Welcome

Computational Understanding of Natural Language (CAP-6640)

Spencer Lyon

Welcome to week 1!

This week we will get to know eachother and get on the same page with respect to what this class is all about.

We will also dive right in to setting up the tools (Python + NLP libraires) and learning foundational concepts.

Each week lecture notes will be distributed as a collection of Jupyter notebooks. The notebooks will follow a strict naming convention, where each notebook has a name such as L@@.##_XXX.md where @@ is a two digit lecture number and ## is a two digit file number and XXX is one or more words describing the content of the notebook. We will work through the notebooks in the order indicated by the ##. The XXX are to provide easier access when reviewing notes after class.

About Me¶

• Spencer Lyon (spencer.lyon@ucf.edu)

• Economics PhD from NYU (2018)

• Love to teach: mostly economics, data science, AI/ML – all have programming/computational element

• Moved to Orlando in July 2018 with wife and 5 (yes!) kids

• Run data and AI consulting practice at Arete Capital Partners

• Working on a couple startups (always...)

About you¶

• Background?

• Progress in program?

• Areas of interest? (meaningful answers here! they matter…)

• Rumors about the course?

About the course¶

• Interdisciplinary by nature

• “Living course”: never been taught, content is flexible

  • More ideas/topics than time!

  • Is centered on the exciting space of NLP, which with the advent of LLMs is moving very quickly

• Heavy emphasis on programming and hands-on work

• Goal: get you to a point where you can build NLP applications and understand the underlying concepts

• Tools: Python, Jupyter Notebooks, various NLP libraries (Spacy, huggingface, langchain, etc.)

• Assessments: programming assignments, projects, participation, oral exams (more on this later)

• Collaboration encouraged, but all submitted work must be your own

• AI policy: use of AI tools is allowed, but must be disclosed and properly cited

• Office hours and communication: reach out via email or during office hours for help or questions

• Looking forward to a great semester together!

• Core concepts

  • Text processing and representation: from raw text to meaningful features

  • Classical NLP and machine learning techniques for language tasks

  • Deep learning evolution: RNNs to transformers to foundation models

  • Building applications with modern LLM APIs and RAG patterns

  • Agentic AI systems: orchestrating LLMs with tools and workflows

• Applications

  • Machine translation and multilingual systems

  • Chatbots and conversational AI

  • Search and information retrieval

  • Sentiment analysis and text classification

  • Summarization and question answering

  • Industry domains: tech, healthcare, finance, legal

Expectations¶

• Study reading assignments before class

• Complete assignments on time -- no exceptions

• Participate in in-class discussions

• Spend ~3-6 hours outside of class per week

• Communication

  • Post all content related questions to class discussion forum

  • Respond to peers’ questions and engage in discussions

  • Personal questions should go directly to me via email

  • I do not use email on Sunday. Other days I will respond within 48 hours.

• Deliverables

  • Homework (~8 – 30%)

  • Exam (2 – 30%)

  • Projects (2 – 30%)

  • Citizenship (throughout - 10%)

    • First best: attend class in person, actively participate

    • Acceptable: attend virtually, but keep video on and be ready to speak up when called on

    • Unacceptable: attend virtually, but keep video off and/or don’t participate in discussions

Tools/Resources¶

• Core text: lecture notes and assignments

• Lecture notes are accessible via the course website

• Lecture notes AND assignments in Jupyter notebooks

• All course administration will happen through webcourses (Canvas)

  • Assignments me <-> you

  • Feedback on assignments me -> you

  • Discussion me + you <-> me + you

• Official grades will be visible on canvas