Reza Farahbakhsh

Khouloud Mnassri, Praboda Rajapaksha, Reza Farahbakhsh et al.

With the freedom of communication provided in online social media, hate speech has increasingly generated. This leads to cyber conflicts affecting social life at the individual and national levels. As a result, hateful content classification is becoming increasingly demanded for filtering hate content before being sent to the social networks. This paper focuses on classifying hate speech in social media using multiple deep models that are implemented by integrating recent transformer-based language models such as BERT, and neural networks. To improve the classification performances, we evaluated with several ensemble techniques, including soft voting, maximum value, hard voting and stacking. We used three publicly available Twitter datasets (Davidson, HatEval2019, OLID) that are generated to identify offensive languages. We fused all these datasets to generate a single dataset (DHO dataset), which is more balanced across different labels, to perform multi-label classification. Our experiments have been held on Davidson dataset and the DHO corpora. The later gave the best overall results, especially F1 macro score, even it required more resources (time execution and memory). The experiments have shown good results especially the ensemble models, where stacking gave F1 score of 97% on Davidson dataset and aggregating ensembles 77% on the DHO dataset.

Khouloud Mnassri, Praboda Rajapaksha, Reza Farahbakhsh et al.

The rise of emergence of social media platforms has fundamentally altered how people communicate, and among the results of these developments is an increase in online use of abusive content. Therefore, automatically detecting this content is essential for banning inappropriate information, and reducing toxicity and violence on social media platforms. The existing works on hate speech and offensive language detection produce promising results based on pre-trained transformer models, however, they considered only the analysis of abusive content features generated through annotated datasets. This paper addresses a multi-task joint learning approach which combines external emotional features extracted from another corpora in dealing with the imbalanced and scarcity of labeled datasets. Our analysis are using two well-known Transformer-based models, BERT and mBERT, where the later is used to address abusive content detection in multi-lingual scenarios. Our model jointly learns abusive content detection with emotional features by sharing representations through transformers' shared encoder. This approach increases data efficiency, reduce overfitting via shared representations, and ensure fast learning by leveraging auxiliary information. Our findings demonstrate that emotional knowledge helps to more reliably identify hate speech and offensive language across datasets. Our hate speech detection Multi-task model exhibited 3% performance improvement over baseline models, but the performance of multi-task models were not significant for offensive language detection task. More interestingly, in both tasks, multi-task models exhibits less false positive errors compared to single task scenario.

Cong Huy Nguyen, Son Dinh Nguyen, Guanlin Li et al.

Automated medical report generation for 3D PET/CT imaging is fundamentally challenged by the high-dimensional nature of volumetric data and a critical scarcity of annotated datasets, particularly for low-resource languages. Current black-box methods map whole volumes to reports, ignoring the clinical workflow of analyzing localized Regions of Interest (RoIs) to derive diagnostic conclusions. In this paper, we bridge this gap by introducing VietPET-RoI, the first large-scale 3D PET/CT dataset with fine-grained RoI annotation for a low-resource language, comprising 600 PET/CT samples and 1,960 manually annotated RoIs, paired with corresponding clinical reports. Furthermore, to demonstrate the utility of this dataset, we propose HiRRA, a novel framework that mimics the professional radiologist diagnostic workflow by employing graph-based relational modules to capture dependencies between RoI attributes. This approach shifts from global pattern matching toward localized clinical findings. Additionally, we introduce new clinical evaluation metrics, namely RoI Coverage and RoI Quality Index, that measure both RoI localization accuracy and attribute description fidelity using LLM-based extraction. Extensive evaluation demonstrates that our framework achieves SOTA performance, surpassing existing models by 19.7% in BLEU and 4.7% in ROUGE-L, while achieving a remarkable 45.8% improvement in clinical metrics, indicating enhanced clinical reliability and reduced hallucination. Our code and dataset are available on GitHub.

Sergey Berezin, Reza Farahbakhsh, Noel Crespi

We introduce a novel class of adversarial attacks on toxicity detection models that exploit language models' failure to interpret spatially structured text in the form of ASCII art. To evaluate the effectiveness of these attacks, we propose ToxASCII, a benchmark designed to assess the robustness of toxicity detection systems against visually obfuscated inputs. Our attacks achieve a perfect Attack Success Rate (ASR) across a diverse set of state-of-the-art large language models and dedicated moderation tools, revealing a significant vulnerability in current text-only moderation systems.

Sergey Berezin, Reza Farahbakhsh, Noel Crespi

The fundamental problem in toxicity detection task lies in the fact that the toxicity is ill-defined. This causes us to rely on subjective and vague data in models' training, which results in non-robust and non-accurate results: garbage in - garbage out. This work suggests a new, stress-level-based definition of toxicity designed to be objective and context-aware. On par with it, we also describe possible ways of applying this new definition to dataset creation and model training.

Aditya Narayan Sankaran, Reza Farahbakhsh, Noel Crespi

Abusive speech detection is becoming increasingly important as social media shifts towards voice-based interaction, particularly in multilingual and low-resource settings. Most current systems rely on automatic speech recognition (ASR) followed by text-based hate speech classification, but this pipeline is vulnerable to transcription errors and discards prosodic information carried in speech. We investigate whether Contrastive Language-Audio Pre-training (CLAP) can support abusive speech detection directly from audio. Using the ADIMA dataset, we evaluate CLAP-based representations under few-shot supervised contrastive adaptation in cross-lingual and leave-one-language-out settings, with zero-shot prompting included as an auxiliary analysis. Our results show that CLAP yields strong cross-lingual audio representations across ten Indic languages, and that lightweight projection-only adaptation achieves competitive performance with respect to fully supervised systems trained on complete training data. However, the benefits of few-shot adaptation are language-dependent and not monotonic with shot size. These findings suggest that contrastive audio-text models provide a promising basis for cross-lingual audio abuse detection in low-resource settings, while also indicating that transfer remains incomplete and language-specific in important ways.

Sergey Berezin, Reza Farahbakhsh, Noel Crespi

We introduce a simple yet efficient sentence-level attack on black-box toxicity detector models. By adding several positive words or sentences to the end of a hateful message, we are able to change the prediction of a neural network and pass the toxicity detection system check. This approach is shown to be working on seven languages from three different language families. We also describe the defence mechanism against the aforementioned attack and discuss its limitations.

Minh Khoi Nguyen, Dai Lam Le, Amir Reza Jafari et al.

Large vision-language models (VLMs) demonstrate strong performance in medical image understanding, but frequently generate clinically plausible yet incorrect statements, raising significant safety concerns. Existing medical hallucination benchmarks primarily focus on 2D imaging with one-shot diagnostic questions, offering limited insight into whether predictions are grounded in correct localization and abnormality identification, allowing critical reasoning errors to remain hidden behind seemingly correct diagnoses. We introduce Med-StepBench, the first large-scale benchmark for step-wise hallucination detection in 3D oncological PET/CT, comprising over 12,000 images and more than 1,000,000 image-statement pairs across volumetric and multi-view 2D data, which decomposes clinical reasoning into four expert-designed diagnostic stages. Using clinician-verified annotations, we perform the first step-level evaluation of general-purpose and medical VLMs, revealing systematic failure modes obscured by aggregate accuracy metrics. Furthermore, we show that current VLMs are highly susceptible to adversarial yet clinically plausible intermediate explanations, which significantly amplify hallucinations despite contradictory visual evidence. Together, our findings highlight fundamental limitations in grounding multi-step clinical reasoning and establish Med-StepBench as a rigorous benchmark for developing safer and more reliable medical VLMs.

Hrishikesh Dutta, Roberto Minerva, Reza Farahbakhsh et al.

Modern IoT and sensor networks generate vast amounts of data, posing significant challenges for storage, transmission, and real-time processing. Traditional approaches, such as compressive sensing and machine learning-based compression, often suffer from computational inefficiencies and irreversible data loss. This paper introduces Information Density as a quantitative metric to support sensor deployment and enable AI-driven virtual sensing. We propose a framework that leverages spatial, temporal and inter-modal correlations among sensor signals to perform sensing tasks even in the absence of physical sensors. Two complementary measures: (i) Phase in Eigen Space and (ii) Mutual Information, are developed to quantify and assess information density, enabling the selection of optimal sensor configurations across both intra-modality and cross-modality scenarios. Validated using real-world data from Madrid's smart city infrastructure, this framework demonstrates the feasibility of replacing physical sensors with virtual ones under bounded error conditions (e.g., achieving $<3.21\%$ mean error with a single sensor). The results highlight the potential for scalable and energy-efficient sensing systems in smart environments.

Sergey Berezin, Reza Farahbakhsh, Noel Crespi

We present a novel class of jailbreak adversarial attacks on LLMs, termed Task-in-Prompt (TIP) attacks. Our approach embeds sequence-to-sequence tasks (e.g., cipher decoding, riddles, code execution) into the model's prompt to indirectly generate prohibited inputs. To systematically assess the effectiveness of these attacks, we introduce the PHRYGE benchmark. We demonstrate that our techniques successfully circumvent safeguards in six state-of-the-art language models, including GPT-4o and LLaMA 3.2. Our findings highlight critical weaknesses in current LLM safety alignments and underscore the urgent need for more sophisticated defence strategies. Warning: this paper contains examples of unethical inquiries used solely for research purposes.

Aditya Narayan Sankaran, Reza Farahbakhsh, Noel Crespi

Online abusive content detection, particularly in low-resource settings and within the audio modality, remains underexplored. We investigate the potential of pre-trained audio representations for detecting abusive language in low-resource languages, in this case, in Indian languages using Few Shot Learning (FSL). Leveraging powerful representations from models such as Wav2Vec and Whisper, we explore cross-lingual abuse detection using the ADIMA dataset with FSL. Our approach integrates these representations within the Model-Agnostic Meta-Learning (MAML) framework to classify abusive language in 10 languages. We experiment with various shot sizes (50-200) evaluating the impact of limited data on performance. Additionally, a feature visualization study was conducted to better understand model behaviour. This study highlights the generalization ability of pre-trained models in low-resource scenarios and offers valuable insights into detecting abusive language in multilingual contexts.

Shaghayegh Najari, Davood Rafiee, Mostafa Salehi et al.

Social bots play a significant role in many online social networks (OSN) as they imitate human behavior. This fact raises difficult questions about their capabilities and potential risks. Given the recent advances in Generative AI (GenAI), social bots are capable of producing highly realistic and complex content that mimics human creativity. As the malicious social bots emerge to deceive people with their unrealistic content, identifying them and distinguishing the content they produce has become an actual challenge for numerous social platforms. Several approaches to this problem have already been proposed in the literature, but the proposed solutions have not been widely evaluated. To address this issue, we evaluate the behavior of a text-based bot detector in a competitive environment where some scenarios are proposed: \textit{First}, the tug-of-war between a bot and a bot detector is examined. It is interesting to analyze which party is more likely to prevail and which circumstances influence these expectations. In this regard, we model the problem as a synthetic adversarial game in which a conversational bot and a bot detector are engaged in strategic online interactions. \textit{Second}, the bot detection model is evaluated under attack examples generated by a social bot; to this end, we poison the dataset with attack examples and evaluate the model performance under this condition. \textit{Finally}, to investigate the impact of the dataset, a cross-domain analysis is performed. Through our comprehensive evaluation of different categories of social bots using two benchmark datasets, we were able to demonstrate some achivement that could be utilized in future works.

Sergey Berezin, Reza Farahbakhsh, Noel Crespi

Most toxicity detection models treat toxicity as an intrinsic property of text, overlooking the role of context in shaping its impact. Drawing on interdisciplinary research, we reconceptualise toxicity as a socially emergent stress signal. We introduce a new framework for toxicity detection, including a formal definition and metric, and validate our approach on a novel dataset, demonstrating improved contextual sensitivity and adaptability.

Aditya Narayan Sankaran, Reza Farahbakhsh, Noel Crespi

Code switching, particularly between Korean and English, has become a defining feature of modern K-pop, reflecting both aesthetic choices and global market strategies. This paper is a primary investigation into the linguistic strategies employed in K-pop songs that achieve global chart success, with a focus on the role of code-switching and English lyric usage. A dataset of K-pop songs that appeared on the Billboard Hot 100 and Global 200 charts from 2017 to 2025, spanning 14 groups and 8 solo artists, was compiled. Using this dataset, the proportion of English and Korean lyrics, the frequency of code-switching, and other stylistic features were analysed. It was found that English dominates the linguistic landscape of globally charting K-pop songs, with both male and female performers exhibiting high degrees of code-switching and English usage. Statistical tests indicated no significant gender-based differences, although female solo artists tend to favour English more consistently. A classification task was also performed to predict performer gender from lyrics, achieving macro F1 scores up to 0.76 using multilingual embeddings and handcrafted features. Finally, differences between songs charting on the Hot 100 versus the Global 200 were examined, suggesting that, while there is no significant gender difference in English, higher English usage may be more critical for success in the US-focused Hot 100. The findings highlight how linguistic choices in K-pop lyrics are shaped by global market pressures and reveal stylistic patterns that reflect performer identity and chart context.

Amir Reza Jafari, Behnam Heidary, Reza Farahbakhsh et al.

These days different platforms such as social media provide their clients from different backgrounds and languages the possibility to connect and exchange information. It is not surprising anymore to see comments from different languages in posts published by international celebrities or data providers. In this era, understanding cross languages content and multilingualism in natural language processing (NLP) are hot topics, and multiple efforts have tried to leverage existing technologies in NLP to tackle this challenging research problem. In this survey, we provide a comprehensive overview of the existing literature with a focus on transfer learning techniques in multilingual tasks. We also identify potential opportunities for further research in this domain.

Marzieh Mozafari, Reza Farahbakhsh, Noel Crespi

Disparate biases associated with datasets and trained classifiers in hateful and abusive content identification tasks have raised many concerns recently. Although the problem of biased datasets on abusive language detection has been addressed more frequently, biases arising from trained classifiers have not yet been a matter of concern. Here, we first introduce a transfer learning approach for hate speech detection based on an existing pre-trained language model called BERT and evaluate the proposed model on two publicly available datasets annotated for racism, sexism, hate or offensive content on Twitter. Next, we introduce a bias alleviation mechanism in hate speech detection task to mitigate the effect of bias in training set during the fine-tuning of our pre-trained BERT-based model. Toward that end, we use an existing regularization method to reweight input samples, thereby decreasing the effects of high correlated training set' s n-grams with class labels, and then fine-tune our pre-trained BERT-based model with the new re-weighted samples. To evaluate our bias alleviation mechanism, we employ a cross-domain approach in which we use the trained classifiers on the aforementioned datasets to predict the labels of two new datasets from Twitter, AAE-aligned and White-aligned groups, which indicate tweets written in African-American English (AAE) and Standard American English (SAE) respectively. The results show the existence of systematic racial bias in trained classifiers as they tend to assign tweets written in AAE from AAE-aligned group to negative classes such as racism, sexism, hate, and offensive more often than tweets written in SAE from White-aligned. However, the racial bias in our classifiers reduces significantly after our bias alleviation mechanism is incorporated. This work could institute the first step towards debiasing hate speech and abusive language detection systems.

Koosha Zarei, Reza Farahbakhsh, Noel Crespi

Impersonators on Online Social Networks such as Instagram are playing an important role in the propagation of the content. These entities are the type of nefarious fake accounts that intend to disguise a legitimate account by making similar profiles. In addition to having impersonated profiles, we observed a considerable engagement from these entities to the published posts of verified accounts. Toward that end, we concentrate on the engagement of impersonators in terms of active and passive engagements which is studied in three major communities including ``Politician'', ``News agency'', and ``Sports star'' on Instagram. Inside each community, four verified accounts have been selected. Based on the implemented approach in our previous studies, we have collected 4.8K comments, and 2.6K likes across 566 posts created from 3.8K impersonators during 7 months. Our study shed light into this interesting phenomena and provides a surprising observation that can help us to understand better how impersonators engaging themselves inside Instagram in terms of writing Comments and leaving Likes.

Marzieh Mozafari, Reza Farahbakhsh, Noel Crespi

Generated hateful and toxic content by a portion of users in social media is a rising phenomenon that motivated researchers to dedicate substantial efforts to the challenging direction of hateful content identification. We not only need an efficient automatic hate speech detection model based on advanced machine learning and natural language processing, but also a sufficiently large amount of annotated data to train a model. The lack of a sufficient amount of labelled hate speech data, along with the existing biases, has been the main issue in this domain of research. To address these needs, in this study we introduce a novel transfer learning approach based on an existing pre-trained language model called BERT (Bidirectional Encoder Representations from Transformers). More specifically, we investigate the ability of BERT at capturing hateful context within social media content by using new fine-tuning methods based on transfer learning. To evaluate our proposed approach, we use two publicly available datasets that have been annotated for racism, sexism, hate, or offensive content on Twitter. The results show that our solution obtains considerable performance on these datasets in terms of precision and recall in comparison to existing approaches. Consequently, our model can capture some biases in data annotation and collection process and can potentially lead us to a more accurate model.

Praboda Rajapaksha, Reza Farahbakhsh, Noel Crespi et al.

Social networking sites (SNSs) facilitate the sharing of ideas and information through different types of feedback including publishing posts, leaving comments and other type of reactions. However, some comments or feedback on SNSs are inconsiderate and offensive, and sometimes this type of feedback has a very negative effect on a target user. The phenomenon known as flaming goes hand-in-hand with this type of posting that can trigger almost instantly on SNSs. Most popular users such as celebrities, politicians and news media are the major victims of the flaming behaviors and so detecting these types of events will be useful and appreciated. Flaming event can be monitored and identified by analyzing negative comments received on a post. Thus, our main objective of this study is to identify a way to detect flaming events in SNS using a sentiment prediction method. We use a deep Neural Network (NN) model that can identity sentiments of variable length sentences and classifies the sentiment of SNSs content (both comments and posts) to discover flaming events. Our deep NN model uses Word2Vec and FastText word embedding methods as its training to explore which method is the most appropriate. The labeled dataset for training the deep NN is generated using an enhanced lexicon based approach. Our deep NN model classifies the sentiment of a sentence into five classes: Very Positive, Positive, Neutral, Negative and Very Negative. To detect flaming incidents, we focus only on the comments classified into the Negative and Very Negative classes. As a use-case, we try to explore the flaming phenomena in the news media domain and therefore we focused on news items posted by three popular news media on Facebook (BBCNews, CNN and FoxNews) to train and test the model.

Reza Farahbakhsh, Angel Cuevas, Ruben Cuevas et al.

Today's Internet traffic is mostly dominated by multimedia content and the prediction is that this trend will intensify in the future. Therefore, main Internet players, such as ISPs, content delivery platforms (e.g. Youtube, Bitorrent, Netflix, etc) or CDN operators, need to understand the evolution of multimedia content availability and popularity in order to adapt their infrastructures and resources to satisfy clients requirements while they minimize their costs. This paper presents a thorough analysis on the evolution of multimedia content available in BitTorrent. Specifically, we analyze the evolution of four relevant metrics across different content categories: content availability, content popularity, content size and user's feedback. To this end we leverage a large-scale dataset formed by 4 snapshots collected from the most popular BitTorrent portal, namely The Pirate Bay, between Nov. 2009 and Feb. 2012. Overall our dataset is formed by more than 160k content that attracted more than 185M of download sessions.

Saeedreza Shehnepoor, Mostafa Salehi, Reza Farahbakhsh et al.

Nowadays, a big part of people rely on available content in social media in their decisions (e.g. reviews and feedback on a topic or product). The possibility that anybody can leave a review provide a golden opportunity for spammers to write spam reviews about products and services for different interests. Identifying these spammers and the spam content is a hot topic of research and although a considerable number of studies have been done recently toward this end, but so far the methodologies put forth still barely detect spam reviews, and none of them show the importance of each extracted feature type. In this study, we propose a novel framework, named NetSpam, which utilizes spam features for modeling review datasets as heterogeneous information networks to map spam detection procedure into a classification problem in such networks. Using the importance of spam features help us to obtain better results in terms of different metrics experimented on real-world review datasets from Yelp and Amazon websites. The results show that NetSpam outperforms the existing methods and among four categories of features; including review-behavioral, user-behavioral, reviewlinguistic, user-linguistic, the first type of features performs better than the other categories.