Expectation-based theories of sentence processing posit that processing difficulty is determined by predictability in context. While predictability quantified via surprisal has gained empirical support, this representation-agnostic measure leaves open the question of how to best approximate the human comprehender's latent probability model. This article first describes an incremental left-corner parser that incorporates information about common linguistic abstractions such as syntactic categories, predicate-argument structure, and morphological rules as a computational-level model of sentence processing. The article then evaluates a variety of structural parsers and deep neural language models as cognitive models of sentence processing by comparing the predictive power of their surprisal estimates on self-paced reading, eye-tracking, and fMRI data collected during real-time language processing. The results show that surprisal estimates from the proposed left-corner processing model deliver comparable and often superior fits to self-paced reading and eye-tracking data when compared to those from neural language models trained on much more data. This may suggest that the strong linguistic generalizations made by the proposed processing model may help predict humanlike processing costs that manifest in latency-based measures, even when the amount of training data is limited. Additionally, experiments using Transformer-based language models sharing the same primary architecture and training data show a surprising negative correlation between parameter count and fit to self-paced reading and eye-tracking data. These findings suggest that large-scale neural language models are making weaker generalizations based on patterns of lexical items rather than stronger, more humanlike generalizations based on linguistic structure.
more »
« less
Predicting Long-Term Citations from Short-Term Linguistic Influence
A standard measure of the influence of a research paper is the number of times it is cited. However, papers may be cited for many reasons, and citation count offers limited information about the extent to which a paper affected the content of subsequent publications. We therefore propose a novel method to quantify linguistic influence in timestamped document collections. There are two main steps: first, identify lexical and semantic changes using contextual embeddings and word frequencies; second, aggregate information about these changes into per-document influence scores by estimating a high-dimensional Hawkes process with a low-rank parameter matrix. We show that this measure of linguistic influence is predictive of future citations: the estimate of linguistic influence from the two years after a paper’s publication is correlated with and predictive of its citation count in the following three years. This is demonstrated using an online evaluation with incremental temporal training/test splits, in comparison with a strong baseline that includes predictors for initial citation counts, topics, and lexical features.
more »
« less
- NSF-PAR ID:
- 10382963
- Date Published:
- Journal Name:
- Findings of the Association for Computational Linguistics: EMNLP 2022
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
We introduce a new general framework for sign recognition from monocular video using limited quantities of annotated data. The novelty of the hybrid framework we describe here is that we exploit state-of-the art learning methods while also incorporating features based on what we know about the linguistic composition of lexical signs. In particular, we analyze hand shape, orientation, location, and motion trajectories, and then use CRFs to combine this linguistically significant information for purposes of sign recognition. Our robust modeling and recognition of these sub-components of sign production allow an efficient parameterization of the sign recognition problem as compared with purely data-driven methods. This parameterization enables a scalable and extendable time-series learning approach that advances the state of the art in sign recognition, as shown by the results reported here for recognition of isolated, citation-form, lexical signs from American Sign Language (ASL).more » « less
-
We introduce a new general framework for sign recognition from monocular video using limited quantities of annotated data. The novelty of the hybrid framework we describe here is that we exploit state-of-the art learning methods while also incorporating features based on what we know about the linguistic composition of lexical signs. In particular, we analyze hand shape, orientation, location, and motion trajectories, and then use CRFs to combine this linguistically significant information for purposes of sign recognition. Our robust modeling and recognition of these sub-components of sign production allow an efficient parameterization of the sign recognition problem as compared with purely data-driven methods. This parameterization enables a scalable and extendable time-series learning approach that advances the state of the art in sign recognition, as shown by the results reported here for recognition of isolated, citation-form, lexical signs from American Sign Language (ASL).more » « less
-
more » « less
Abstract Diachronic word embeddings—vector representations of words over time—offer remarkable insights into the evolution of language and provide a tool for quantifying sociocultural change from text documents. Prior work has used such embeddings to identify shifts in the meaning of individual words. However, simply knowing that a word has changed in meaning is insufficient to identify the instances of word usage that convey the historical meaning or the newer meaning. In this study, we link diachronic word embeddings to documents, by situating those documents as leaders or laggards with respect to ongoing semantic changes. Specifically, we propose a novel method to quantify the degree of semantic progressiveness in each word usage, and then show how these usages can be aggregated to obtain scores for each document. We analyze two large collections of documents, representing legal opinions and scientific articles. Documents that are scored as semantically progressive receive a larger number of citations, indicating that they are especially influential. Our work thus provides a new technique for identifying lexical semantic leaders and demonstrates a new link between progressive use of language and influence in a citation network.
-
more » « less
Abstract Following widespread availability of computerized databases, much research has correlated bibliometric measures from papers or patents to subsequent success, typically measured as the number of publications or citations. Building on this large body of work, we ask the following questions: given available bibliometric information in one year, along with the combined theories on sources of creative breakthroughs from the literatures on creativity and innovation, how accurately can we explain the impact of authors in a given research community in the following year? In particular, who is most likely to publish, publish highly cited work, and even publish a highly cited outlier? And, how accurately can these existing theories predict breakthroughs using only contemporaneous data? After reviewing and synthesizing (often competing) theories from the literatures, we simultaneously model the collective hypotheses based on available data in the year before RNA interference was discovered. We operationalize author impact using publication count, forward citations, and the more stringent definition of being in the top decile of the citation distribution. Explanatory power of current theories altogether ranges from less than 9% for being top cited to 24% for productivity. Machine learning (ML) methods yield similar findings as the explanatory linear models, and tangible improvement only for non-linear Support Vector Machine models. We also perform predictions using only existing data until 1997, and find lower predictability than using explanatory models. We conclude with an agenda for future progress in the bibliometric study of creativity and look forward to ML research that can explain its models.
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
