Research

Research Area

Natural Language Processing

• Linguistic Intelligence: Linguistic Analysis, Language Representation, and Tokenization

• Cross-lingual NLP: Machine Translation, Multilingual Modeling, and Cross-lingual Transfer Learning

• Language Generation: Controllable Generation, Structured Text Generation, and Natural Language Generation

• Neuro-Symbolic NLP: Knowledge Integration, Compositional Reasoning, and Symbolic Language Understanding

Foundation Models

• Foundation Model Development: Data Collection, Preprocessing, Tokenization, Pretraining, Post-training, and Model Architecture

• Evaluation & Ecosystem Model Evaluation, Benchmark Development, Safety & Fairness Assessment, and Open Leaderboards

• Advanced Capabilities: Long-Context Modeling, Knowledge Editing, Code & Mathematical Reasoning, Multilingual & Multimodal Intelligence, and Agentic AI

Trustworthy Language Intelligence

• Reliability: LLM Evaluation, Benchmark Reliability, LLM-as-a-Judge, and Robustness

• Interpretability: Mechanistic Interpretability, Representation Analysis, and Model Behavior Understanding

• Controllability: Parameter Region Identification, Capability Transplantation, and Model Editing

• Safety: AI Safety, Red Teaming, Risk Assessment, and Safe Agentic Systems

• Privacy: Privacy-Preserving Language Models, Secure Inference, and Data Protection