ÌÇÐÄlogo

We invite you to submit your latest research to the Information Theory for Large Language Models (IT4LLM) workshop at ISIT2026, held July 3 in Guangzhou, China. This workshop explores how principled information-theoretic frameworks can advance the efficiency, reliability, and interpretability of large language models. As a researcher pushing the boundaries of intelligent systems, your work at the intersection of information theory and AI is exactly what we need to advance the next generation of efficient, reliable, and transparent language models.

Why Submit to IT4LLM?

• Elite Speaker Lineup: Present your work alongside keynotes from Prof. Yuejie Chi (Yale University) and Prof. Yingbin Liang (The Ohio State University).

• Cross-Disciplinary Panel: Engage with our distinguished panel including Prof. Deniz Gündüz (Imperial College London), Prof. Jiantao Jiao (UC Berkeley), and Prof. Ioannis Kontoyiannis (Cambridge University).

• Interdisciplinary Collaboration: Connect with researchers from information theory, machine learning, and AI communities to tackle fundamental challenges in LLM efficiency and interpretability.

Our core topics include, but are not limited to:

• Information-theoretic foundations of LLMs  
• Memory and retrieval in LLMs
• Reinforcement learning and LLM reasoning
• Generalization bounds for LLM training
• Uncertainty/hallucination detection in LLMs
• Multimodal LLM information fusion
• Energy-efficient LLMs via info-theoretic optimization

Submission Details:

• Deadline: April 7, 2026
• Notification: April 21, 2026
• Final Manuscripts: April 28, 2026
• Submission Link: 
• Workshop Date: July 3, 2026 (Guangzhou, China)
• Website: 

Join us to bridge information theory and AI, and help build the theoretical foundations for the next generation of transparent and efficient language models.