Nitisha Aggarwal
Author directory2026
Team CPS Lab at SHROOM-Visions: Parameter-Efficient Vision-Language Hallucination Detection with Threshold Tuning
Anika Srivastava | Vishal Kumar | Harshita Sharma | Nitisha Aggarwal | Geetika Jain Saxena | Amit Pundir | Sanjeev Singh
The Proceedings of the Third Workshop on Uncertainty-Aware NLP
Anika Srivastava | Vishal Kumar | Harshita Sharma | Nitisha Aggarwal | Geetika Jain Saxena | Amit Pundir | Sanjeev Singh
The Proceedings of the Third Workshop on Uncertainty-Aware NLP
Fine-grained large-vision-language hallucination detection presents the challenge of locating hallucinated content in image-grounded responses and classifying them with corresponding hallucination categories. Our method for the SHROOM-Visions 2026 Shared Task explores a parameter-efficient solution by fine-tuning a lightweight token-level classification head on the frozen weights of the Qwen3-VL-8B-Instruct vision-language model. We utilized category-specific thresholds calculated based on validation character-level IoU and designed a span merging post-processing step to facilitate span localization with fewer fragmented predictions. We also establish a possible connection between the annotator agreement and the operating thresholds. Our analysis suggests that, within our experimental setup, hallucination categories that had higher inter-annotator agreement tended to have higher optimal decision thresholds whereas categories with lower agreement tended to have lower optimal decision thresholds. Our parameter-efficient method achieved ranks 18th, 16th, 16th, and 15th place on the English, French, Italian, and Chinese tracks, respectively, showing promising performance for multilingual fine-grained hallucination detection.
CPS Lab 1@MMLow 2026: Hierarchical Multimodal Classification of Tamil Political Memes via Logit-Conditioned Stance-Target Prediction
Vishal Kumar | Anika Srivastava | Harshita Sharma | Nitisha Aggarwal | Geetika Jain Saxena | Amit Pundir | Sanjeev Singh
Proceedings of the Fifth Workshop on Multimodal Machine Learning in Low-resource Languages
Vishal Kumar | Anika Srivastava | Harshita Sharma | Nitisha Aggarwal | Geetika Jain Saxena | Amit Pundir | Sanjeev Singh
Proceedings of the Fifth Workshop on Multimodal Machine Learning in Low-resource Languages
Political memes have become an effective tool for online political communication and engagement. Their interpretation is particularly challenging in low-resource and code-mixed languages such as Tamil, where memes often include Tamil-English, use culturally relevant lexicon, incorporate political entities, and utilize references to prominent personalities and parties. This paper presents our submission to the MMLow shared task at AACL-IJCNLP 2026 on multi-level classification of Tamil political memes, which requires jointly predicting (i) a coarse-grained stance label (Support vs. Troll) and (ii) a fine-grained target label identifying whether the meme concerns a political party or a person. We proposed a lightweight hierarchical classifier built on frozen, pretrained multimodal encoders: images are represented using CLIP (ViT-B/32) and text using the multilingual BGE-M3 embedding model, with the two modalities combined via early, non-parametric concatenation fusion. A shared encoder feeds two prediction heads, where the target-identification head is explicitly conditioned on the stance head’s output logits via a gradient-detachment mechanism, allowing the model to respect the task’s label hierarchy without corrupting the coarse-level classifier’s gradients. To address pronounced class imbalance in both label levels, we applied balanced loss weighting at each level. On the test set of unlabeled data, our system achieved 3rd place, with macro-F1 scores of 0.856 for stance classification and 0.532 for target identification, yielding a final score of 0.694, indicating promising performance on the Tamil political meme classification problem.
CPSLab2@MMLow 2026: A Hierarchical Multimodal Framework with Stratified K-Fold Training for Political Meme Classification
Anika Srivastava | Vishal Kumar | Harshita Sharma | Nitisha Aggarwal | Geetika Jain Saxena | Amit Pundir | Sanjeev Singh
Proceedings of the Fifth Workshop on Multimodal Machine Learning in Low-resource Languages
Anika Srivastava | Vishal Kumar | Harshita Sharma | Nitisha Aggarwal | Geetika Jain Saxena | Amit Pundir | Sanjeev Singh
Proceedings of the Fifth Workshop on Multimodal Machine Learning in Low-resource Languages
Political memes communicate opinions using the interplay between visual and textual information, posing a challenging multimodal classification task. We present a hierarchical multimodal framework for political meme classification, which decomposes the task into stance and target entity prediction. Our framework was evaluated on the data from the Multi-Level Political Meme Classification Shared Task at MMLow @ AACL-IJCNLP 2026. The task comprised two levels of classification: (i) predicting the political stance of a meme as either Support/Praise or Troll/Oppose, and (ii) identifying the target of the meme combined with the stance. We approached the task by first predicting the Level-1 labels followed by predicting the target entity (Person/Party) and then combining them in a deterministic manner to predict the final Level-2 labels. Our method used a SigLIP vision transformer for stance classification and a hybrid SigLIP-XLM-RoBERTa architecture for target classification with late fusion. We trained our models using 5-fold stratified cross-validation with checkpoint ensembling. We also performed ablation experiments to analyze the contribution of individual modalities, observing that visual features are more informative for stance prediction while textual features contain more information for target identification, which supports the multimodal task at hand. Our system achieved a macro F1 score of 0.9027 and 0.6694 for Level-1 and Level-2 classification, respectively, achieving a final score of 0.7860, ranking first in the shared task.
2025
teamiic@DravidianLangTech2025-NAACL 2025: Transformer-Based Multimodal Feature Fusion for Misogynistic Meme Detection in Low-Resource Dravidian Language
Harshita Sharma | Simran | Vajratiya Vajrobol | Nitisha Aggarwal
Proceedings of the Fifth Workshop on Speech, Vision, and Language Technologies for Dravidian Languages
Harshita Sharma | Simran | Vajratiya Vajrobol | Nitisha Aggarwal
Proceedings of the Fifth Workshop on Speech, Vision, and Language Technologies for Dravidian Languages
Misogyny has become a pervasive issue in digital spaces. Misleading gender stereotypes are getting communicated through digital content.This content is majorly displayed as a text-and-image memes. With the growing prevalence of online content, it is essential to develop automated systems capable of detecting such harmful content to ensure safer online environments. This study focuses on the detection of misogynistic memes in two Dravidian languages, Tamil and Malayalam. The proposed model utilizes a pre-trained XLM-RoBERTa (XLM-R) model for text analysis and a Vision Transformer (ViT) for image feature extraction. A custom neural network classifier was trained on integrating the outputs of both modalities to form a unified representation. This model predicts whether the meme represents misogyny or not. This follows an early-fusion strategy since features of both modalities are combined before feeding into the classification model. This approach achieved promising results using a macro F1-score of 0.84066 on the Malayalam test dataset and 0.68830 on the Tamil test dataset. In addition, it is worth noting that this approach secured Rank 7 and 11 in Malayalam and Tamil classification respectively in the shared task of Misogyny Meme Detection (MMD). The findings demonstrate that the multimodal approach significantly enhances the accuracy of detecting misogynistic content compared to text-only or image-only models.
2023
iicteam@LT-EDI-2023: Leveraging pre-trained Transformers for Fine-Grained Depression Level Detection in Social Media
Vajratiya Vajrobol | Nitisha Aggarwal | Karanpreet Singh
Proceedings of the Third Workshop on Language Technology for Equality, Diversity and Inclusion
Vajratiya Vajrobol | Nitisha Aggarwal | Karanpreet Singh
Proceedings of the Third Workshop on Language Technology for Equality, Diversity and Inclusion
Depression is a prevalent mental illness characterized by feelings of sadness and a lack of interest in daily activities. Early detection of depression is crucial to prevent severe consequences, making it essential to observe and treat the condition at its onset. At ACL-2022, the DepSign-LT-EDI project aimed to identify signs of depression in individuals based on their social media posts, where people often share their emotions and feelings. Using social media postings in English, the system categorized depression signs into three labels: “not depressed,” “moderately depressed,” and “severely depressed.” To achieve this, our team has applied MentalRoBERTa, a model trained on big data of mental health. The test results indicated a macro F1-score of 0.439, ranking the fourth in the shared task.
IIC_Team@Multimodal Hate Speech Event Detection 2023: Detection of Hate Speech and Targets using Xlm-Roberta-base
Karanpreet Singh | Vajratiya Vajrobol | Nitisha Aggarwal
Proceedings of the 6th Workshop on Challenges and Applications of Automated Extraction of Socio-political Events from Text
Karanpreet Singh | Vajratiya Vajrobol | Nitisha Aggarwal
Proceedings of the 6th Workshop on Challenges and Applications of Automated Extraction of Socio-political Events from Text
Hate speech has emerged as a pressing issue on social media platforms, fueled by the increasing availability of multimodal data and easy internet access. Addressing this problem requires collaborative efforts from researchers, policymakers, and online platforms. In this study, we investigate the detection of hate speech in multimodal data, comprising text-embedded images, by employing advanced deep learning models. The main objective is to identify effective strategies for hate speech detection and content moderation. We conducted experiments using four state-of-the-art classifiers: XLM-Roberta-base, BiLSTM, XLNet base cased, and ALBERT, on the CrisisHateMM[4] dataset, consisting of over 4700 text-embedded images related to the Russia-Ukraine conflict. The best findings reveal that XLM-Roberta-base exhibits superior performance, outperforming other classifiers across all evaluation metrics, including an impressive F1 score of 84.62 for sub-task 1 and 69.73 for sub-task 2. The future scope of this study lies in exploring multimodal approaches to enhance hate speech detection accuracy, integrating ethical considerations to address potential biases, promoting fairness, and safeguarding user rights. Additionally, leveraging larger and more diverse datasets will contribute to developing more robust and generalised hate speech detection solutions.