Which course? Discourse! Teaching Discourse and Generation in the Era of LLMs

2026/02/02 by Junyi Jessy Li, Yang Janet Liu, Kanishka Misra +2 · 1 voice
Computer Science · #cs.CL

paper · pdf

arxiv published 2026/02/02 · arxiv updated 2026/02/09

Abstract

The field of NLP has undergone vast, continuous transformations over the past few years, sparking debates going beyond discipline boundaries. This begs important questions in education: how do we design courses that bridge sub-disciplines in this shifting landscape? This paper explores this question from the angle of discourse processing, an area with rich linguistic insights and computational models for the intentional, attentional, and coherence structure of language. Discourse is highly relevant for open-ended or long-form text generation, yet this connection is under-explored in existing undergraduate curricula. We present a new course, "Computational Discourse and Natural Language Generation". The course is collaboratively designed by a team with complementary expertise and was offered for the first time in Fall 2025 as an upper-level undergraduate course, cross-listed between Linguistics and Computer Science. Our philosophy is to deeply integrate the theoretical and empirical aspects, and create an exploratory mindset inside the classroom and in the assignments. This paper describes the course in detail and concludes with takeaways from an independent survey as well as our vision for future directions.

Citations

• Large Language Model Hacking: Quantifying the Hidden Risks of Using LLMs for Text Annotation
• PrefPalette: Personalized Preference Modeling with Latent Attributes
• Is It JUST Semantics? A Case Study of Discourse Particle Understanding in LLMs
• Frankentext: Stitching random text fragments into long-form narratives
• Pixels Versus Priors: Controlling Knowledge Priors in Vision-Language Models through Visual Counterfacts
• LLMs Get Lost In Multi-Turn Conversation
• Explanatory Summarization with Discourse-Driven Planning
• Finding Flawed Fictions: Evaluating Complex Reasoning in Language Models via Plot Hole Detection
• GUM-SAGE: A Novel Dataset and Approach for Graded Entity Salience Prediction
• QUDsim: Quantifying Discourse Similarities in LLM-Generated Text
• Behavioral Analysis of Information Salience in Large Language Models
• Are Language Models Models?
• How linguistics learned to stop worrying and love the language models
• The Alternative Annotator Test for LLM-as-a-Judge: How to Statistically Justify Replacing Human Annotators with LLMs
• Explicating the Implicit: Argument Detection Beyond Sentence Boundaries
• LLMs instead of Human Judges? A Large Scale Empirical Study across 20 NLP Evaluation Tasks
• Probing out-of-distribution generalization in machine learning for materials
• MiniCheck: Efficient Fact-Checking of LLMs on Grounding Documents
• Which questions should I answer? Salience Prediction of Inquisitive Questions
• Large language models that replace human participants can harmfully misportray and flatten identity groups
• InfoLossQA: Characterizing and Recovering Information Loss in Text Simplification
• QUDEVAL: The Evaluation of Questions Under Discussion Discourse Parsing
• BooookScore: A systematic exploration of book-length summarization in the era of LLMs
• Art or Artifice? Large Language Models and the False Promise of Creativity
• Llama 2: Open Foundation and Fine-Tuned Chat Models
• Lost in the Middle: How Language Models Use Long Contexts
• Factually Consistent Summarization via Reinforcement Learning with Textual Entailment Feedback
• Elaborative Simplification as Implicit Questions Under Discussion
• ChatGPT outperforms crowd workers for text-annotation tasks
• Large Language Models Are State-of-the-Art Evaluators of Translation Quality
• ClarifyDelphi: Reinforced Clarification Questions with Defeasibility Rewards for Social and Moral Situations
• Emergent World Representations: Exploring a Sequence Model Trained on a Synthetic Task
• minicons: Enabling Flexible Behavioral and Representational Analyses of Transformer Language Models
• How Do We Answer Complex Questions: Discourse Structure of Long-form Answers
• Training language models to follow instructions with human feedback
• DiscoScore: Evaluating Text Generation with BERT and Discourse Coherence
• Discourse Comprehension: A Question Answering Framework to Represent Sentence Connections
• Pragmatic competence of pre-trained language models through the lens of discourse connectives
• Recursively Summarizing Books with Human Feedback
• Implicit Representations of Meaning in Neural Language Models
• Is Incoherence Surprising? Targeted Evaluation of Coherence Prediction from Language Models
• RoFormer: Enhanced Transformer with Rotary Position Embedding
• QADiscourse -- Discourse Relations as QA Pairs: Representation, Crowdsourcing and Baselines
• Learning to summarize from human feedback
• BLEURT: Learning Robust Metrics for Text Generation
• Scaling Laws for Neural Language Models
• BLiMP: The Benchmark of Linguistic Minimal Pairs for English
• Discourse-Aware Neural Extractive Text Summarization
• The Curious Case of Neural Text Degeneration
• BERTScore: Evaluating Text Generation with BERT
• Neural Language Models as Psycholinguistic Subjects: Representations of\n Syntactic State
• Assessing the Ability of LSTMs to Learn Syntax-Sensitive Dependencies
• Modeling Local Coherence: An Entity-Based Approach
• Anaphora and Discourse Structure

Discussions

• Then Jessy and I will be at TeachNLP workshop on Sunday, where Jessy will present our new course on discourse ;) (arxiv.org/abs/2602.02878) Excited to catch up with old friends and new! [bsky, 3 points, 0 comments]

Related

• Natural Language Generation
• From Words to Wisdom: Discourse Annotation and Baseline Models for Student Dialogue Understanding
• Training an NLP Scholar at a Small Liberal Arts College: A Backwards Designed Course Proposal
• “Who's Doing the Thinking Here?”: A Pedagogy-First Approach to Integrating Large Language Models in Higher Education
• Natural Language Generation