Automatic Identification of Narrative Diegesis and Point of View

Eisenberg, Joshua D.; Finlayson, Mark A.

doi:10.18653/v1/w16-5705

Cited by 9 publications

(5 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…beings) because we are interested in extracting the information most closely coupled with people or characters. We were less concerned In the prior section we showed the PEDC contains a significant amount of personal events by running the POV and diegesis extractors from Eisenberg and Finlayson (2016). We found that the PEDC contains 56% first-person narrators, and 32% homodiegetic narrators.…”

Section: What Are Personal Events?mentioning

confidence: 85%

“…We back our claim (that dialogue is different than newswire) by comparing the distributions of narrative point of view (POV) and diegesis of the narrators in PEDC with the Reuters-21578 newswire corpus. 2 Eisenberg and Finlayson (2016) found that narrators in newswire texts from the Reuters-21,578 corpus use the first-person POV less than 1% of the time, and are homodiegetic less than 1% of the time. However, in the 14 episodes (1,028 utterances) of This American Life, we found that 56% narrators were first-person, and 32% narrators were homodiegetic.…”

Section: Motivationmentioning

confidence: 99%

“…We found these distributions in PEDC by using the automatic POV and diegesis extractors from Eisenberg and Finlayson (2016), which were open sourced. 3 Comparing the distributions of POV and diegesis for the PEDC to that of newswire demonstrates how different spoken dialogue is.…”

Section: Motivationmentioning

confidence: 99%

See 2 more Smart Citations

Automatic extraction of personal events from dialogue

Eisenberg

Sheriff²

2020

Proceedings of the First Joint Workshop on Narrative Understanding, Storylines, and Events

Self Cite

View full text Add to dashboard Cite

In this paper we introduce the problem of extracting events from dialogue. Previous work on event extraction focused on newswire, however we are interested in extracting events from spoken dialogue. To ground this study, we annotated dialogue transcripts from fourteen episodes of the podcast This American Life. This corpus contains 1,038 utterances, made up of 16,962 tokens, of which 3,664 represent events. The agreement for this corpus has a Cohen's κ of 0.83. We have open sourced this corpus for the NLP community. With this corpus in hand, we trained support vector machines (SVM) to correctly classify these phenomena with 0.68 F1, when using episodefold cross-validation. This is nearly 100% higher F1 than the baseline classifier. The SVM models achieved performance of over 0.75 F1 on some testing folds. We report the results for SVM classifiers trained with four different types of features (verb classes, part of speech tags, named entities, and semantic role labels), and different machine learning protocols (under-sampling and trigram context). This work is grounded in narratology and computational models of narrative. It is useful for extracting events, plot, and story content from spoken dialogue.

show abstract

Section: What Are Personal Events?mentioning

confidence: 85%

Section: Motivationmentioning

confidence: 99%

See 1 more Smart Citation

Automatic extraction of personal events from dialogue

Eisenberg

Sheriff²

2020

Proceedings of the First Joint Workshop on Narrative Understanding, Storylines, and Events

Self Cite

View full text Add to dashboard Cite

show abstract

“…An annotation scheme, NarrativeML, was developed to facilitate the annotation of these concepts. More recently, Eisenberg and Finlayson (2016) trained SVM models to distinguish between first-person and third-person point of view in a narrative, as well as for determining whether the narrator is involved in the story, i.e. diegesis.…”

Section: Related Workmentioning

confidence: 99%

Changing the Narrative Perspective: From Deictic to Anaphoric Point of View

Chen

Bunescu

2021

Preprint

View full text Add to dashboard Cite

We introduce the task of changing the narrative point of view, where characters are assigned a narrative perspective that is different from the one originally used by the writer.The resulting shift in the narrative point of view alters the reading experience and can be used as a tool in fiction writing or to generate types of text ranging from educational to self-help and self-diagnosis. We introduce a benchmark dataset containing a wide range of types of narratives annotated with changes in point of view from deictic (first or second person) to anaphoric (third person) and describe a pipeline for processing raw text that relies on a neural architecture for mention selection. Evaluations on the new benchmark dataset show that the proposed architecture substantially outperforms the baselines by generating mentions that are less ambiguous and more natural.

show abstract

“…This chapter was adapted from an article I published in the Proceedings of the Second Workshop on Computing News Storylines(Eisenberg, November 2016).…”

mentioning

confidence: 99%

Automatic Extraction of Narrative Structure from Long Form Text

Eisenberg¹

View full text Add to dashboard Cite

Automatic understanding of stories is a long-held goal of artificial intelligence and natural language processing research communities. Stories explain the human experience. Understanding stories promotes the understanding of both individuals and groups of people; cultures, societies, families, organizations, governments, and corporations, to name a few. People use stories to share information. Stories are told-by narrators-in linguistic bundles of words called narratives. My work has given computers understanding of some aspects of narrative structure. Specifically, where are the boundaries of a narrative in a text. This is the task of determining where a narrative begins and ends, a non-trivial task, because people rarely tell one story at a time. People don't specifically announce when we are starting or stopping our stories: We interrupt each other. We tell stories within stories. Before my work, computers had no awareness of narrative boundaries, essentially where stories begin and end. My programs can extract narrative boundaries from novels and short stories with an F 1 of 0.65. viii Before this I worked on teaching computers to identify which paragraphs of text have story content, with an F 1 of 0.75 (which is state of the art). Additionally, I have taught computers to identify the narrative point of view (POV; how the narrator refers to theirself) and diegesis (how is the narrator involved in the story's action) with F 1 of over 0.90 for both narrative characteristics. For the narrative POV, diegesis, and narrative level extractors I ran annotation studies, with high agreement, that allowed me to teach computational models to identify structural elements of narrative through supervised machine learning. My work has given computers the ability to find where stories begin and end in raw text. This will allow for further, automatic analysis, like extraction of plot, intent, event causality, and event coreference. These are difficult when there are multiple narratives in one text. There are two key contributions in my work: 1) my identification of features that accurately extract information about narrators, and narrative levels and 2) the gold-standard data generated from running annotation studies on identifying these same elements of narrative structure.

show abstract

Automatic Identification of Narrative Diegesis and Point of View

Cited by 9 publications

References 9 publications

Automatic extraction of personal events from dialogue

Automatic extraction of personal events from dialogue

Changing the Narrative Perspective: From Deictic to Anaphoric Point of View

Automatic Extraction of Narrative Structure from Long Form Text

Contact Info

Product

Resources

About