Tamoor of University of Copenhagen

Large Language Models (LLMs) have shown remarkable capabilities in language generation, yet their reasoning often falls short on complex, multi-step problems. This talk explores Chain of Thought (CoT) prompting—a powerful technique that guides models to generate intermediate reasoning steps before arriving at an answer. We will cover key developments, practical examples in math and commonsense tasks, and recent innovations like self-consistency and Tree-of-Thought prompting. The session will also highlight current limitations and future directions, including the fusion of CoT with retrieval and symbolic tools. Whether you're working with LLMs in research or applications, this talk will offer insights into unlocking their deeper reasoning potential.

Biography:

Ameer Tamoor Khan is a postdoctoral researcher at the Department of Plant and Environmental Sciences, University of Copenhagen. He holds a PhD from The Hong Kong Polytechnic University and has authored 30+ peer-reviewed publications across AI, robotics, and computer vision. His research focuses on integrating machine learning with robotics, large language models, and bio-inspired optimization for real-world applications in agriculture and autonomous systems. He has led and contributed to projects like HalmTDH, SAM2-based segmentation, and LLM-guided robotic control, and is actively engaged in cross-disciplinary innovation and international collaborations.

Email:

Address:Department of Plant and Environmental Sciences, University of Copenhagen, Copenhagen, Denmark