Predicting Email and Article Clickthroughs with Domain-adaptive Language Models

Jaidka, Kokil; Goyal, Tanya; Chhaya, Niyati

doi:10.1145/3201064.3201071

Cited by 11 publications

(5 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Due to the length of news headlines, here the frequency of each clickbait phrase or expression is not considered in our feature set as it leads to many zeros in our feature matrix. Such dictionaries have been successfully applied in clickbait detection [9,21,42].…”

Section: Semantic-levelmentioning

confidence: 99%

Fake News Early Detection

et al. 2020

View full text Add to dashboard Cite

Massive dissemination of fake news and its potential to erode democracy has increased the demand for accurate fake news detection. Recent advancements in this area have proposed novel techniques that aim to detect fake news by exploring how it propagates on social networks. Nevertheless, to detect fake news at an early stage, i.e., when it is published on a news outlet but not yet spread on social media, one cannot rely on news propagation information as it does not exist. Hence, there is a strong need to develop approaches that can detect fake news by focusing on news content. In this article, a theorydriven model is proposed for fake news detection. The method investigates news content at various levels: lexicon-level, syntax-level, semantic-level, and discourse-level. We represent news at each level, relying on well-established theories in social and forensic psychology. Fake news detection is then conducted within a supervised machine learning framework. As an interdisciplinary research, our work explores potential fake news patterns, enhances the interpretability in fake news feature engineering, and studies the relationships among fake news, deception/disinformation, and clickbaits. Experiments conducted on two real-world datasets indicate the proposed method can outperform the state-of-the-art and enable fake news early detection when there is limited content information. CCS Concepts: • Human-centered computing → Collaborative and social computing theory, concepts and paradigms; • Computing methodologies → Natural language processing; Machine learning; • Security and privacy → Social aspects of security and privacy; • Applied computing → Sociology; Computer forensics;

show abstract

Section: Semantic-levelmentioning

confidence: 99%

Fake News Early Detection

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Ensuring such data's confidentiality and ethical use is paramount, and future studies should rigorously adhere to privacy guidelines and data protection regulations. For instance, technologies developed in [44] offer a way to fine-tune existing predictive models with an organization's data, which is increasingly being considered in recent paradigms that adapt Large Language Models for commercial or enterprise AI goals [85].…”

Section: Discussionmentioning

confidence: 99%

“…Prior work on email behavior has focused on both marketing and professional emails. Studies that predict whether recipients will open marketing emails have focused on measuring the sentiment in the subject line [43] as well as other linguistic features and explored the need to adapt to new domains [44]. On the other hand, in professional communication, email responses have been predicted based on linguistic signals [45,46].…”

Section: Email Behaviormentioning

confidence: 99%

Building a Multimodal Classifier of Email Behavior: Towards a Social Network Understanding of Organizational Communication

Shah,

Jaidka,

Ungar

et al. 2023

Information

Self Cite

View full text Add to dashboard Cite

Within organizational settings, communication dynamics are influenced by various factors, such as email content, historical interactions, and interpersonal relationships. We introduce the Email MultiModal Architecture (EMMA) to model these dynamics and predict future communication behavior. EMMA uses data related to an email sender’s social network, performance metrics, and peer endorsements to predict the probability of receiving an email response. Our primary analysis is based on a dataset of 0.6 million corporate emails from 4320 employees between 2012 and 2014. By integrating features that capture a sender’s organizational influence and likability within a multimodal structure, EMMA offers improved performance over models that rely solely on linguistic attributes. Our findings indicate that EMMA enhances email reply prediction accuracy by up to 12.5% compared to leading text-centric models. EMMA also demonstrates high accuracy on other email datasets, reinforcing its utility and generalizability in diverse contexts. Our findings recommend the need for multimodal approaches to better model communication patterns within organizations and teams and to better understand how relationships and histories shape communication trajectories.

show abstract

“…Under this theory, the factor that actually matters here is whether the topic of the subject line matches with the recipients' preferences. The longer subject lines might get lower open rates but higher action rates, because only those who are interested in it will open it [22,55]. (2) Top section: The traditional theory is that users would pay more attention to the top positions during browsing [44].…”

Section: Bulk Emailmentioning

confidence: 99%

Multi-Objective Personalization in Multi-Stakeholder Organizational Bulk E-mail

Kong

Zhang

Sun

et al. 2022

Proc. ACM Hum.-Comput. Interact.

View full text Add to dashboard Cite

Bulk email is often used in organizations to communicate "important-to-organization'' messages such as policy changes, organizational plans, and administrative updates. However, normal employees may prefer messages more relevant to their jobs or interests. Organizations face the challenge of balancing prioritizing the messages they prefer employees to know (tactical goals) while maintaining employees' positive experiences with these bulk emails, then they continue to read these emails in the future (strategic goals). Could personalization help organizations achieve these tactical and strategic goals? In an 8-week field experiment with a university newsletter, we implemented a 4x5x5 factorial design on personalizing subject lines, top news, and message order based on both the employees' and the organization's preferences. We measured these designs' influences on the open/interest/recognition/read-in-detail rate of the whole newsletter and the single messages within it. We found that ''important-to-organization'' messages only got higher recognition rates when being put on subject lines / top news (tactical goal). Mixing them with employee-preferred messages in top news did not bring further improvement to their own recognition rates but could improve the whole newsletter's recognition rate. Only when the top news solely contained the employee-preferred messages were the employees slightly more interested in the newsletter (strategic goal). We further analyze on which topics the employees and the organization's preferences conflicted. Finally, we discuss the design suggestions for organizational bulk email.

show abstract

Predicting Email and Article Clickthroughs with Domain-adaptive Language Models

Cited by 11 publications

References 23 publications

Fake News Early Detection

Fake News Early Detection

Building a Multimodal Classifier of Email Behavior: Towards a Social Network Understanding of Organizational Communication

Multi-Objective Personalization in Multi-Stakeholder Organizational Bulk E-mail

Contact Info

Product

Resources

About