Papers by Melissa Roemmele
LLMs Behind the Scenes: Enabling Narrative Scene Illustration (2025.emnlp-main)
Copied to clipboard
| Challenge: | Generative AI has established the ability to readily transform content from one medium to another. |
| Approach: | They propose a pipeline that uses large language models to prompt text-to-image models to generate scenes for story text. |
| Outcome: | The proposed pipeline synthesizes illustrations for scenes in a story corpus using human annotation tasks. |
AnswerQuest: A System for Generating Question-Answer Items from Multi-Paragraph Documents (2021.eacl-demos)
Copied to clipboard
| Challenge: | Existing systems that generate and answer questions in a question-and-answer format can facilitate reading comprehension. |
| Approach: | They propose a system that integrates question answering and question generation tasks to produce a list of Q&A items for a text. |
| Outcome: | The proposed system generates a catalog of Q&A items for a text. |
AbLit: A Resource for Analyzing and Generating Abridged Versions of English Literature (2023.eacl-main)
Copied to clipboard
| Challenge: | Creating an abridged version of a text requires shortening it while maintaining its linguistic qualities. |
| Approach: | They propose an abridgement task that requires shortening and maintaining linguistic qualities of a text while maintaining its linguistic quality. |
| Outcome: | The proposed dataset captures passage-level alignments between original and abridged texts . it can be used to generate a bridge and shorten the original text . |
From Test-Taking to Test-Making: Examining LLM Authoring of Commonsense Assessment Items (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) excel in answering questions pertaining to commonsense reasoning and inference. |
| Approach: | They prompt LLMs to generate items in the style of a benchmark for commonsense reasoning . they find that LLM authors that answer COPA items are more successful . |
| Outcome: | The authors' responses to their own items and their own generated items are better than those of the original LLMs. |