beijing
Joint PhD Program, 2017-2018

Nanyang Technological University, Singapore

Projects

FLM Family

The FLM series of large models are a set of large models completed by the Cofe-AI team, including FLM-2, FLM-101B, and FreeLM. The core technology mainly consists of growth techniques, loss prediction technologies, and the FreeLM framework, among others.

Arabic Language Model (ALM 1.0)

We develop and open Arabic Language Model (ALM).

Artificial Intelligence Fundamental Model Support Platform and Evaluation Technology (National Key R&D Program)

This project focuses on the development of large models, with the aim of promoting the advancement of cutting-edge technologies for ultra-large-scale intelligent models in China and facilitating the empowerment of artificial intelligence for economic and social development. It aims to build an internationally leading open-source technology system for fundamental models and establish an open innovation ecosystem centered around foundational models in artificial intelligence.

Implicit Sentiment Analysis on Complicated Web Text

National Science Foundation of China (NSFC, 62106249)

Selected Publications

Masked Structural Growth for 2x Faster Language Model Pre-training

To lower the computional cost of training large model, we focus on speeding up pre-training by progressively growing from a small Transformer structure to a large one.

Yiqun Yao, Zheng Zhang, Jing Li, Yequan Wang

52B to 1T: Lessons Learned via Tele-FLM Series

As scaling laws underscore the potential of increasing model sizes, the academic community has intensified its investigations into LLMs with capacities exceeding 50 billion parameters. This technical report builds on our prior work with Tele-FLM (also known as FLM-2), a publicly available 52-billion-parameter model.

Xiang Li, Yiqun Yao, Xin Jiang, Xuezhi Fang, China Telecom, Yequan Wang, Zhongjiang He, Zhongyuan Wang, Xuelong Li, Tiejun Huang

Research without Re-search: Maximal Update Parametrization Yields Accurate Loss Prediction across Scales

We find that Maximal Update parametrization (uP) enables accurate fitting of scaling laws for hyperparameters close to common loss basins, without any search. Thus, different models can be directly compared on large scales with loss prediction even before the training starts. We propose a new paradigm as a first step towards reliable academic research for any model scale without heavy computation.

Yiqun Yao, Yequan Wang

Research without Re-search: Maximal Update Parametrization Yields Accurate Loss Prediction across Scales

Sentiment Analysis by Capsules

In this paper, we propose RNN-Capsule, a capsule model based on Recurrent Neural Network (RNN) for sentiment analysis.

Yequan Wang, Aixin Sun, Jialong Han, Ying Liu, Xiaoyan Zhu

Attention-based LSTM for Aspect-level Sentiment Classification

To reveal that the sentiment polarity of a sentence is not only determined by the content but is also highly related to the concerned aspect, we propose an Attention-based Long Short-Term Memory Network for aspect-level sentiment classification.

Yequan Wang, Minlie Huang, Li Zhao, Xiaoyan Zhu

Attention-based LSTM for Aspect-level Sentiment Classification

Recent Publications

Yiqun Yao, Zheng Zhang, Jing Li, Yequan Wang (2028). Masked Structural Growth for 2x Faster Language Model Pre-training.

PDF Cite Code Project Project

Xiang Li, Yiqun Yao, Xin Jiang, Xuezhi Fang, China Telecom, Yequan Wang, Zhongjiang He, Zhongyuan Wang, Xuelong Li, Tiejun Huang (2024). 52B to 1T: Lessons Learned via Tele-FLM Series.

PDF Cite Project Tele-FLM-1T Tele-FLM

Yiqun Yao, Yequan Wang (2023). Research without Re-search: Maximal Update Parametrization Yields Accurate Loss Prediction across Scales.

PDF Cite Project

Yequan Wang, Hengran Zhang, Aixin Sun, Xuying Meng (2022). CORT: A New Baseline for Comparative Opinion Classification by Dual Prompts. In Findings of the EMNLP 2022.

PDF Cite Code Dataset Project

Ting Lin, Aixin Sun, Yequan Wang (2022). EDU-Capsule: aspect-based sentiment analysis at clause level. KAIS.

PDF Cite Code Slides DOI

Yequan Wang, Xiang Li, Aixin Sun, Xuying Meng, Huaming Liao, Jiafeng Guo (2022). CofeNet: Context and Former-Label Enhanced Net for Complicated Quotation Extraction. In COLING 2022.

PDF Cite Code Dataset Project

Xuying Meng, Yequan Wang, Runxin Ma, Haitong Luo, Xiang Li, Yujun Zhang (2022). Packet Representation Learning for Traffic Classification. In KDD 2022.

PDF Cite Code DOI

Siqi Fan, Yequan Wang, Jing Li, Zheng Zhang, Shuo Shang, Peng Han (2022). Interactive Information Extraction by Semantic Information Graph. In IJCAI 2022.

PDF Cite Code Dataset

Yiyi Liu, Yequan Wang, Aixin Sun, Xuying Meng, Jing Li, Jiafeng Guo (2022). A Dual-Channel Framework for Sarcasm Recognition by Detecting Sentiment Conflict. In Findings of NAACL 2022.

PDF Cite Code Dataset

Xi Lin, Yequan Wang, Xiaokui Xiao, Zengxiang Li, Sourav S. Bhowmick (2019). Path Travel Time Estimation using Attribute-related Hybrid Trajectories Network. In CIKM 2019.

PDF Cite DOI