Welcome to the inclusive NLP (iNLP) Lab at SUTD!
This page contains the schedule of our weekly reading group - where we share and discuss the latest research in natural language processing and large language models.
👉 If you'd like to join us or give a talk (to make this reading group more inclusive :), please feel free to reach out.
| Date | Presenter | Content | Slides |
|---|---|---|---|
| 28/01/2026 | Zhibo Man | Exploring Multi-Agent Frameworks for Research with LLMs: Paper2Review & Paper2Rebuttal | Slides |
| 04/02/2026 | Pengyang Shao | LLM Unlearning: From Balanced Disentanglement to Local Entropy Maximization | Slides |
| 11/02/2026 | Yuyang Dai | Exploring Silicon-Based Societies: An Early Study of the Moltbook Agent Community | Slides |
| 25/02/2026 | Long P. Hoang | Steering Transferability in Activation Space | Slides |
| 04/03/2026 | Ryner Tan | Distillation: Techniques and Defenses | Slides |
| 18/03/2026 | Shaoyang Xu | From "Artificial Hivemind" to Diversity in LLM Research | Slides |
| 25/03/2026 | Chen Huang | Reasoning Models Struggle to Control their Chains of Thought | Slides |
| 01/04/2026 | Yi Feng | Benchmarks & Agentic RL in the Claw Era | Slides |
| 08/04/2026 | Le Viet Hai | Discussion on Recent Claude Leak | Slides |
| 06/05/2026 | Suhyun Lee | From “What Is Your AI Agent Buying?” to “Sparks of Rationality” | Slides |
| 13/05/2026 | Matthew Christopher Pohadi | Paradigm Routing, ToTool Use, and Multimodal Search | Slides |
| 03/06/2026 | Long P. Hoang | Diffusion Language Models | Slides |
| 10/06/2026 | Chen Huang | When AI Builds Itself | Slides |
| 17/06/2026 | Shaoyang Xu | Anthropic's Fable 5 | Slides |
For previous meeting schedules and slides, please see:
- When: Weekly (usually Wednesday 10.30am - 12pm).
- Where: Hybrid (SUTD campus + online Ms Teams link).
The inclusive NLP (iNLP) Lab at SUTD is dedicated to advancing the next generation of Natural Language Processing (NLP) systems that are not only powerful, but also inclusive and trustworthy.
- Inclusive → We design models that support diverse languages, cultures, and users. This includes research on multilingual LLMs, accessibility through efficient and compressed models, and human-centered design for broader adoption.
- Trustworthy → We ensure that NLP systems are safe, robust, and reliable across settings. Our work includes LLM safety frameworks, fair and comprehensive evaluation, and mechanistic studies to better understand knowledge and reasoning inside LLMs.
- Fundamentals → Beyond application-driven projects, we pursue fundamental questions in machine learning and language modeling, exploring how models represent, reason, and interact with humans.