Zhun Sun

Specially appointed associate professor (a.k.a. non-tenured casual laborer) at Language AI research center, Tohoku University. Funded by the Accelerating large-scale computational neuroscience by LLM project in the Brain/MINDS 2.0 program supported by Japan Agency for Medical Research and Development (AMED). Currently interested in the following topics:

1. Philosophy of Large Language Models (Philosophy of Science, Epistemology);
2. Post-Training of Large Language Models (Reinforcement Learning, Chain-of-Thought Reasoning)
3. Multimodal Large Language Models (Modality Fusion, Visual Reasoning)
4. Non-Linguistic Large Language Models (Code & Symbolic Reasoning, Latent Space Reasoning)
5. Neuroscience and Large Language Models (Biomimetic Interpretability)

Received Ph.D from Graduate School of Information Sciences, Tohoku University. Worked as a senior researcher at Tencent and Baidu in last 4 years before joining the Tohoku NLP Group, having great passion in ridiculing these two companies. Had experience in the following topics:

1. Robust representation learning, adversarial attack and defense.
2. Generative model, e.g., VAE, GAN and diffusion.
3. Tensor decomposition, tensor network.
4. Contrastive learning, both unimodal and multimodal.
5. Heterozygote of arbitrary topics above.

This site is my past blog storage.