This presentation explores AI-aided design, starting with the fundamental concept of generative models like pix2pix, which learn to create new images based on paired input and output data. It then discusses more complex applications such as creating 3D objects from sketches and generating maps using geographical data. The speaker also introduces Large Language Models (LLMs), explaining their architecture, the evolution from RNNs to Transformers, and the concept of fine-tuning and embedding to customize their behavior and knowledge. The presentation concludes by demonstrating practical uses of LLMs, including local execution of models and exploring various types of machine learning problems such as classification and regression, showcasing how AI models can be applied to different datasets like medical or financial information.https://namjulee.github.io/njs-lab-public/work?id=2025-introductionToDesignComputation

Ver todos los episodios