Sentiment_Classification
-
When Correct Demonstrations Hurt: Rethinking the Role of Exemplars in In-Context Learning
2026-05-25 21:52:49Chenghao Qiu, Chunli Peng, Yufeng Yang, Kuan-Hao Huang, Yi ZhouAbstract
In-context learning (ICL) is often motivated by the intuition that demonstrations help because they provide correct input-output examples. However, we reveal a counterintuitive phenomenon: correctness does not guarantee exemplar utility, and some correct demonstrations can even reduce ICL accuracy. To study this correctness-utility gap, we introduce task-preserving perturbations, where only the exemplar input is changed, while the example remains a correct instance of the same task. Concretely, each perturbed exemplar is assigned the target induced by the task mapping. This framework covers both label-updating perturbations, where task-relevant semantics change and targets are recomputed, and stricter target-preserving perturbations, where the original target remains valid. We formalize the resulting failure mode as contextual evidence shift: task-preserving perturbations can change the effective mixture of evidence used by the model for contextual inference, thereby separating exemplar correctness from exemplar utility. Across sentiment classification, logical reasoning, and math word problems, we find that task-preserving perturbed demonstrations can substantially degrade ICL performance, especially for smaller models, harder tasks, and higher perturbation ratios. Our results show that robust ICL requires evaluating not only whether demonstrations are correct, but also how they influence contextual inference. Code is available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2605.26350
PDF
-
Grammatically-Guided Sparse Attention for Efficient and Interpretable Transformers
2026-05-23 11:11:19Spandan PratyushAbstract
The quadratic complexity of self-attention in Transformer models remains a significant bottleneck for processing long sequences and deploying large language models efficiently. For this approach, there has been significant research into Sparse Attention, and Deepseek Sparse Attention has combined various methods of creating segments of tokens to reduce the time complexity. This paper introduces a novel approach, Grammatically-Guided Sparse Attention, which constrains attention computations based on the grammatical roles of tokens. By leveraging Parts-of-Speech (POS) tags, attention masks are dynamically generated that enforce linguistically coherent connections between tokens, reducing the computational graph without sacrificing essential linguistic dependencies. Two masking strategies are proposed and evaluated: a hard mask that strictly allows only predefined grammatical interactions, and a soft mask that biases attention towards these interactions. The experiments, conducted on the SST-2 sentiment classification task using a DistilBERT-like architecture, demonstrate that Grammatically-Guided Sparse Attention maintains comparable accuracy to full attention while significantly reducing the theoretical computational overhead. Preliminary results show accuracy values of 0.8200 for hard masking and 0.8165 for soft masking, closely matching the 0.8200 of full attention, providing a path towards more efficient, interpretable, and linguistically-informed Transformer architectures.
Abstract (translated)
URL
https://arxiv.org/abs/2605.24518
PDF
-
SentimentLens: Reconciling Sentiment and Ratings via Dual-Modality in the Hospitality Sector
2026-05-22 17:21:27Dineth Jayakody, Pasindu Thenahandi, Sampath JayarathnaAbstract
Online travel platforms generate vast volumes of user-generated hotel reviews, offering rich opportunities to understand traveler experiences at scale. However, transforming unstructured textual feedback into structured, actionable insights remains a challenging task. This paper presents SentimentLens, a scalable analysis system based on Aspect-Based Sentiment Analysis that performs knowledge extraction from unstructured hotel reviews and organizes them into interpretable service categories. SentimentLens integrates aspect term extraction, aspect sentiment classification, semantic category assignment, and multi-level analytical modules to support region-level, hotel-level, and category-level evaluation. The system is designed to operate across different geographic contexts and hospitality settings. To demonstrate its practical utility, we apply SentimentLens to a large real-world dataset of over 10,000 publicly available hotel reviews. Through extensive analysis, the framework reveals how traveler sentiment varies across regions, service categories, and hotel archetypes. We further implement a cross-modal reconciliation of textual sentiment and numerical ratings to identify latent operational conflicts, structural inconsistencies in service quality, and high-impact improvement opportunities using importance--performance and entropy-based analyses. The results show that SentimentLens effectively transforms large-scale unstructured reviews into actionable intelligence, supporting data-driven decision-making for hospitality management and tourism policy. While demonstrated using a national case study, the proposed system is generalizable to other destinations and review-driven service domains.
Abstract (translated)
URL
https://arxiv.org/abs/2606.00084
PDF
-
RADAR: Relative Angular Divergence Across Representations
2026-05-21 20:51:00Xavier Cadet, Mateusz Nowak, Peter ChinAbstract
Machine learning methods rely on data. However, gathering suitable data can be challenging due to availability constraints, cost, or the need for domain expertise. Expanding datasets with additional sources is a common response to limited data, yet this practice does not always improve downstream performance and can sometimes lead to a loss of performance, known as negative transfer. We propose RADAR, a simple, geometrically grounded metric for estimating cross-domain transferability in foundation models. RADAR analyzes the layer-wise evolution of representations by measuring angular alignments and relative changes in distance along layer-to-layer displacement trajectories, and by comparing empirical distributions of within-domain and cross-domain dynamics. We hypothesize that domain transferability is related to the divergence between these trajectory distributions. We evaluate the metric across multiple modalities, including cross-lingual sentiment classification with text embedding models and cross-domain image classification with foundation vision models. Across several settings, RADAR provides competitive predictive performance relative to existing transferability metrics on several vision and text benchmarks, with particularly strong results when domain transitions are smooth or cleanly separated. Our ablations further suggest that the effectiveness of transferability estimation depends on the geometry of the model's internal representation space, with different modalities favoring different topological formulations.
Abstract (translated)
URL
https://arxiv.org/abs/2605.23028
PDF
-
From TF-IDF to Transformers: A Comparative and Ensemble Approach to Sentiment Classification
2026-05-21 05:00:12Dip Biswas Shanto, Mitali Yadav, Prajwal Panth, Suresh Chandra SatapathyAbstract
Sentiment analysis, also referred to as opinion mining, primarily tries to extract opinion from any text-based data. In the context of movie reviews and critics, sentimental analysis can be a helpful tool to predict whether a movie review is generally positive or negative. It can be difficult for the ML models to understand the context or metaphysical sentiment accurately, as ML models rely largely on statistical word representations. The objective of this paper is to examine and categorise movie reviews into positive and negative sentiments. Diverse machine learning models are considered in doing so, and Natural Language Processing (NLP) methodologies are employed for data preprocessing and model assessment. The IMDb dataset is used. Specifically, Naive Bayes, Logistic Regression, Support Vector Machines (SVM), LightGBM, LSTM, and transformer-based models such as RoBERTa and DistilBERT were evaluated. After a lot of testing with accuracy, precision, recall, F1-score, and ROC-AUC, RoBERTa performed better than all the other models, with an accuracy of 93.02%. A soft voting ensemble that combined all the models also improved classification performance, showing that model ensembling works well for sentiment analysis.
Abstract (translated)
URL
https://arxiv.org/abs/2605.22003
PDF
-
Beyond Sentiment Classification: A Generative Framework for Emotion Intensity Evaluation in Text
2026-05-15 20:32:29Francesco A. Fabozzi, Dasol Kim, William N. GoetzmannAbstract
We introduce a novel approach to emotion modeling that shifts the focus from identification to evaluation, addressing the limitations of discrete classification in applied domains such as finance. By constructing a dataset of emotional intensity scores and fine-tuning open-weight generative language models to output continuous values from 0-100, we demonstrate a more expressive, generalizable framework for sentiment and emotion analysis. Our findings not only outperform classification baselines but also reveal surprising generalization capabilities and transfer effects to related constructs such as sentiment and arousal. This work contributes to the interdisciplinary recontextualization of NLP by introducing emotion intensity evaluation as an alternative to classification, arguing that this shift better aligns with the needs of domains--such as finance--where the degree of emotional content is central to interpretation and decision-making.
Abstract (translated)
URL
https://arxiv.org/abs/2605.16613
PDF
-
From Llama to Cria: Scaling Down Neural Networks via Neuron-Level Spectral Structural Importance Evaluation
2026-05-14 10:58:41Yongyu WangAbstract
This paper proposes a neuron pruning framework based on neuron-level spectral structural importance evaluation. Given a trained neural network, we record the hidden states of each hidden layer during inference and model neurons as graph nodes, with hidden states treated as graph signals. Using ideas from graph signal processing, we infer layer-wise input and output graphs that characterize the structural relationships among neurons before and after each layer transformation. We then evaluate the spectral structural importance of neurons by analyzing the transformation between these graphs based on spectral graph theory. Neurons with high spectral structural importance are regarded as strongly involved in the internal representation transformation and are therefore preserved, while neurons with low importance scores are selected as pruning candidates. The pruning process is conducted iteratively until a predefined effective parameter reduction target is reached. Instead of fine-tuning after every pruning step, the proposed strategy first removes low-importance neurons to obtain a compact architecture and then applies a final recovery fine-tuning stage to restore task performance. By connecting neuron pruning with graph signal processing and spectral structural analysis, the proposed framework offers a principled way to reduce neural network size while maintaining solution quality. Experimental results on CIFAR-10 image classification and SST-2 sentiment classification show that our method can effectively remove low-importance neurons and achieve compact networks with competitive performance after recovery fine-tuning.
Abstract (translated)
URL
https://arxiv.org/abs/2605.18860
PDF
-
FinMoji: A Framework for Emoji-driven Sentiment Analysis in Financial Social Media
2026-05-10 10:39:14Ahmed Mahrous, Roberto Di PietroAbstract
This paper explores the use of emojis in financial sentiment analysis, focusing on the social media platform StockTwits. Emojis, increasingly prevalent in digital communication, have potential as compact indicators of investor sentiment, which can be critical for predicting market trends. Our study examines whether emojis alone can serve as reliable proxies for financial sentiment and how they compare with traditional text-based analysis. We conduct a series of experiments using logistic regression and transformer models. We further analyze the performance, computational efficiency, and data requirements of emoji-based versus text-based sentiment classification. Using a balanced dataset of about 528,000 emoji-containing StockTwits posts, we find that emoji-only models achieve F1 approximately 0.75, lower than text-emoji combined models, which achieve F1 approximately 0.88, but with far lower computational cost. This is a useful feature in time-sensitive settings such as high-frequency trading. Furthermore, certain emojis and emoji pairs exhibit strong predictive power for market sentiment, demonstrating over 90 percent accuracy in predicting bullish or bearish trends. Finally, our research reveals large statistical differences in emoji usage between financial and general social media contexts, stressing the need for domain-specific sentiment analysis models.
Abstract (translated)
URL
https://arxiv.org/abs/2605.09469
PDF
-
A Comparative Analysis of Classical Machine Learning and Deep Learning Approaches for Sentiment Classification on IMDb Movie Reviews
2026-05-08 14:44:36Erma Daniar Safitri, Lia Hana Ichisasmita, Citra Agustin, Luluk Muthoharoh, Ardika Satria, Martin Clinton Tosima ManullangAbstract
This paper presents a comparative study of classical machine learning and deep learning methods for sentiment classification on the IMDb movie reviews dataset. The machine learning pipeline uses TF-IDF features and PyCaret AutoML to evaluate Logistic Regression, Naïve Bayes, and Support Vector Machine, while the deep learning pipeline implements BiLSTM and BiLSTM with an attention mechanism. Experimental results show that classical machine learning, especially SVM, achieves the best performance with an accuracy of 0.8530, outperforming the deep learning models in this study. The BiLSTM with Attention model improves over the standard BiLSTM and reaches an accuracy of 0.706, indicating better contextual modeling. The paper concludes that although deep learning can capture sequential dependencies, classical machine learning remains a strong baseline when combined with effective feature engineering such as TF-IDF, particularly under limited data and computational resources.
Abstract (translated)
URL
https://arxiv.org/abs/2605.07811
PDF
-
A Comparative Analysis of Machine Learning and Deep Learning Models for Tweet Sentiment Classification: A Case Study on the Sentiment140 Dataset
2026-05-06 13:21:07Vita Anggraini, Cintya Bella, Bastian, Luluk Muthoharoh, Ardika Satria, Martin C. T. ManullangAbstract
The exponential growth of social media has created an urgent need for automated systems to analyze unstructured public sentiment in real time. This study compares a traditional Logistic Regression model using TF-IDF features with a deep learning Bidirectional Long Short-Term Memory (BiLSTM) architecture on a 10,000-tweet subset of the Sentiment140 dataset. Experimental results show that Logistic Regression outperformed BiLSTM, achieving an accuracy of 73.5% compared with 69.17%, while the deep learning model exhibited mild overfitting. These findings suggest that for medium-scale informal text data, classical machine learning with robust feature extraction can outperform more complex deep learning approaches. Finally, the trained models were integrated into an interactive web application using Streamlit and deployed on Hugging Face Spaces for public access.
Abstract (translated)
URL
https://arxiv.org/abs/2605.04888
PDF
-
Sentiment Analysis of Indonesian Spotify Reviews Using Machine Learning and BiLSTM
2026-05-05 07:28:42Uliano Wilyam Purba, Andre Hadiman Rotua Parhusip, Sahid Maulana, Luluk Muthoharoh, Ardika Satria, Martin C. T. ManullangAbstract
This paper benchmarks classical machine learning and deep learning approaches for three-class sentiment classification of Indonesian Spotify reviews. Using 100,000 scraped reviews and 70,155 cleaned samples, the study compares Support Vector Machine, Multinomial Naive Bayes, and Decision Tree models with a two-layer BiLSTM. Both approaches use the same preprocessing pipeline, including slang normalization, stopword removal, and stemming. Decision Tree achieves the best performance among the classical models, while BiLSTM attains the highest weighted F1-score overall but fails on the minority neutral class. The paper concludes that BiLSTM is stronger for overall sentiment detection, whereas machine learning with SMOTE provides more balanced three-class performance.
Abstract (translated)
URL
https://arxiv.org/abs/2605.03443
PDF
-
Benchmarking Logistic Regression, SVM, Naive Bayes, and IndoBERT Fine-Tuning for Sentiment Analysis on Indonesian Product Reviews
2026-05-05 07:20:42Nabila Zakiyah Zahra, Salwa Farhanatussaidah, Nasywa Nur Afifah, Luluk Muthoharoh, Ardika Satria, Martin C. T. ManullangAbstract
The exponential growth of e-commerce platforms in Indonesia has generated a massive volume of user-generated product reviews. Analyzing the sentiment of these reviews is critical for measuring customer satisfaction and identifying product issues at scale. This paper benchmarks traditional Machine Learning (ML) approaches against a Transformer-based Deep Learning model for a three-class sentiment analysis task (positive, neutral, negative) on the Tokopedia Product Reviews 2025 dataset. We implemented Term Frequency-Inverse Document Frequency (TF-IDF) feature extraction coupled with three algorithms: Logistic Regression, Linear Support Vector Machine (SVM), and Multinomial Naive Bayes as robust baselines. Subsequently, we fine-tuned the IndoBERT model (indobenchmark/indobert-base-p1) for contextual sequence classification. To computationally address the severe class imbalance inherent in e-commerce feedback, we applied balanced class weights for the baseline models and engineered a custom weighted cross-entropy loss function within the IndoBERT training loop, following the broader motivation of imbalanced-learning research. Our comprehensive evaluation using Accuracy, Macro F1-score, and Weighted F1-score revealed that the traditional Linear SVC model significantly outperformed the IndoBERT model in our experimental setup, achieving an Accuracy of 97.60% and a Macro F1-score of 0.5510, compared to IndoBERT's 88.70% and 0.5088. Detailed analysis indicates that this performance gap was primarily driven by discrepancies in the data sampling regimes, where baselines utilized the full corpus while the Transformer was constrained to a sampled subset. Finally, we demonstrate the practical viability of our pipeline by deploying the final sentiment classification model as an interactive Gradio web application.
Abstract (translated)
URL
https://arxiv.org/abs/2605.03439
PDF
-
Enhancing Game Review Sentiment Classification on Steam Platform with Attention-Based BiLSTM
2026-05-02 08:17:06Abit Ahmad Oktarian, Fadhil Fitra Wijaya, Dhafin Razaqa Luthfi, Luluk Muthoharoh, Ardika Satria, Martin Clinton Tosima ManullangAbstract
This paper investigates sentiment classification of Steam game reviews using an attention-based Bidirectional Long Short-Term Memory (BiLSTM) model. Using a dataset of 50,000 reviews sampled from a larger Steam review corpus, the authors compare a traditional machine learning baseline based on TF-IDF and PyCaret AutoML with a deep learning approach implemented in PyTorch. The proposed BiLSTM+Attention model is trained with class-weighted cross-entropy to address class imbalance and achieves 83% accuracy and 85% weighted F1-score on the test set, with 90% recall for negative reviews. The paper also presents attention visualizations to show interpretability by highlighting sentiment-bearing words. The study concludes that the BiLSTM+Attention model is effective for analyzing user sentiment in Steam reviews and useful for helping developers understand player feedback.
Abstract (translated)
URL
https://arxiv.org/abs/2605.01315
PDF
-
Sentiment Analysis of AI Adoption in Indonesian Higher Education Using Machine Learning and Transformer-Based Models
2026-04-30 05:25:05Happy Syahrul Ramadhan, Ahmad Sahidin Akbar, Karin Yehezkiel Sinaga, Luluk Muthoharoh, Ardika Satria, Martin C. T. ManullangAbstract
This study analyzes Indonesian student opinions on the adoption of artificial intelligence in higher education using two approaches: TF-IDF-based machine learning and Transformer-based deep learning. The dataset consists of 2,295 labeled samples, combining 1,154 student opinions with additional lexical sentiment data. LightGBM, Random Forest, and Support Vector Machine (SVM) are evaluated as machine learning models, while DistilBERT is fine-tuned for binary sentiment classification. The results show that SVM achieves the best performance among the machine learning models with 82.14% test accuracy and F1-score, while DistilBERT performs best overall with 84.78% accuracy and 84.75% F1-score. These findings indicate that Transformer-based models better capture contextual information, although SVM remains a competitive and efficient alternative for sentiment classification.
Abstract (translated)
URL
https://arxiv.org/abs/2604.27439
PDF
-
Benchmarking Logistic Regression, SVM, and LightGBM Against BiLSTM with Attention for Sentiment Analysis on Indonesian Product Reviews
2026-04-28 10:00:42Razin Hafid Hamdi, Ivana Margareth Hutabarat, Hanna Gresia Sinaga, Luluk Muthoharoh, Ardika Satria, Martin C. T. ManullangAbstract
Sentiment analysis of product reviews on e-commerce platforms plays a critical role in automatically understanding customer satisfaction and providing actionable insights for sellers seeking to improve product quality. This paper presents a comprehensive benchmarking study comparing a Machine Learning (ML) approach via the PyCaret AutoML framework against a Deep Learning (DL) approach based on a Bidirectional Long Short-Term Memory (BiLSTM) architecture with an Attention mechanism for binary sentiment classification on Indonesian product reviews. The dataset comprises 19,728 samples balanced equally between positive and negative reviews. For the ML approach, three prominent algorithms were evaluated via 10-fold stratified cross-validation: Logistic Regression (LR), Support Vector Machine (SVM) with a linear kernel, and Light Gradient Boosting Machine (LightGBM). Logistic Regression achieved the best ML performance with an accuracy of 97.26\% and an F1-score of 97.26\%. The BiLSTM with Attention model, evaluated on 3,946 held-out test samples, achieved an accuracy of 97.24\% and an F1-score of 97.24\%. These comparative results demonstrate that traditional ML algorithms with proper preprocessing and feature extraction can compete closely with, and even marginally outperform, more complex sequential DL architectures on high-dimensional datasets, while simultaneously offering greater computational efficiency.
Abstract (translated)
URL
https://arxiv.org/abs/2604.25452
PDF
-
Does Welsh media need a review? Detecting bias in Nation.Cymru's political reporting
2026-04-19 21:52:59Cai Parry-JonesAbstract
Wales' political landscape has been marked by growing accusations of bias in Welsh media. This paper takes the first computational step toward testing those claims by examining this http URL, a prominent Welsh political news outlet. I use a two-stage natural language processing (NLP) pipeline: (1) a robustly optimized BERT approach (RoBERTa) bias detector for efficient bias discovery and (2) a large language model (LLM) for target-attributed sentiment classification of bias labels from (1). A primary analysis of 15,583 party mentions across 2022-2026 news articles finds that Reform UK attracts biased framing at twice the rate of Plaid Cymru and over three times as negative in mean sentiment (p<0.001). A secondary analysis across four parties across both news and opinion articles shows that Plaid Cymru is the outlier, receiving markedly more favourable framing than any other party. These findings provide evidence of measurable differential framing in a single Welsh political media outlet, supporting calls for a broader review of Welsh media coverage. Furthermore, the two-stage pipeline offers a low-cost, replicable framework for extending this analysis to other Welsh outlets, as well as media ecosystems outside of Wales.
Abstract (translated)
URL
https://arxiv.org/abs/2604.17628
PDF
-
Compressing Sequences in the Latent Embedding Space: $K$-Token Merging for Large Language Models
2026-04-16 15:32:45Zihao Xu, John Harvill, Ziwei Fan, Yizhou Sun, Hao Ding, Hao WangAbstract
Large Language Models (LLMs) incur significant computational and memory costs when processing long prompts, as full self-attention scales quadratically with input length. Token compression aims to address this challenge by reducing the number of tokens representing inputs. However, existing prompt-compression approaches primarily operate in token space and overlook inefficiencies in the latent embedding space. In this paper, we propose K-Token Merging, a latent-space compression framework that merges each contiguous block of K token embeddings into a single embedding via a lightweight encoder. The compressed sequence is processed by a LoRA-adapted LLM, while generation remains in the original vocabulary. Experiments on structural reasoning (Textualized Tree), sentiment classification (Amazon Reviews), and code editing (CommitPackFT) show that K-Token Merging lies on the Pareto frontier of performance vs. compression, achieving up to 75% input length reduction with minimal performance degradation.
Abstract (translated)
URL
https://arxiv.org/abs/2604.15153
PDF
-
Applied Explainability for Large Language Models: A Comparative Study
2026-04-15 13:07:29Venkata Abhinandan KancharlaAbstract
Large language models (LLMs) achieve strong performance across many natural language processing tasks, yet their decision processes remain difficult to interpret. This lack of transparency creates challenges for trust, debugging, and deployment in real-world systems. This paper presents an applied comparative study of three explainability techniques: Integrated Gradients, Attention Rollout, and SHAP, on a fine-tuned DistilBERT model for SST-2 sentiment classification. Rather than proposing new methods, the focus is on evaluating the practical behavior of existing approaches under a consistent and reproducible setup. The results show that gradient-based attribution provides more stable and intuitive explanations, while attention-based methods are computationally efficient but less aligned with prediction-relevant features. Model-agnostic approaches offer flexibility but introduce higher computational cost and variability. This work highlights key trade-offs between explainability methods and emphasizes their role as diagnostic tools rather than definitive explanations. The findings provide practical insights for researchers and engineers working with transformer-based NLP systems. This is a preprint and has not undergone peer review.
Abstract (translated)
URL
https://arxiv.org/abs/2604.15371
PDF
-
From Prediction to Justification: Aligning Sentiment Reasoning with Human Rationale via Reinforcement Learning
2026-04-15 01:55:40Shihao Zhang, Ziwei Wang, Jie Zhou, Yulan Wu, Qin Chen, Zhikai Lei, Liyang Yu, Liang Dou, Liang HeAbstract
While Aspect-based Sentiment Analysis (ABSA) systems have achieved high accuracy in identifying sentiment polarities, they often operate as "black boxes," lacking the explicit reasoning capabilities characteristic of human affective cognition. Humans do not merely categorize sentiment; they construct causal explanations for their judgments. To bridge this gap, we propose ABSA-R1, a large language model framework designed to mimic this ``reason-before-predict" cognitive process. By leveraging reinforcement learning (RL), ABSA-R1 learns to articulate the why behind the what, generating natural language justifications that ground its sentiment predictions. We introduce a Cognition-Aligned Reward Model (formerly sentiment-aware reward model) that enforces consistency between the generated reasoning path and the final emotional label. Furthermore, inspired by metacognitive monitoring, we implement a performance-driven rejection sampling strategy that selectively targets hard cases where the model's internal reasoning is uncertain or inconsistent. Experimental results on four benchmarks demonstrate that equipping models with this explicit reasoning capability not only enhances interpretability but also yields superior performance in sentiment classification and triplet extraction compared to non-reasoning baselines.
Abstract (translated)
URL
https://arxiv.org/abs/2604.13398
PDF
-
MEME-Fusion@CHiPSAL 2026: Multimodal Ablation Study of Hate Detection and Sentiment Analysis on Nepali Memes
2026-04-13 07:37:14Samir Wagle, Reewaj Khanal, Abiral AdhikariAbstract
Hate speech detection in Devanagari-scripted social media memes presents compounded challenges: multimodal content structure, script-specific linguistic complexity, and extreme data scarcity in low-resource settings. This paper presents our system for the CHiPSAL 2026 shared task, addressing both Subtask A (binary hate speech detection) and Subtask B (three-class sentiment classification: positive, neutral, negative). We propose a hybrid cross-modal attention fusion architecture that combines CLIP (ViT-B/32) for visual encoding with BGE-M3 for multilingual text representation, connected through 4-head self-attention and a learnable gating network that dynamically weights modality contributions on a per-sample basis. Systematic evaluation across eight model configurations demonstrates that explicit cross-modal reasoning achieves a 5.9% F1-macro improvement over text-only baselines on Subtask A, while uncovering two unexpected but critical findings: English-centric vision models exhibit near-random performance on Devanagari script, and standard ensemble methods catastrophically degrade under data scarcity (N nearly equal to 850 per fold) due to correlated overfitting. The code can be accessed at this https URL
Abstract (translated)
URL
https://arxiv.org/abs/2604.14218
PDF
