# Whodunnit? Crime Drama as a Case for Natural Language Understanding

arXiv: https://arxiv.org/abs/1710.11601 · HTML: https://mlsanity.com/paper/1710.11601/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 31 Oct 2017.

## Built on by (most cited first, 6 of 6)

- [SummScreen: A Dataset for Abstractive Screenplay Summarization](https://mlsanity.com/paper/2104.07091/) (2021)
- [Large Language Models Fall Short: Understanding Complex Relationships in Detective Narratives](https://mlsanity.com/paper/2402.11051/) (2024)
- [Narration Generation for Cartoon Videos](https://mlsanity.com/paper/2101.06803/) (2021)
- [Analysing the Memorability of a Procedural Crime-Drama TV Series, CSI](https://mlsanity.com/paper/2208.03479/) (2022)
- [Novel Aficionados and Doppelg"angers: a referential task for semantic representations of individual entities](https://mlsanity.com/paper/2104.10270/) (2021)
- [The Challenge and Reward of Fair Play in Narrative: A Computational Approach](https://mlsanity.com/paper/2507.13841/) (2025)

## Abstract (arXiv)

In this paper we argue that crime drama exemplified in television programs such as CSI:Crime Scene Investigation is an ideal testbed for approximating real-world natural language understanding and the complex inferences associated with it. We propose to treat crime drama as a new inference task, capitalizing on the fact that each episode poses the same basic question (i.e., who committed the crime) and naturally provides the answer when the perpetrator is revealed. We develop a new dataset based on CSI episodes, formalize perpetrator identification as a sequence labeling problem, and develop an LSTM-based model which learns from multi-modal data. Experimental results show that an incremental inference strategy is key to making accurate guesses as well as learning from representations fusing textual, visual, and acoustic input.