Papers by Steve Skiena

7 papers
Chapter Ordering in Novels (2022.emnlp-main)

Copied to clipboard

Challenge: a major challenge in research on long-form narrative texts is the cost of annotation . authors propose a new task that reconstructs the original order of chapters in novels without the need for human annotation.
Approach: They propose a task that reconstructs the original order of chapters in novels given a random permutation of the text.
Outcome: The proposed task yields a Spearman correlation of 0.59 on the novel and challenging task, substantially above baseline.
Chapter Captor: Text Segmentation in Novels (2020.emnlp-main)

Copied to clipboard

Challenge: Using a hybrid approach, we identify chapter boundaries in novels . chapter boundaries are typically denoted by formatting conventions such as page breaks, white-space, chapter numbers, and titles.
Approach: They build a project Gutenberg data set of 9,126 English novels to analyze chapter boundaries . they use neural inference and rule matching to recognize chapter title headers .
Outcome: The proposed method achieves an F1 score of 0.77 on the segmentation task . the annotated data reveal interesting historical trends in the chapter structure of novels .
Cleaning Dirty Books: Post-OCR Processing for Previously Scanned Texts (2021.findings-emnlp)

Copied to clipboard

Challenge: a large amount of work is required to clean digitized books for NLP analysis because of errors in the scanned text and duplicate volumes in the corpora.
Approach: They propose methods to handle optical character recognition errors in scanned texts . they identify the canonical version for each of 17,136 repeatedly-scanned books .
Outcome: The proposed method corrects over six times as many errors as it introduces, the authors show . the authors evaluate a collection of 19,347 texts from the Gutenberg dataset and 96,635 from the HathiTrust Library .
What time is it? Temporal Analysis of Novels (2020.emnlp-main)

Copied to clipboard

Challenge: a novel based on the flow of time provides a framework for understanding the text . a computational approach to annotate a book's lines with wall clock times is needed to understand the flow through time.
Approach: They propose to annotate each line of a book with wall clock times . they use a data set of hourly time phrases from 52,183 fictional books .
Outcome: The proposed method improves upon baselines by over two hours and can partition a book into segments that correspond to a particular time-of-day.
Multi-view Models for Political Ideology Detection of News Articles (D18-1)

Copied to clipboard

Challenge: Existing models for automatic detection of political ideology only leverage textual cues to identify the ideology evinced by a news article.
Approach: They propose a novel attention based multi-view model that leverages cues from textual content and the network structure of news articles to identify political ideology.
Outcome: The proposed model outperforms state of the art models by 10 percentage points on a battery of baselines and compares with baselines.
The Trumpiest Trump? Identifying a Subject’s Most Characteristic Tweets (D19-1)

Copied to clipboard

Challenge: characterization scores are associated with popularity of a given short text, but are not always representative of the source.
Approach: They use a dataset of tweets from 15 celebrities to quantify the extent to which a given short text is characteristic of a specific person.
Outcome: The proposed model shows a statistically significant correlation between characterization scores and popularity of the associated texts for 13 of the 15 celebrities in the study.
Simple Neologism Based Domain Independent Models to Predict Year of Authorship (C18-1)

Copied to clipboard

Challenge: Using domain independent models, we date documents based only on neologism usage patterns . nasa models use only 200 input features, compared to state of the art models using 200K features.
Approach: They propose domain independent models to date documents based only on neologism usage patterns.
Outcome: The proposed models can generalize to various domains like News, Fiction, and Non-Fiction with competitive performance.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations