[{"Version":"5","Task":"1","Concept":"1","Class Description":"Case study on fine-tuning an LLM to choose promising news headlines with high click-through. Discussion of applications of fine-tuning LLMs.","Coding Intensity":"Mostly programming without resorting to \"No Code\" tools"},{"Version":"5","Task":"1","Concept":"2","Class Description":"A session where students choose an AI tool and build a workflow. Discussion and sharing of projects.","Coding Intensity":"Mix of \"No Code\" and programming"},{"Version":"5","Task":"2","Concept":"1","Class Description":"Case study on impacts of AI on artists (from Prof. Bell's research). Exploration of a named entity recognition pipeline (to detect artist names in Midjourney prompts) using a hybrid machine learning + LLM workflow.","Coding Intensity":"Mostly programming without resorting to \"No Code\" tools"},{"Version":"5","Task":"2","Concept":"2","Class Description":"A session where students choose an AI tool and build a workflow. Discussion and sharing of projects.","Coding Intensity":"Mix of \"No Code\" and programming"},{"Version":"5","Task":"3","Concept":"1","Class Description":"A session where students choose an AI tool and build a workflow. Discussion and sharing of projects.","Coding Intensity":"Mix of \"No Code\" and programming"},{"Version":"5","Task":"3","Concept":"2","Class Description":"Use an LLM to label video footage with faces and train a smaller model for face detection. This is known as distillation. Technically very challenging.","Coding Intensity":"Mostly programming without resorting to \"No Code\" tools"},{"Version":"5","Task":"4","Concept":"1","Class Description":"Retrieval Augmented Generation (RAG). Explore a case study and practice building a RAG system. Activity: build a small RAG system from a set of documents and see how well it works.","Coding Intensity":"Mix of \"No Code\" and programming"},{"Version":"5","Task":"4","Concept":"2","Class Description":"Case study on fine-tuning an LLM to choose promising news headlines with high click-through. Discussion of applications of fine-tuning LLMs.","Coding Intensity":"Mix of \"No Code\" and programming"},{"Version":"5","Task":"5","Concept":"1","Class Description":"Use an LLM to label video footage with faces and train a smaller model for face detection. This is known as distillation. Technically very challenging.","Coding Intensity":"Mix of \"No Code\" and programming"},{"Version":"5","Task":"5","Concept":"2","Class Description":"Retrieval Augmented Generation (RAG). Explore a case study and practice building a RAG system. Activity: build a small RAG system from a set of documents and see how well it works.","Coding Intensity":"Mostly \"No Code\" Tools"},{"Version":"5","Task":"6","Concept":"1","Class Description":"Retrieval Augmented Generation (RAG). Explore a case study and practice building a RAG system. Activity: build a small RAG system from a set of documents and see how well it works.","Coding Intensity":"Mostly programming without resorting to \"No Code\" tools"},{"Version":"5","Task":"6","Concept":"2","Class Description":"Case study on fine-tuning an LLM to choose promising news headlines with high click-through. Discussion of applications of fine-tuning LLMs.","Coding Intensity":"Mostly \"No Code\" Tools"},{"Version":"5","Task":"7","Concept":"1","Class Description":"Retrieval Augmented Generation (RAG). Explore a case study and practice building a RAG system. Activity: build a small RAG system from a set of documents and see how well it works.","Coding Intensity":"Mostly \"No Code\" Tools"},{"Version":"5","Task":"7","Concept":"2","Class Description":"A session where students choose an AI tool and build a workflow. Discussion and sharing of projects.","Coding Intensity":"Mix of \"No Code\" and programming"},{"Version":"5","Task":"8","Concept":"1","Class Description":"Retrieval Augmented Generation (RAG). Explore a case study and practice building a RAG system. Activity: build a small RAG system from a set of documents and see how well it works.","Coding Intensity":"Mostly programming without resorting to \"No Code\" tools"},{"Version":"5","Task":"8","Concept":"2","Class Description":"Case study on fine-tuning an LLM to choose promising news headlines with high click-through. Discussion of applications of fine-tuning LLMs.","Coding Intensity":"Mostly \"No Code\" Tools"},{"Version":"5","Task":"9","Concept":"1","Class Description":"Case study on impacts of AI on artists (from Prof. Bell's research). Exploration of a named entity recognition pipeline (to detect artist names in Midjourney prompts) using a hybrid machine learning + LLM workflow.","Coding Intensity":"Mostly programming without resorting to \"No Code\" tools"},{"Version":"5","Task":"9","Concept":"2","Class Description":"Build a discrete choice experiment (conjoint) with a product\/service chosen by the students in class. Activity: Explore product feature combinations that increase estimated market share.","Coding Intensity":"Mostly programming without resorting to \"No Code\" tools"},{"Version":"5","Task":"10","Concept":"1","Class Description":"Case study on clustering customers of an Italian online supermarket called Mammapack. Activity: finding clusters and implications for Mammapack.","Coding Intensity":"Mostly \"No Code\" Tools"},{"Version":"5","Task":"10","Concept":"2","Class Description":"Retrieval Augmented Generation (RAG). Explore a case study and practice building a RAG system. Activity: build a small RAG system from a set of documents and see how well it works.","Coding Intensity":"Mostly \"No Code\" Tools"}]