Thailand

In sentiment analysis of longer texts, there may be a variety of topics discussed, of entities mentioned, and of sentiments expressed regarding each entity. We find a lack of studies exploring how such texts express their sentiment towards each entity of interest, and how these sentiments can be modelled. In order to better understand how sentiment regarding persons and organizations (each entity in our scope) is expressed in longer texts, we have collected a dataset of expert annotations where the overall sentiment regarding each entity is identified, together with the sentence-level sentiment for these entities separately. We show that the reader&#39;s perceived sentiment regarding an entity often differs from an arithmetic aggregation of sentiments at the sentence level. Only 70\% of the positive and 55\% of the negative entities receive a correct overall sentiment label when we aggregate the (human-annotated) sentiment labels for the sentences where the entity is mentioned. Our dataset reveals the complexity of entity-specific sentiment in longer texts, and allows for more precise modelling and evaluation of such sentiment expressions.

ACL 2024

Entity-Level Sentiment: More than the Sum of Its Parts

entity-level sentiment analysis

sentiment analysis

In sentiment analysis of longer texts, there may be a variety of topics discussed, of entities mentioned, and of sentiments expressed regarding each entity. We find a lack of studies exploring how such texts express their sentiment towards each entity of interest, and how these sentiments can be modelled. In order to better understand how sentiment regarding persons and organizations (each entity in our scope) is expressed in longer texts, we have collected a dataset of expert annotations where the overall sentiment regarding each entity is identified, together with the sentence-level sentiment for these entities separately. We show that the reader's perceived sentiment regarding an entity often differs from an arithmetic aggregation of sentiments at the sentence level. Only 70\% of the positive and 55\% of the negative entities receive a correct overall sentiment label when we aggregate the (human-annotated) sentiment labels for the sentences where the entity is mentioned. Our dataset reveals the complexity of entity-specific sentiment in longer texts, and allows for more precise modelling and evaluation of such sentiment expressions.

workshop paper

### Welcome!
The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. Our Virtual Poster Sessions will take place online Thursday, August 22, 2024.

You are required to register for this event. **Please register [here](https://2024.aclweb.org/registration). **

If you have already registered, please check your inbox for an email from Underline granting you access to ACL 2024 content.

Please register!

The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. More information will be announced soon.

Trust in media has reached a historical low as consumers increasingly doubt the credibility of the news they encounter. This growing skepticism is exacerbated by the prevalence of opinion-driven articles, which can influence readers’ beliefs to align with the authors’ viewpoints. In response to this trend, this study examines the expression of opinions in news by detecting subjective and objective language. We conduct an analysis of the subjectivity present in various news datasets and evaluate how different language models detect subjectivity and generalize to out-of-distribution data. We also investigate the use of in-context learning (ICL) within large language models (LLMs) and propose a straightforward prompting method that outperforms standard ICL and chain-of-thought (CoT) prompts.

Subjectivity Detection in English News using Large Language Models

Pre-trained language models consider the context of neighboring words and documents but lack any author context of the human generating the text. However, language depends on the author’s states, traits, social, situational, and environmental attributes, collectively referred to as human context (Soni et al., 2024). Human-centered natural language processing requires incorporating human context into language models. Currently, two methods exist: pre-training with 1) group-wise attributes (e.g., over-45-year-olds) or 2) individual traits. Group attributes are simple but coarse — not all 45-year-olds write the same way — while individual traits allow for more personalized representations, but require more complex modeling and data. It is unclear which approach benefits what tasks. We compare pre-training models with human context via 1) group attributes, 2) individual users, and 3) a combined approach on five user- and document-level tasks. Our results show that there is no best approach, but that human-centered language modeling holds avenues for different methods.

Comparing Pre-trained Human Language Models: Is it Better with Human Context as Groups, Individual Traits, or Both?

The growing number of online articles and reviews necessitates innovative techniques for document-level aspect-based sentiment analysis. Capturing the context in which an aspect is mentioned is crucial. Existing models have focused on relatively short reviews and may fail to consider distant contextual information. This is especially so in longer documents where an aspect may be referred to in multiple ways across dispersed sentences. This work introduces a hierarchical Transformer-based architecture that encodes information at different level of granularities with attention aggregation mechanisms to learn the local and global aspect-specific document representations. For empirical validation, we curate two datasets of long documents: one on social issues, and another covering various topics involving trust-related issues. Experimental results show that the proposed architecture outperforms state-of-the-art methods for document-level aspect-based sentiment classification. We also demonstrate the potential applicability of our approach for long document trust prediction.

Modeling Complex Interactions in Long Documents for Aspect-Based Sentiment Analysis

Social media are a critical component of the information ecosystem during public health crises. Understanding the public discourse is essential for effective communication and misinformation mitigation. Computational methods can aid these efforts through online social listening. We combined hierarchical text clustering and sentiment analysis to examine the face mask-wearing discourse in Germany during the COVID-19 pandemic using a dataset of 353,420 German X (formerly Twitter) posts from 2020. For sentiment analysis, we annotated a subsample of the data to train a neural network for classifying the sentiments of posts (neutral, negative, or positive). In combination with clustering, this approach uncovered sentiment patterns of different topics and their subtopics, reflecting the online public response to mask mandates in Germany. We show that our approach can be used to examine long-term narratives and sentiment dynamics and to identify specific topics that explain peaks of interest in the social media discourse.

Guiding Sentiment Analysis with Hierarchical Text Clustering: Analyzing the German X/Twitter Discourse on Face Masks in the 2020 COVID-19 Pandemic

Social media is an integral part of the daily life of an increasingly large number of people worldwide. Used for entertainment, communication and news updates, it constitutes a source of information that has been extensively used to study human behaviour. Unfortunately, the open nature of social media platforms along with the difficult task of supervising their content has led to a proliferation of misinformation posts. 
In this paper, we aim to identify the textual differences between the profiles of user that share misinformation from questionable sources and those that do not. Our goal is to better understand user behaviour in order to be better equipped to combat this issue. To this end, we identify Twitter (X) accounts of potential misinformation spreaders and apply transformer models specialised in social media to extract characteristics such as sentiment, emotion, topic and presence of hate speech. Our results indicate that, while there may be some differences between the behaviour of users that share misinformation and those that do not, there are no large differences when it comes to the type of content shared.

A Multi-Faceted NLP Analysis of Misinformation Spreaders in Twitter

Social media has become a crucial open-access platform enabling individuals to freely express opinions and share experiences. These platforms contain user-generated content facilitating instantaneous communication and feedback. However, leveraging low-resource language data from Twitter can be challenging due to the scarcity and poor quality of content with significant variations in language use, such as slang and code-switching. Automatically identifying tweets in low-resource languages can also be challenging because Twitter primarily supports high-resource languages; low-resource languages often lack robust linguistic and contextual support. This paper analyzes Kenyan code-switched data from Twitter using four transformer-based pretrained models for sentiment and emotion classification tasks using supervised and semi-supervised methods. We detail the methodology behind data collection, the annotation procedure, and the challenges encountered during the data curation phase. Our results show that XLM-R outperforms other models; for sentiment analysis, XLM-R supervised model achieves the highest accuracy (69.2\%) and F1 score (66.1\%), XLM-R semi-supervised (67.2\% accuracy, 64.1\% F1 score). In emotion analysis, DistilBERT supervised leads in accuracy (59.8\%) and F1 score (31\%), mBERT semi-supervised (accuracy (59\% and F1 score 26.5\%). AfriBERTa models show the lowest accuracy and F1 scores. This indicates that the semi-supervised method's performance is constrained by the small labeled dataset.

RideKE: Leveraging Low-resource Twitter User-generated Content for Sentiment and Emotion Detection on Code-switched RHS Dataset

News headlines often evoke sentiment by intentionally portraying entities in particular ways, making targeted sentiment analysis (TSA) of headlines a worthwhile but difficult task. Due to its subjectivity, creating TSA datasets can involve various annotation paradigms, from descriptive to prescriptive, either encouraging or limiting subjectivity. LLMs are a good fit for TSA due to their broad linguistic and world knowledge and in-context learning abilities, yet their performance depends on prompt design. In this paper, we compare the accuracy of state-of-the-art LLMs and fine-tuned encoder models for TSA of news headlines using descriptive and prescriptive datasets across several languages. Exploring the descriptive--prescriptive continuum, we analyze how performance is affected by prompt prescriptiveness, ranging from plain zero-shot to elaborate few-shot prompts. Finally, we evaluate the ability of LLMs to quantify uncertainty via calibration error and comparison to human label variation. We find that LLMs outperform fine-tuned encoders on descriptive datasets, while calibration and F1-score generally improve with increased prescriptiveness, yet the optimal level varies.

LLMs for Targeted Sentiment in News Headlines: Exploring the Descriptive-Prescriptive Dilemma

This paper explores the task of automatic prediction of text spans in a legal problem description that support a legal area label. We use a corpus of problem descriptions written by laypeople in English that is annotated by practising lawyers. Inherent subjectivity exists in our task because legal area categorisation is a complex task, and lawyers often have different views on a problem. Experiments show that training on majority-voted spans outperforms training on disaggregated ones.

To Aggregate or Not to Aggregate. That is the Question: A Case Study on Annotation Subjectivity in Span Prediction

In machine learning, temporal shifts occur when there are differences between training and test splits in terms of time. For streaming data such as news or social media, models are commonly trained on a fixed corpus from a certain period of time, and they can become obsolete due to the dynamism and evolving nature of online content. This paper focuses on temporal shifts in social media and, in particular, Twitter. We propose a unified evaluation scheme to assess the performance of language models (LMs) under temporal shift on standard social media tasks. LMs are tested on five diverse social media NLP tasks under different temporal settings, which revealed two important findings: (i) the decrease in performance under temporal shift is consistent across different models for entity-focused tasks such as named entity recognition or disambiguation, and hate speech detection, but not significant in the other tasks analysed (i.e., topic and sentiment classification); and (ii) continuous pre-training on the test period does not improve the temporal adaptability of LMs.

A Systematic Analysis on the Temporal Generalization of Language Models in Social Media

Sentiment analysis serves as a pivotal component in Natural Language Processing (NLP). Advancements in multilingual pre-trained models such as XLM-R and mT5 have contributed to the increasing interest in cross-lingual sentiment analysis. The recent emergence in Large Language Models (LLM) has significantly advanced general NLP tasks, however, the capability of such LLMs in cross-lingual sentiment analysis has not been fully studied. This work undertakes an empirical analysis to compare the cross-lingual transfer capability of public Small Multilingual Language Models (SMLM) like XLM-R, against English-centric LLMs such as Llama-3, in the context of sentiment analysis across English, Spanish, French and Chinese. Our findings reveal that among public models, SMLMs exhibit superior zero-shot cross-lingual performance relative to LLMs. However, in few-shot cross-lingual settings, public LLMs demonstrate an enhanced adaptive potential. In addition, we observe that proprietary GPT-3.5 and GPT-4 lead in zero-shot cross-lingual capability, but are outpaced by public models in few-shot scenarios.

Premium content

Entity-Level Sentiment: More than the Sum of Its Parts

Next from ACL 2024

Subjectivity Detection in English News using Large Language Models

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES