MimicProp: Learning to Incorporate Lexicon Knowledge into Distributed Word Representation for Social Media Analysis

Yan, Muheng; Lin, Yu‐Ru; Hwa, Rebecca; Ertugrul, Ali Mert; Guo, Meiqi; Chung, Wen-Ting

doi:10.1609/icwsm.v14i1.7339

Cited by 3 publications

(4 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While Sentiment Analysis models have been robust in analyzing political and social issues (Elghazaly 2016), they may not suffice for measuring nuanced resonance aspects. These require both sentiment polarity and targets, linking it to Stance Detection in NLP (Küc ¸ük and Can 2020; Yan et al 2020). While there are efforts in stance detection, such as predicting attitudes towards specific topics (Darwish et al 2020), or predicting the target and attitude (Dey et al 2017), framing label prediction poses challenges that arise from similar frames possessing variable sentiments and the scarcity of frame-labeled samples, making traditional ML training problematic.…”

Section: Related Workmentioning

confidence: 99%

From Posts to Pavement, or Vice Versa? The Dynamic Interplay between Online Activism and Offline Confrontations

Yan,

Chiang,

Lin

2024

ICWSM

View full text Add to dashboard Cite

This study examines how the relationship between social media discourse and offline confrontations in social movements, focusing on the "Black Lives Matter" (BLM) protests following George Floyd's death in 2020. While social media's role in facilitating social movements is well-documented, its relationship with offline confrontations remains understudied. To bridge this gap, we curated a dataset comprising 108,443 Facebook posts and 1,406 offline BLM protest events. Our analysis categorized online media framing into "consonance" (alignment) and "dissonance" (misalignment) with the perspectives of different involved parties. Our findings indicate a reciprocal relationship between online activism support and offline confrontational occurrences. Online support for the BLM, in particular, was associated with less property damage and fewer confrontational protests in the days that followed. Conversely, offline confrontations amplified online support for the protesters. By illuminating this dynamic, we highlight the multifaceted influence of social media on social movements. Not only does it serve as a platform for information dissemination and mobilization but also plays a pivotal role in shaping public discourse about offline confrontations.

show abstract

Section: Related Workmentioning

confidence: 99%

From Posts to Pavement, or Vice Versa? The Dynamic Interplay between Online Activism and Offline Confrontations

Yan,

Chiang,

Lin

2024

ICWSM

View full text Add to dashboard Cite

show abstract

“…We make use of a publicly available Twitter dataset from a prior study [99]. This dataset contains more than 600k Twitter users that have been identified with liberal or conservative leaning based on their following behaviors [98,99]. From this data, we have identified a total of 3,100 tweets and 2,256 users that are related to "gun rights" or "gun control" discussions.…”

Section: Application Scenario and Datamentioning

confidence: 99%

“…(1) a word2vec [59] embedding trained on a standard Twitter corpus [9], and (2) the attribute-aligned embedding trained with MimicProp algorithm [98] that is optimized for sociolinguistic lexicons. We concatenate the two equal-sized embeddings to generate a 600-dimensional vector for each word.…”

Section: Generating Language Cues Viamentioning

confidence: 99%

Tribe or Not? Critical Inspection of Group Differences Using TribalGram

Ahn

Yan

Lin

et al. 2022

ACM Trans. Interact. Intell. Syst.

View full text Add to dashboard Cite

With the rise of AI and data mining techniques, group profiling and group-level analysis have been increasingly used in many domains, including policy making and direct marketing. In some cases, the statistics extracted from data may provide insights to a group’s shared characteristics; in others, the group-level analysis can lead to problems, including stereotyping and systematic oppression. How can analytic tools facilitate a more conscientious process in group analysis? In this work, we identify a set of accountable group analytics design guidelines to explicate the needs for group differentiation and preventing overgeneralization of a group. Following the design guidelines, we develop TribalGram , a visual analytic suite that leverages interpretable machine learning algorithms and visualization to offer inference assessment, model explanation, data corroboration, and sense-making. Through the interviews with domain experts, we showcase how our design and tools can bring a richer understanding of “groups” mined from the data.

show abstract

“…Text-based embedding learning has been previously employed to understand phenomenon on social media (Alam, Joty, and Imran 2018;Chen, McKeever, and Delany 2019;Yan et al 2020;Bahgat, Wilson, and Magdy 2020), including politics on social media (Oliveira et al 2018;Hemphill and Schöpke-Gonzalez 2020). Some of the existing methods in the literature, which aim at embedding social media users based on the written content, concatenate all the posts of a user as a single document and then train a document level embedding model Pan 2017, 2018;Benton, Arora, and Dredze 2016) 1 .…”

Section: Introductionmentioning

confidence: 99%

American Politicians Diverge Systematically, Indian Politicians do so Chaotically: Text Embeddings as a Window into Party Polarization

Budhiraja¹,

Sharma²,

Agrawal³

et al. 2021

ICWSM

View full text Add to dashboard Cite

Conversations on polarization are increasingly central to discussions of politics and society, but the schisms between parties and states can be hard to identify systematically in what politicians say. In this paper, we demonstrate the use of representation learning as a window into political dialogue on social media through the tweets authored by politicians on Twitter. We propose to embed politicians in a space such that their output embedding vectors represent the content similarity between the two politicians based on their tweets. We further propose a short-text based embedding technique to overcome some of the shortcomings of the previous methods. The learnt embeddings for politicians of India and the United States show two trends. First, that in the US case, we find a clear distinction between Democrats and Republicans, which is also reflected in the coalescing of the states that lean towards each party placing likewise in a graphical space. However, in the Indian case, the federal structure, multiparty system, and linguistic differences clearly manifest in the coalescing political discourse in the largely monolingual north and the scattered regional states. Our work shows ways in which machine learning methods can offer a window into thinking about how polarized political discourses manifest in what politicians say online.

show abstract

MimicProp: Learning to Incorporate Lexicon Knowledge into Distributed Word Representation for Social Media Analysis

Cited by 3 publications

References 26 publications

From Posts to Pavement, or Vice Versa? The Dynamic Interplay between Online Activism and Offline Confrontations

From Posts to Pavement, or Vice Versa? The Dynamic Interplay between Online Activism and Offline Confrontations

Tribe or Not? Critical Inspection of Group Differences Using TribalGram

American Politicians Diverge Systematically, Indian Politicians do so Chaotically: Text Embeddings as a Window into Party Polarization

Contact Info

Product

Resources

About