Researcher profile

Preslav Nakov

Preslav Nakov contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate

Computation and Language Artificial Intelligence Machine Learning Information Retrieval Social and Information Networks Computer Vision Multimedia Neural and Evolutionary Computing cs.CY

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author

32works

0followers

9topics

4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.

Building this graph slice

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2026arXiv

A Multi-View Media Profiling Suite: Resources, Evaluation, and Analysis

News outlets shape public opinion at a scale that makes automated detection of political bias and factuality essential. However, the field still lacks unified resources, comprehensive evaluations across diverse approaches, and systematic analyses of the representations and fusion strategies that matter most, especially under label sparsity and dataset diversity. In addition, there is little empirical work reporting broad, observation-driven findings about what consistently works, what fails, and why. We address these gaps through four main contributions. First, we introduce MBFC-2025, a large-scale label set covering approximately 2,600 outlets from Media Bias/Fact Check (MBFC). Second, we construct multiview representations for ACL-2020 (Panayotov et al., 2022), which includes around 900 outlets, as well as for MBFC-2025. These representations span Alexa graphs, hyperlink graphs, LLM-derived graphs, articles, and Wikipedia descriptions. Third, we provide a systematic evaluation and analysis of embedding views and fusion strategies, including a reinforcement learning-based fusion variant. Fourth, we conduct extensive experiments that achieve state-of-the-art results on ACL-2020 and establish strong benchmarks on MBFC-2025.

preprint2026arXiv

Cultural Benchmarking of LLMs in Standard and Dialectal Arabic Dialogues

There is a significant gap in evaluating cultural reasoning in LLMs using conversational datasets that capture culturally rich and dialectal contexts. Most Arabic benchmarks focus on short text snippets in Modern Standard Arabic (MSA), overlooking the cultural nuances that naturally arise in dialogues. To address this gap, we introduce ArabCulture-Dialogue, a culturally grounded conversational dataset covering 13 Arabic-speaking countries, in both MSA and each country's respective dialect, spanning 12 daily-life topics and 54 fine-grained subtopics. We utilize the dataset to form three benchmarking tasks: (i) multiple-choice cultural reasoning, (ii) machine translation between MSA and dialects, and (iii) dialect-steering generation. Our experiments indicate that the performance gap between MSA and Arabic dialects still exists, whereby the models perform worse on all three tasks in the dialectal setup, compared to the MSA one.

preprint2026arXiv

EngTrace: A Symbolic Benchmark for Verifiable Process Supervision of Engineering Reasoning

Large Language Models (LLMs) are increasingly entering specialized, safety-critical engineering workflows governed by strict quantitative standards and immutable physical laws, making rigorous evaluation of their reasoning capabilities imperative. However, existing benchmarks such as MMLU, MATH, and HumanEval assess isolated cognitive skills, failing to capture the physically grounded reasoning central to engineering, where scientific principles, quantitative modeling, and practical constraints must converge. To enable verifiable process supervision in engineering, we introduce EngTrace, a symbolic benchmark comprising 90 templates across three major engineering branches, nine core domains and 20 distinct areas. Through domain-aware parameterization, we generate 1,350 unique, contamination-resistant test cases to stress-test generalization. Moving beyond outcome matching, we introduce a verifiable two-stage evaluation framework that uses a tiered protocol to validate intermediate reasoning traces alongside final answers through automated procedural checks and a heterogeneous AI Tribunal. Our evaluation of 24 leading LLMs reveals a distinct trade-off between numeric precision and trace fidelity, identifying a complexity cliff where abstract mathematical pre-training fails to translate into the integrative reasoning required for advanced engineering tasks.

preprint2026arXiv

FMI_SU_Yotkova_Kastreva at SemEval-2026 Task 13: Lightweight Detection of LLM-Generated Code via Stylometric Signals

SemEval-2026 Task 13 investigates machine-generated code detection across multiple programming languages and application scenarios, asking participating systems to generalize to unseen languages and domains. This paper describes our participation in Subtask A (binary classification) and explores both pretrained code encoders and lightweight feature-based methods. We design ratio-based features that are less sensitive to snippet length. To support the extraction of descriptiveness-related signals, we use parsing engines and a programming-language classifier. Additionally, we train a separate code-vs-text line classifier to identify raw natural language segments embedded within samples. We combine a shallow decision tree with heuristic rules derived from data analysis to produce the final predictions. Our approach is computationally efficient, requires only CPU resources for training, and achieves near-instant inference time, offering a lightweight alternative to large pretrained models.

preprint2026arXiv

How Does Prefix Matter in Reasoning Model Tuning?

Recent alignment studies commonly remove introductory boilerplate phrases from supervised fine-tuning (SFT) datasets. This work challenges that assumption. We hypothesize that safety- and reasoning-oriented prefix sentences serve as lightweight alignment signals that can guide model decoding toward safer and more coherent responses. To examine this, we fine-tune three R1 series models across three core model capabilities: reasoning (mathematics, coding), safety, and factuality, systematically varying prefix inclusion from 0% to 100%. Results show that prefix-conditioned SFT improves both safety and reasoning performance, yielding up to +6% higher Safe@1 accuracy on adversarial benchmarks (WildJailbreak, StrongReject) and +7% improvement on GSM8K reasoning. However, factuality and coding tasks show marginal or negative effects, indicating that prefix-induced narrowing of the search space benefits structured reasoning. Token-level loss analysis further reveals that prefix tokens such as "revised" and "logically" incur higher gradient magnitudes, acting as alignment anchors that stabilize reasoning trajectories. Our findings suggest that prefix conditioning offers a scalable and interpretable mechanism for improving reasoning safety, serving as an implicit form of alignment that complements traditional reward-based methods.

preprint2026arXiv

Instruction-Guided Poetry Generation in Arabic and Its Dialects

Poetry has long been a central art form for Arabic speakers, serving as a powerful medium of expression and cultural identity. While modern Arabic speakers continue to value poetry, existing research on Arabic poetry within Large Language Models (LLMs) has primarily focused on analysis tasks such as interpretation or metadata prediction, e.g., rhyme schemes and titles. In contrast, our work addresses the practical aspect of poetry creation in Arabic by introducing controllable generation capabilities to assist users in writing poetry. Specifically, we present a large-scale, carefully curated instruction-based dataset in Modern Standard Arabic (MSA) and various Arabic dialects. This dataset enables tasks such as writing, revising, and continuing poems based on predefined criteria, including style and rhyme, as well as performing poetry analysis. Our experiments show that fine-tuning LLMs on this dataset yields models that can effectively generate poetry that is aligned with user requirements, based on both automated metrics and human evaluation with native Arabic speakers. The data and the code are available at https://github.com/mbzuai-nlp/instructpoet-ar

preprint2026arXiv

MuDRiC: Multi-Dialect Reasoning for Arabic Commonsense Validation

Commonsense validation evaluates whether a sentence aligns with everyday human understanding, a critical capability for developing robust natural language understanding systems. While substantial progress has been made in English, the task remains underexplored in Arabic, particularly given its rich linguistic diversity. Existing Arabic resources have primarily focused on Modern Standard Arabic (MSA), leaving regional dialects underrepresented despite their prevalence in spoken contexts. To bridge this gap, we present two key contributions. We introduce MuDRiC, an extended Arabic commonsense dataset incorporating multiple dialects. To the best of our knowledge, this is the first Arabic multi-dialect commonsense reasoning dataset. We further propose a novel method adapting Graph Convolutional Networks (GCNs) to Arabic commonsense reasoning, which enhances semantic relationship modeling for improved commonsense validation. Our experimental results demonstrate that this approach consistently outperforms the baseline of direct language model fine-tuning. Overall, our work enhances Arabic natural language understanding by providing a foundational dataset and a new method for handling its complex variations. Data and code are available at https://github.com/KareemElozeiri/MuDRiC.

preprint2026arXiv

The Cylindrical Representation Hypothesis for Language Model Steering

Steering is a widely used technique for controlling large language models, yet its effects are often unstable and hard to predict. Existing theoretical accounts are largely based on the Linear Representation Hypothesis (LRH). While LRH assumes that concepts can be orthogonalized for lossless control, this idealized mapping fails in real representations and cannot account for the observed unpredictability of steering. By relaxing LRH's orthogonality assumption while preserving linear representations, we show that overlapping concept contributions naturally yield a sample-specific axis-orthogonal structure. We formalize this as the Cylindrical Representation Hypothesis (CRH). In CRH, a central axis captures the main difference between concept absence and presence and drives concept generation. A surrounding normal plane controls steering sensitivity by determining how easily the axis can activate the target concept. Within this plane, only specific sensitive sectors strongly facilitate concept activation, while other sectors can suppress or delay it. While the surrounding normal plane can be reliably identified from difference vectors, the sensitive sector cannot, introducing intrinsic uncertainty at the sector level. This uncertainty provides a principled explanation for why steering outcomes often fluctuate even when using well-aligned directions. Our experiments verify the existence of the cylindrical structure and demonstrate that CRH provides a valid and practical way to interpret model steering behavior in real settings: https://github.com/mbzuai-nlp/CRH.

preprint2026arXiv

The Geometry of Forgetting: Temporal Knowledge Drift as an Independent Axis in LLM Representations

Large language models confidently produce outdated answers, and no existing method can detect them. We show this is not an engineering failure but a structural one: temporal drift, whether a stored fact has changed since training, is encoded as a direction in the residual stream geometrically orthogonal to both correctness and uncertainty. Any method operating on correctness or uncertainty signals is therefore blind to drift by construction. We verify this across six instruction-tuned models. A linear probe trained directly on drift labels achieves AUROC $0.83$--$0.95$; methods based on token entropy, semantic entropy, CCS, and SAPLMA all remain near chance ($0.49$--$0.57$). Five tests confirm the geometric orthogonality: weight cosines ($|\cos| \leq 0.14$), score correlations ($|r| \leq 0.20$), bidirectional null-space projection ($|Δ| \leq 0.008$), iterative null-space projection with $k{=}10$, and difference-of-means dissociation. Mechanistically, the MLP retrieval circuit produces identical dynamics for stale recall and confabulation ($r > 0.81$, six models), explaining why output confidence cannot separate them. A cross-cutoff experiment holds inputs constant and varies only the model: the probe fires on the model whose training predates the fact's transition and stays silent otherwise ($P(A{>}B) = 0.975$--$0.998$, twelve model pairs), confirming it reads model-internal knowledge state rather than input properties. Our code and datasets will be publicly released.

preprint2026arXiv

YaPO: Learnable Sparse Activation Steering Vectors for Domain Adaptation

Steering Large Language Models (LLMs) through activation interventions has emerged as a lightweight alternative to fine-tuning for alignment and personalization. Recent work on Bi-directional Preference Optimization (BiPO) shows that dense steering vectors can be learned directly from preference data in a Direct Preference Optimization (DPO) fashion, enabling control over truthfulness, hallucinations, and safety behaviors. However, dense steering vectors often entangle multiple latent factors due to neuron multi-semanticity, limiting their effectiveness and stability in fine-grained settings such as cultural alignment, where closely related values and behaviors (e.g., among Middle Eastern cultures) must be distinguished. In this paper, we propose Yet another Policy Optimization (YaPO), a \textit{reference-free} method that learns \textit{sparse steering vectors} in the latent space of a Sparse Autoencoder (SAE). By optimizing sparse codes, YaPO produces disentangled, interpretable, and efficient steering directions. Empirically, we show that YaPO converges faster, achieves stronger performance, and exhibits improved training stability compared to dense steering baselines. Beyond cultural alignment, YaPO generalizes to a range of alignment-related behaviors, including hallucination, wealth-seeking, jailbreak, and power-seeking. Importantly, YaPO preserves general knowledge, with no measurable degradation on MMLU. Overall, our results show that YaPO provides a general recipe for efficient, stable, and fine-grained alignment of LLMs, with broad applications to controllability and domain adaptation. The associated code and data are publicly available\footnote{https://github.com/MBZUAI-Paris/YaPO}.

preprint2022arXiv

A Neighbourhood Framework for Resource-Lean Content Flagging

We propose a novel framework for cross-lingual content flagging with limited target-language data, which significantly outperforms prior work in terms of predictive performance. The framework is based on a nearest-neighbour architecture. It is a modern instantiation of the vanilla k-nearest neighbour model, as we use Transformer representations in all its components. Our framework can adapt to new source-language instances, without the need to be retrained from scratch. Unlike prior work on neighbourhood-based approaches, we encode the neighbourhood information based on query--neighbour interactions. We propose two encoding schemes and we show their effectiveness using both qualitative and quantitative analysis. Our evaluation results on eight languages from two different datasets for abusive language detection show sizable improvements of up to 9.5 F1 points absolute (for Italian) over strong baselines. On average, we achieve 3.6 absolute F1 points of improvement for the three languages in the Jigsaw Multilingual dataset and 2.14 points for the WUL dataset.

preprint2022arXiv

A Survey on Stance Detection for Mis- and Disinformation Identification

Understanding attitudes expressed in texts, also known as stance detection, plays an important role in systems for detecting false information online, be it misinformation (unintentionally false) or disinformation (intentionally false information). Stance detection has been framed in different ways, including (a) as a component of fact-checking, rumour detection, and detecting previously fact-checked claims, or (b) as a task in its own right. While there have been prior efforts to contrast stance detection with other related tasks such as argumentation mining and sentiment analysis, there is no existing survey on examining the relationship between stance detection and mis- and disinformation detection. Here, we aim to bridge this gap by reviewing and analysing existing work in this area, with mis- and disinformation in focus, and discussing lessons learnt and future challenges.

preprint2022arXiv

Detecting and Understanding Harmful Memes: A Survey

The automatic identification of harmful content online is of major concern for social media platforms, policymakers, and society. Researchers have studied textual, visual, and audio content, but typically in isolation. Yet, harmful content often combines multiple modalities, as in the case of memes, which are of particular interest due to their viral nature. With this in mind, here we offer a comprehensive survey with a focus on harmful memes. Based on a systematic analysis of recent literature, we first propose a new typology of harmful memes, and then we highlight and summarize the relevant state of the art. One interesting finding is that many types of harmful memes are not really studied, e.g., such featuring self-harm and extremism, partly due to the lack of suitable datasets. We further find that existing datasets mostly capture multi-class scenarios, which are not inclusive of the affective spectrum that memes can represent. Another observation is that memes can propagate globally through repackaging in different languages and that they can also be multilingual, blending different cultures. We conclude by highlighting several challenges related to multimodal semiotics, technological constraints, and non-trivial social engagement, and we present several open-ended aspects such as delineating online harm and empirically examining related frameworks and assistive interventions, which we believe will motivate and drive future research.

preprint2022arXiv

Detecting the Role of an Entity in Harmful Memes: Techniques and Their Limitations

Harmful or abusive online content has been increasing over time, raising concerns for social media platforms, government agencies, and policymakers. Such harmful or abusive content can have major negative impact on society, e.g., cyberbullying can lead to suicides, rumors about COVID-19 can cause vaccine hesitance, promotion of fake cures for COVID-19 can cause health harms and deaths. The content that is posted and shared online can be textual, visual, or a combination of both, e.g., in a meme. Here, we describe our experiments in detecting the roles of the entities (hero, villain, victim) in harmful memes, which is part of the CONSTRAINT-2022 shared task, as well as our system for the task. We further provide a comparative analysis of different experimental settings (i.e., unimodal, multimodal, attention, and augmentation). For reproducibility, we make our experimental code publicly available. \url{https://github.com/robi56/harmful_memes_block_fusion}

preprint2022arXiv

DISARM: Detecting the Victims Targeted by Harmful Memes

Internet memes have emerged as an increasingly popular means of communication on the Web. Although typically intended to elicit humour, they have been increasingly used to spread hatred, trolling, and cyberbullying, as well as to target specific individuals, communities, or society on political, socio-cultural, and psychological grounds. While previous work has focused on detecting harmful, hateful, and offensive memes, identifying whom they attack remains a challenging and underexplored area. Here we aim to bridge this gap. In particular, we create a dataset where we annotate each meme with its victim(s) such as the name of the targeted person(s), organization(s), and community(ies). We then propose DISARM (Detecting vIctimS targeted by hARmful Memes), a framework that uses named entity recognition and person identification to detect all entities a meme is referring to, and then, incorporates a novel contextualized multimodal deep neural network to classify whether the meme intends to harm these entities. We perform several systematic experiments on three test setups, corresponding to entities that are (a) all seen while training, (b) not seen as a harmful target on training, and (c) not seen at all on training. The evaluation results show that DISARM significantly outperforms ten unimodal and multimodal systems. Finally, we show that DISARM is interpretable and comparatively more generalizable and that it can reduce the relative error rate for harmful target identification by up to 9 points absolute over several strong multimodal rivals.

preprint2022arXiv

Leaf: Multiple-Choice Question Generation

Testing with quiz questions has proven to be an effective way to assess and improve the educational process. However, manually creating quizzes is tedious and time-consuming. To address this challenge, we present Leaf, a system for generating multiple-choice questions from factual text. In addition to being very well suited for the classroom, Leaf could also be used in an industrial setting, e.g., to facilitate onboarding and knowledge sharing, or as a component of chatbots, question answering systems, or Massive Open Online Courses (MOOCs). The code and the demo are available on https://github.com/KristiyanVachev/Leaf-Question-Generation.

preprint2022arXiv

On the Effect of Dropping Layers of Pre-trained Transformer Models

Transformer-based NLP models are trained using hundreds of millions or even billions of parameters, limiting their applicability in computationally constrained environments. While the number of parameters generally correlates with performance, it is not clear whether the entire network is required for a downstream task. Motivated by the recent work on pruning and distilling pre-trained models, we explore strategies to drop layers in pre-trained models, and observe the effect of pruning on downstream GLUE tasks. We were able to prune BERT, RoBERTa and XLNet models up to 40%, while maintaining up to 98% of their original performance. Additionally we show that our pruned models are on par with those built using knowledge distillation, both in terms of size and performance. Our experiments yield interesting observations such as, (i) the lower layers are most critical to maintain downstream task performance, (ii) some tasks such as paraphrase detection and sentence similarity are more robust to the dropping of layers, and (iii) models trained using a different objective function exhibit different learning patterns and w.r.t the layer dropping.

preprint2022arXiv

QCRI's COVID-19 Disinformation Detector: A System to Fight the COVID-19 Infodemic in Social Media

Fighting the ongoing COVID-19 infodemic has been declared as one of the most important focus areas by the World Health Organization since the onset of the COVID-19 pandemic. While the information that is consumed and disseminated consists of promoting fake cures, rumors, and conspiracy theories to spreading xenophobia and panic, at the same time there is information (e.g., containing advice, promoting cure) that can help different stakeholders such as policy-makers. Social media platforms enable the infodemic and there has been an effort to curate the content on such platforms, analyze and debunk them. While a majority of the research efforts consider one or two aspects (e.g., detecting factuality) of such information, in this study we focus on a multifaceted approach, including an API,\url{https://app.swaggerhub.com/apis/yifan2019/Tanbih/0.8.0/} and a demo system,\url{https://covid19.tanbih.org}, which we made freely and publicly available. We believe that this will facilitate researchers and different stakeholders. A screencast of the API services and demo is available.\url{https://youtu.be/zhbcSvxEKMk}

preprint2022arXiv

TeamX@DravidianLangTech-ACL2022: A Comparative Analysis for Troll-Based Meme Classification

The spread of fake news, propaganda, misinformation, disinformation, and harmful content online raised concerns among social media platforms, government agencies, policymakers, and society as a whole. This is because such harmful or abusive content leads to several consequences to people such as physical, emotional, relational, and financial. Among different harmful content \textit{trolling-based} online content is one of them, where the idea is to post a message that is provocative, offensive, or menacing with an intent to mislead the audience. The content can be textual, visual, a combination of both, or a meme. In this study, we provide a comparative analysis of troll-based memes classification using the textual, visual, and multimodal content. We report several interesting findings in terms of code-mixed text, multimodal setting, and combining an additional dataset, which shows improvements over the majority baseline.

preprint2022arXiv

The Role of Context in Detecting Previously Fact-Checked Claims

Recent years have seen the proliferation of disinformation and fake news online. Traditional approaches to mitigate these issues is to use manual or automatic fact-checking. Recently, another approach has emerged: checking whether the input claim has previously been fact-checked, which can be done automatically, and thus fast, while also offering credibility and explainability, thanks to the human fact-checking and explanations in the associated fact-checking article. Here, we focus on claims made in a political debate and we study the impact of modeling the context of the claim: both on the source side, i.e., in the debate, as well as on the target side, i.e., in the fact-checking explanation document. We do this by modeling the local context, the global context, as well as by means of co-reference resolution, and multi-hop reasoning over the sentences of the document describing the fact-checked claim. The experimental results show that each of these represents a valuable information source, but that modeling the source-side context is most important, and can yield 10+ points of absolute improvement over a state-of-the-art model.

preprint2022arXiv

The Spread of Propaganda by Coordinated Communities on Social Media

Large-scale manipulations on social media have two important characteristics: (i) use of propaganda to influence others, and (ii) adoption of coordinated behavior to spread it and to amplify its impact. Despite the connection between them, these two characteristics have so far been considered in isolation. Here we aim to bridge this gap. In particular, we analyze the spread of propaganda and its interplay with coordinated behavior on a large Twitter dataset about the 2019 UK general election. We first propose and evaluate several metrics for measuring the use of propaganda on Twitter. Then, we investigate the use of propaganda by different coordinated communities that participated in the online debate. The combination of the use of propaganda and coordinated behavior allows us to uncover the authenticity and harmfulness of the different communities. Finally, we compare our measures of propaganda and coordination with automation (i.e., bot) scores and Twitter suspensions, revealing interesting trends. From a theoretical viewpoint, we introduce a methodology for analyzing several important dimensions of online behavior that are seldom conjointly considered. From a practical viewpoint, we provide new insights into authentic and inauthentic online activities during the 2019 UK general election.

preprint2020arXiv

A Survey on Computational Propaganda Detection

Propaganda campaigns aim at influencing people's mindset with the purpose of advancing a specific agenda. They exploit the anonymity of the Internet, the micro-profiling ability of social networks, and the ease of automatically creating and managing coordinated networks of accounts, to reach millions of social network users with persuasive messages, specifically targeted to topics each individual user is sensitive to, and ultimately influencing the outcome on a targeted issue. In this survey, we review the state of the art on computational propaganda detection from the perspective of Natural Language Processing and Network Analysis, arguing about the need for combined efforts between these communities. We further discuss current challenges and future research directions.

preprint2020arXiv

aschern at SemEval-2020 Task 11: It Takes Three to Tango: RoBERTa, CRF, and Transfer Learning

We describe our system for SemEval-2020 Task 11 on Detection of Propaganda Techniques in News Articles. We developed ensemble models using RoBERTa-based neural architectures, additional CRF layers, transfer learning between the two subtasks, and advanced post-processing to handle the multi-label nature of the task, the consistency between nested spans, repetitions, and labels from similar spans in training. We achieved sizable improvements over baseline fine-tuned RoBERTa models, and the official evaluation ranked our system 3rd (almost tied with the 2nd) out of 36 teams on the span identification subtask with an F1 score of 0.491, and 2nd (almost tied with the 1st) out of 31 teams on the technique classification subtask with an F1 score of 0.62.

preprint2020arXiv

Can We Spot the "Fake News" Before It Was Even Written?

Given the recent proliferation of disinformation online, there has been also growing research interest in automatically debunking rumors, false claims, and "fake news." A number of fact-checking initiatives have been launched so far, both manual and automatic, but the whole enterprise remains in a state of crisis: by the time a claim is finally fact-checked, it could have reached millions of users, and the harm caused could hardly be undone. An arguably more promising direction is to focus on fact-checking entire news outlets, which can be done in advance. Then, we could fact-check the news before it was even written: by checking how trustworthy the outlets that published it is. We describe how we do this in the Tanbih news aggregator, which makes readers aware of what they are reading. In particular, we develop media profiles that show the general factuality of reporting, the degree of propagandistic content, hyper-partisanship, leading political ideology, general frame of reporting, and stance with respect to various claims and topics.

preprint2020arXiv

CheckThat! at CLEF 2020: Enabling the Automatic Identification and Verification of Claims in Social Media

We describe the third edition of the CheckThat! Lab, which is part of the 2020 Cross-Language Evaluation Forum (CLEF). CheckThat! proposes four complementary tasks and a related task from previous lab editions, offered in English, Arabic, and Spanish. Task 1 asks to predict which tweets in a Twitter stream are worth fact-checking. Task 2 asks to determine whether a claim posted in a tweet can be verified using a set of previously fact-checked claims. Task 3 asks to retrieve text snippets from a given set of Web pages that would be useful for verifying a target tweet's claim. Task 4 asks to predict the veracity of a target tweet's claim using a set of Web pages and potentially useful snippets in them. Finally, the lab offers a fifth task that asks to predict the check-worthiness of the claims made in English political debates and speeches. CheckThat! features a full evaluation framework. The evaluation is carried out using mean average precision or precision at rank k for ranking tasks, and F1 for classification tasks.

preprint2020arXiv

Overview of CheckThat! 2020: Automatic Identification and Verification of Claims in Social Media

We present an overview of the third edition of the CheckThat! Lab at CLEF 2020. The lab featured five tasks in two different languages: English and Arabic. The first four tasks compose the full pipeline of claim verification in social media: Task 1 on check-worthiness estimation, Task 2 on retrieving previously fact-checked claims, Task 3 on evidence retrieval, and Task 4 on claim verification. The lab is completed with Task 5 on check-worthiness estimation in political debates and speeches. A total of 67 teams registered to participate in the lab (up from 47 at CLEF 2019), and 23 of them actually submitted runs (compared to 14 at CLEF 2019). Most teams used deep neural networks based on BERT, LSTMs, or CNNs, and achieved sizable improvements over the baselines on all tasks. Here we describe the tasks setup, the evaluation results, and a summary of the approaches used by the participants, and we discuss some lessons learned. Last but not least, we release to the research community all datasets from the lab as well as the evaluation scripts, which should enable further research in the important tasks of check-worthiness estimation and automatic claim verification.

preprint2020arXiv

Predicting the Topical Stance of Media and Popular Twitter Users

Discovering the stances of media outlets and influential people on current, debatable topics is important for social statisticians and policy makers. Many supervised solutions exist for determining viewpoints, but manually annotating training data is costly. In this paper, we propose a cascaded method that uses unsupervised learning to ascertain the stance of Twitter users with respect to a polarizing topic by leveraging their retweet behavior; then, it uses supervised learning based on user labels to characterize both the general political leaning of online media and of popular Twitter users, as well as their stance with respect to the target polarizing topic. We evaluate the model by comparing its predictions to gold labels from the Media Bias/Fact Check website, achieving 82.6% accuracy.

preprint2020arXiv

Prta: A System to Support the Analysis of Propaganda Techniques in the News

Recent events, such as the 2016 US Presidential Campaign, Brexit and the COVID-19 "infodemic", have brought into the spotlight the dangers of online disinformation. There has been a lot of research focusing on fact-checking and disinformation detection. However, little attention has been paid to the specific rhetorical and psychological techniques used to convey propaganda messages. Revealing the use of such techniques can help promote media literacy and critical thinking, and eventually contribute to limiting the impact of "fake news" and disinformation campaigns. Prta (Propaganda Persuasion Techniques Analyzer) allows users to explore the articles crawled on a regular basis by highlighting the spans in which propaganda techniques occur and to compare them on the basis of their use of propaganda techniques. The system further reports statistics about the use of such techniques, overall and over time, or according to filtering criteria specified by the user based on time interval, keywords, and/or political orientation of the media. Moreover, it allows users to analyze any text or URL through a dedicated interface or via an API. The system is available online: https://www.tanbih.org/prta

preprint2020arXiv

Team Alex at CLEF CheckThat! 2020: Identifying Check-Worthy Tweets With Transformer Models

While misinformation and disinformation have been thriving in social media for years, with the emergence of the COVID-19 pandemic, the political and the health misinformation merged, thus elevating the problem to a whole new level and giving rise to the first global infodemic. The fight against this infodemic has many aspects, with fact-checking and debunking false and misleading claims being among the most important ones. Unfortunately, manual fact-checking is time-consuming and automatic fact-checking is resource-intense, which means that we need to pre-filter the input social media posts and to throw out those that do not appear to be check-worthy. With this in mind, here we propose a model for detecting check-worthy tweets about COVID-19, which combines deep contextualized text representations with modeling the social context of the tweet. We further describe a number of additional experiments and comparisons, which we believe should be useful for future research as they provide some indication about what techniques are effective for the task. Our official submission to the English version of CLEF-2020 CheckThat! Task 1, system Team_Alex, was ranked second with a MAP score of 0.8034, which is almost tied with the wining system, lagging behind by just 0.003 MAP points absolute.

preprint2020arXiv

That is a Known Lie: Detecting Previously Fact-Checked Claims

The recent proliferation of "fake news" has triggered a number of responses, most notably the emergence of several manual fact-checking initiatives. As a result and over time, a large number of fact-checked claims have been accumulated, which increases the likelihood that a new claim in social media or a new statement by a politician might have already been fact-checked by some trusted fact-checking organization, as viral claims often come back after a while in social media, and politicians like to repeat their favorite statements, true or false, over and over again. As manual fact-checking is very time-consuming (and fully automatic fact-checking has credibility issues), it is important to try to save this effort and to avoid wasting time on claims that have already been fact-checked. Interestingly, despite the importance of the task, it has been largely ignored by the research community so far. Here, we aim to bridge this gap. In particular, we formulate the task and we discuss how it relates to, but also differs from, previous work. We further create a specialized dataset, which we release to the research community. Finally, we present learning-to-rank experiments that demonstrate sizable improvements over state-of-the-art retrieval and textual similarity approaches.

preprint2020arXiv

Unsupervised User Stance Detection on Twitter

We present a highly effective unsupervised framework for detecting the stance of prolific Twitter users with respect to controversial topics. In particular, we use dimensionality reduction to project users onto a low-dimensional space, followed by clustering, which allows us to find core users that are representative of the different stances. Our framework has three major advantages over pre-existing methods, which are based on supervised or semi-supervised classification. First, we do not require any prior labeling of users: instead, we create clusters, which are much easier to label manually afterwards, e.g., in a matter of seconds or minutes instead of hours. Second, there is no need for domain- or topic-level knowledge either to specify the relevant stances (labels) or to conduct the actual labeling. Third, our framework is robust in the face of data skewness, e.g., when some users or some stances have greater representation in the data. We experiment with different combinations of user similarity features, dataset sizes, dimensionality reduction methods, and clustering algorithms to ascertain the most effective and most computationally efficient combinations across three different datasets (in English and Turkish). We further verified our results on additional tweet sets covering six different controversial topics. Our best combination in terms of effectiveness and efficiency uses retweeted accounts as features, UMAP for dimensionality reduction, and Mean Shift for clustering, and yields a small number of high-quality user clusters, typically just 2--3, with more than 98\% purity. The resulting user clusters can be used to train downstream classifiers. Moreover, our framework is robust to variations in the hyper-parameter values and also with respect to random initialization.

preprint2020arXiv

What Was Written vs. Who Read It: News Media Profiling Using Text Analysis and Social Media Context

Predicting the political bias and the factuality of reporting of entire news outlets are critical elements of media profiling, which is an understudied but an increasingly important research direction. The present level of proliferation of fake, biased, and propagandistic content online, has made it impossible to fact-check every single suspicious claim, either manually or automatically. Alternatively, we can profile entire news outlets and look for those that are likely to publish fake or biased content. This approach makes it possible to detect likely "fake news" the moment they are published, by simply checking the reliability of their source. From a practical perspective, political bias and factuality of reporting have a linguistic aspect but also a social context. Here, we study the impact of both, namely (i) what was written (i.e., what was published by the target medium, and how it describes itself on Twitter) vs. (ii) who read it (i.e., analyzing the readers of the target medium on Facebook, Twitter, and YouTube). We further study (iii) what was written about the target medium on Wikipedia. The evaluation results show that what was written matters most, and that putting all information sources together yields huge improvements over the current state-of-the-art.

Institution

Affiliation not imported yet

This author record came from a source that does not expose affiliation metadata. Once the author claims the profile or we enrich the record from another provider, this section will link to the concrete institution.

Topic footprint

Fields this researcher appears in

Computation and Language Artificial Intelligence Machine Learning Information Retrieval Social and Information Networks Computer Vision Multimedia Neural and Evolutionary Computing cs.CY

Source provenance