Skip to content

Explore papers

Filter by research area, year and venue. Every filter is in the address bar, so you can share a view as a link.

LLMs & generative AI Attention Is All You NeedThe authors introduce the Transformer, a sequence transduction model based entirely on self-attention mechanisms that dispenses with recurrence and convolution. The model achieves state-of-the-art BLEU scores on machine translation tasks while being significantly more parallelizable and requiring less training time. Vaswani et al. · Advances in Neural Information Processing… · 2017

Detection methodsDatasets & evaluation A Survey on Hate Speech Detection using Natural Language ProcessingThis survey systematically reviews automatic hate speech detection methods, covering feature extraction techniques (surface features, embeddings, sentiment analysis, lexical resources, linguistic features, knowledge bases), classification approaches (mostly supervised learning with SVMs and neural networks), annotation practices, and key… Schmidt & Wiegand · Fifth International Workshop on Natural… · 2017

Spread & networksDetection methods Topological Recurrent Neural Network for Diffusion PredictionWang et al. propose Topo-LSTM, a novel LSTM architecture for modeling dynamic directed acyclic graphs (DAGs) to predict which node will be activated next in information cascades. The key insight is that information diffusion on networks should be modeled explicitly as diffusion topologies—graphs showing how information spreads from… Wang et al. · ICDM 2017 · 2017

Detection methodsSpread & networks Some Like it Hoax: Automated Fake News Detection in Social NetworksHoaxes on Facebook can be detected with 99% accuracy using user interaction patterns (who "liked" posts) rather than content analysis. Two algorithms—logistic regression and harmonic boolean crowdsourcing—achieve high accuracy even with minimal labeled training data, and knowledge transfers across different Facebook communities. Tacchini et al. · Technical Report, UC Santa Cruz · 2017

Spread & networksDetection methods Turing at SemEval-2017 Task 8: Sequential Approach to Rumour Stance Classification with Branch-LSTMThis paper proposes Branch-LSTM, a sequential neural model for rumour stance classification in Twitter conversations. By decomposing conversation threads into linear branches and modeling them with LSTM layers, the approach captures conversational context while assigning stance labels (Support/Deny/Query/Comment) to each tweet. Kochkina et al. · 11th International Workshop on Semantic… · 2017

Bots & coordinated campaignsDetection methods The Paradigm-Shift of Social Spambots: Evidence, Theories, and Tools for the Arms RaceThis paper provides empirical evidence of a paradigm shift in spambot design on Twitter, where new "social spambots" closely mimic genuine user behavior and successfully evade detection. Neither Twitter's automated systems, human annotators, nor established academic detection tools can effectively identify these accounts; however,… Cresci et al. · WWW 2017 Companion · 2017

Detection methodsLLMs & generative AI CSI: A Hybrid Deep Model for Fake News DetectionCSI is a three-module deep learning model that detects fake news by jointly modeling article text, temporal patterns of user engagement (response), and user behavior patterns (source). The Capture module uses LSTM to learn temporal response signals; the Score module learns user suspiciousness from co-engagement patterns; the Integrate… Ruchansky et al. · CIKM'17 · 2017

Detection methodsFact-checking & verification Truth of Varying Shades: Analyzing Language in Fake News and Political Fact-CheckingThis paper performs a linguistic analysis of fake news across four categories—satire, hoaxes, propaganda, and trusted news—using lexical features (LIWC, subjectivity, hedging, intensifiers). Experiments show that fake news articles use more first-person pronouns, exaggerative language (superlatives, modal adverbs), and hedging, while… Rashkin et al. · 2017 Conference on Empirical Methods in… · 2017

Detection methodsSpread & networks Detect Rumors in Microblog Posts Using Propagation Structure via Kernel LearningThis paper addresses rumor detection in microblog posts by modeling information diffusion as propagation trees and proposing a kernel-based method (PTK) that captures high-order structural patterns differentiating rumors from non-rumors. The approach extends to context-sensitive variants (cPTK) and achieves superior early detection… Ma et al. · ACL · 2017

Spread & networksDatasets & evaluation Quantifying Controversy on Social MediaThis work presents a graph-based pipeline to measure controversy in social media discussions. The authors build conversation graphs from Twitter interactions (retweets, follows, mentions), partition them to identify opposing sides, and apply multiple controversy measures—with a random-walk-based metric (RWC) emerging as the most reliable… Garimella et al. · ACM Transactions on Social Computing · 2017

Spread & networksDatasets & evaluation SemEval-2017 Task 8: RumourEval: Determining rumour veracity and support for rumoursRumourEval is a shared task benchmark for detecting and verifying rumours in social media. It introduces two subtasks: (a) stance classification (SDQC: Support/Deny/Query/Comment) of replies to rumourous claims, and (b) veracity prediction (true/false) of source tweets. Derczynski et al. · 11th International Workshop on Semantic… · 2017

Psychology & interventionsPlatforms, policy & society Inoculating the Public against Misinformation about Climate ChangeLarge-scale experiments (N=2,167) demonstrate that misinformation about scientific consensus on climate change nullifies the effect of consensus messaging. Pre-emptive inoculation—warning people about misleading tactics and providing refutations—preserves two-thirds of the consensus effect. van der Linden et al. · Global Challenges · 2017

Platforms, policy & societyFact-checking & verification Defining "Fake News": A typology of scholarly definitionsThis paper reviews 34 academic studies (2003–2017) that use the term "fake news" and proposes a unified typology based on two dimensions: level of facticity (reliance on facts) and author's intention to deceive. The typology identifies six distinct types—satire, parody, fabrication, manipulation, advertising, and propaganda—clarifying… Tandoc et al. · Digital Journalism · 2017

Bots & coordinated campaignsPlatforms, policy & society Detecting Bots on Russian Political TwitterAutomated Twitter accounts (bots) represent 50% of activity in Russian political discourse. Using supervised ensemble classifiers on 3.8M accounts and 18M tweets from 2014–2015, the authors achieve 95% precision in detecting bots while maintaining 75% recall. Stukal et al. · Big Data · 2017

Spread & networksPsychology & interventions Spreadable Spectacle in Digital Culture: Civic Expression, Fake News, and the Role of Media Literacies in "Post-Fact" SocietyThis paper explores how spectacle—media constructs outside ordinary routine that capitalize on polarization—spreads through digital culture during the 2016 U.S. election. Using the Pizzagate conspiracy theory as a case study, the authors argue that citizen-driven spectacle leverages homophilous networks, memes, and appropriated cultural… Mihailidis & Viotty · American Behavioral Scientist · 2017

Psychology & interventions The Psychology of Conspiracy TheoriesPeople are drawn to conspiracy theories via three distinct psychological motives: epistemic (seeking understanding and certainty), existential (seeking control and security), and social (maintaining positive self and group image). However, experimental evidence suggests conspiracy theories ultimately frustrate rather than satisfy these… Douglas et al. · Current Directions in Psychological Science · 2017

Spread & networksPsychology & interventions Emotion shapes the diffusion of moralized content in social networksMoral-emotional language (words combining moral and emotional content) increases message diffusion in social networks by ~20% per word. This effect is stronger within political in-groups than out-groups, explaining how moral and political ideas spread within polarized "echo chambers" on platforms like Twitter. Brady et al. · National Academy of Sciences · 2017

Spread & networksPlatforms, policy & society Information Disorder: Toward an Interdisciplinary Framework for Research and Policy MakingThis Council of Europe report introduces a comprehensive framework for understanding "information disorder" via three types (mis-, dis-, mal-information) defined by falseness and intent-to-harm. It examines the agents, messages, and interpreters involved, the creation-production-distribution lifecycle, and proposes 34 recommendations… Wardle & Derakhshan · Council of Europe report · 2017

Spread & networksBots & coordinated campaigns The Spread of True and False News OnlineAnalyzes 126,000 verified true and false news cascades from Twitter (2006–2017) covering 3 million users and 4.5 million shares. False news diffuses significantly farther, faster, deeper, and more broadly than truth across all categories, with false political news spreading fastest. Vosoughi et al. · MIT Initiative on the Digital Economy… · 2017

Psychology & interventionsPlatforms, policy & society Neutralizing misinformation through inoculation: Exposing misleading argumentation techniques reduces their influencePre-emptively exposing people to flawed arguments before they encounter misinformation (inoculation) reduces that misinformation's influence. Two experiments tested this on climate change: false-balance media coverage (equal weight to consensus vs. contrarian scientists) and fake-expert arguments (using unqualified scientists to… Cook et al. · PLOS ONE · 2017

Spread & networksPlatforms, policy & society Social Media and Fake News in the 2016 ElectionAllcott and Gentzkov provide the first comprehensive empirical evidence on fake news exposure during the 2016 US election. Drawing on a curated database of 156 fake news articles, web traffic data, and a post-election survey of 1,200 adults, they estimate the average American saw and remembered 1.14 fake news articles during the election… Allcott & Gentzkov · Journal of Economic Perspectives · 2017

Spread & networksPsychology & interventions Filter Bubbles, Echo Chambers, and Online News ConsumptionThis paper examines whether technological changes (social media, search engines) increase ideological segregation in online news consumption by analyzing web-browsing data from 50,000 active news readers. They find that while social media and search are associated with higher segregation than direct browsing, the overall effects are… Flaxman et al. · Public Opinion Quarterly · 2016

Datasets & evaluationFact-checking & verification Are You a Racist or Am I Seeing Things? Annotator Influence on Hate Speech Detection on TwitterThis paper compares hate speech detection systems trained on expert annotations (from feminist and anti-racism activists) versus amateur annotations (crowdsourced via CrowdFlower) on 6,909 tweets. Systems trained on expert annotations significantly outperform those trained on amateur data, and amateur annotators systematically label more… Waseem · 2016 EMNLP Workshop on Natural Language… · 2016

Bots & coordinated campaignsDatasets & evaluation BotOrNot: A System to Evaluate Social BotsBotOrNot is a publicly available service and API that evaluates whether a Twitter account is controlled by a human or automated bot. Using 1,000+ features across six categories (network, user, friends, temporal, content, sentiment), the system applies Random Forest classification to achieve 0.95 AUC on a labeled dataset of 15k bots and… Davis et al. · 25th International Conference Companion on… · 2016