F2IND-IT! -- Multimodal Fuzzy Fake Indian News Detection using Images and Text

Jeevaraj S.; Khushi Singh; Kushal Trivedi; Murtuza Shaikh

arxiv: 2605.17115 · v1 · pith:DAHNCAI7new · submitted 2026-05-16 · 💻 cs.AI

F2IND-IT! -- Multimodal Fuzzy Fake Indian News Detection using Images and Text

Kushal Trivedi , Murtuza Shaikh , Khushi Singh , Jeevaraj S. This is my paper

Pith reviewed 2026-05-20 14:52 UTC · model grok-4.3

classification 💻 cs.AI

keywords fake news detectionmultimodal learningANFISattention fusionIndian mediaimage and text analysis

0 comments

The pith

A multimodal system fuses image features, text embeddings, fuzzy inference and attention to detect fake Indian news more accurately.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents a framework that processes news images and text together to identify misinformation in Indian media. Visual details come from a convolutional network while text meaning is captured by a distilled transformer model; these are then fed into a fuzzy inference system that outputs a reliability score and an attention module that learns how much to trust each source before final classification. If the approach holds, it would give automated tools a practical edge in regions where media bias and regional variation make single-modality checks unreliable. The work evaluates the full pipeline on the IFND collection and reports gains over earlier single-stream or simpler fusion methods. The central goal is to show that this specific combination of components handles the complexity of Indian fake news better than prior techniques.

Core claim

The proposed multimodal architecture that combines ResNet-50 for visual feature extraction, DistilBERT for textual semantic embeddings, an Adaptive Neuro-Fuzzy Inference System to produce a fuzzy reliability score, and a lightweight attention-based fusion module achieves superior accuracy, precision, recall, and F1 performance on the IFND dataset compared with previous research.

What carries the argument

The multimodal fusion pipeline that extracts visual and textual representations, routes them through neuro-fuzzy inference for a reliability score, and applies learned attention weights to combine the modalities for classification.

If this is right

The combined visual-textual-fuzzy-attention system records higher accuracy, precision, recall, and F1 scores than earlier methods on the IFND dataset.
The architecture validates the benefit of using both image and text modalities together with neuro-fuzzy scoring for this detection task.
Comparative experiments on the same benchmark confirm the pipeline outperforms prior single-modality or non-fuzzy approaches.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same fusion strategy could be retrained on news from other countries or languages to test whether the visual-text-fuzzy combination transfers beyond the Indian context.
Deploying the model in a live social-media filter would allow early flagging of suspect posts before they reach large audiences.
The fuzzy reliability score could be surfaced to human moderators as a numeric uncertainty measure rather than a hard true-or-false label.

Load-bearing premise

The IFND dataset is treated as an unbiased and representative sample of real-world Indian fake news, and the reported performance numbers are assumed to reflect generalization to new examples rather than overfitting to the training split.

What would settle it

Re-training the same architecture on a new, independently collected set of Indian news images and text and finding that its accuracy, precision, recall, and F1 scores fall to the level of simpler text-only or image-only baselines would falsify the central claim.

Figures

Figures reproduced from arXiv: 2605.17115 by Jeevaraj S., Khushi Singh, Kushal Trivedi, Murtuza Shaikh.

read the original abstract

Biased manipulation of facts across regional and national media outlets complicates misinformation detection in diverse landscapes like India. This paper introduces a novel multimodal framework combining visual and textual modalities for enhanced fake news detection on Indian media. The architecture utilizes a ResNet-50 Convolutional Neural Network to extract visual features from news images, a DistilBERT encoder to obtain textual semantic embeddings, and an Adaptive Neuro-Fuzzy Inference System (ANFIS) to generate a fuzzy reliability score. A lightweight attention-based fusion module assigns learnable weights to each modality prior to classification. Evaluated on the IFND dataset, the proposed model is validated through an in-depth comparative analysis against previous research. Experimental results demonstrate superior performance across accuracy, precision, recall, and $F_1$-scores, confirming the efficacy of the architecture.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This paper wires together ResNet-50, DistilBERT, ANFIS and attention fusion for fake news on the IFND dataset, but the superiority claims rest on thin validation evidence.

read the letter

This paper wires together four established components—ResNet-50 for image features, DistilBERT for text embeddings, ANFIS for a fuzzy reliability score, and a lightweight attention fusion step—then tests the pipeline on the IFND dataset. The central claim is that the combined system outperforms prior work on accuracy, precision, recall, and F1. Nothing in the architecture is mathematically new; the contribution is the specific assembly aimed at Indian media with its linguistic diversity and regional biases. That focus is reasonable for an applied setting where English-only models often fall short. The fuzzy inference layer adds a modest interpretability angle by producing a reliability score rather than a hard binary output, which could be useful in practice. The attention module lets the model learn modality weights, a standard but sensible choice here. The evaluation is the clear weak spot. The reported gains come from what appears to be a single train-test split on one dataset, with no error bars, no repeated runs across seeds, no k-fold or cross-validation protocol, and no statistical significance tests against the baselines. Multimodal models with learnable fusion have enough capacity to exploit collection artifacts or class imbalances in the IFND data, so the performance edge is hard to trust without those checks. The assumption that IFND is a representative benchmark for real-world Indian fake news also goes unexamined in the provided details. This work is aimed at applied researchers building misinformation tools for non-English or culturally specific media environments. A reader already working on multimodal pipelines might borrow the fusion pattern or the ANFIS integration, but the paper will not shift broader methods or benchmarks. I would send it for peer review so the authors can supply proper generalization tests and ablations; the core idea is coherent enough to warrant referee feedback even if the current evidence is limited.

Referee Report

2 major / 2 minor

Summary. The manuscript introduces F2IND-IT, a multimodal architecture for detecting fake news in Indian media. It extracts visual features with ResNet-50, textual embeddings with DistilBERT, computes a fuzzy reliability score via ANFIS, and fuses modalities with a learnable attention module. The central claim is that this model achieves superior accuracy, precision, recall, and F1-score on the IFND dataset relative to prior methods.

Significance. If the performance gains prove robust under proper statistical validation and generalization checks, the hybrid deep-learning-plus-fuzzy approach could provide a practical contribution to regional misinformation detection. The explicit combination of ANFIS with attention fusion is a distinctive design choice that merits further exploration if the empirical support is strengthened.

major comments (2)

[Abstract / Experimental Results] Abstract and Experimental Results section: the superiority claim is asserted without any reported train/test split ratios, k-fold or repeated-run protocol, variance across random seeds, error bars, or statistical significance tests against baselines. These omissions make it impossible to determine whether the reported gains reflect genuine improvement or dataset-specific fitting on IFND.
[Results] Results section (tables/figures reporting metrics): no ablation studies isolating the contribution of the ANFIS component or the attention fusion module are described, nor are baseline details (architectures, hyperparameters, or training procedures) provided. Without these, the claim that the full multimodal pipeline is responsible for the gains cannot be evaluated.

minor comments (2)

[Abstract] The abstract refers to an 'in-depth comparative analysis' but does not list the specific prior methods or their reported scores in the provided text; adding an explicit comparison table would improve clarity.
[Method] Notation for the attention weights and ANFIS membership parameters is introduced but not accompanied by the exact equations or initialization details in the visible sections.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We appreciate the referee's thorough review and constructive suggestions for improving the experimental section of our manuscript. We address each major comment in detail below and have updated the manuscript to incorporate the recommended enhancements for better reproducibility and analysis.

read point-by-point responses

Referee: [Abstract / Experimental Results] Abstract and Experimental Results section: the superiority claim is asserted without any reported train/test split ratios, k-fold or repeated-run protocol, variance across random seeds, error bars, or statistical significance tests against baselines. These omissions make it impossible to determine whether the reported gains reflect genuine improvement or dataset-specific fitting on IFND.

Authors: We agree with the referee that the experimental protocol details are essential for validating the performance claims. The current manuscript does not include these specifics, which is an oversight. In the revised manuscript, we will add a comprehensive description of the data splitting strategy, the use of repeated runs with different random seeds, error bars representing standard deviation, and results of statistical tests to confirm that the improvements are significant and not due to overfitting on the IFND dataset. revision: yes
Referee: [Results] Results section (tables/figures reporting metrics): no ablation studies isolating the contribution of the ANFIS component or the attention fusion module are described, nor are baseline details (architectures, hyperparameters, or training procedures) provided. Without these, the claim that the full multimodal pipeline is responsible for the gains cannot be evaluated.

Authors: We concur that ablation studies and detailed baseline information are crucial to attribute the performance gains to the proposed components. The original submission lacks these analyses. We will revise the Results section to include ablation studies that isolate the effects of the ANFIS component and the attention fusion module. Additionally, we will provide complete specifications for all compared baseline methods, including their architectures, hyperparameters, and training procedures, to allow for proper evaluation of the multimodal pipeline's contributions. revision: yes

Circularity Check

0 steps flagged

No circularity detected in derivation or claims

full rationale

The paper describes a multimodal architecture (ResNet-50 + DistilBERT + ANFIS + attention fusion) and reports empirical performance metrics on the IFND dataset with comparisons to prior work. No mathematical derivation chain, equations, or self-definitional reductions are present in the provided text. Performance results are standard empirical evaluations on an external benchmark rather than quantities forced by construction from fitted parameters or self-citations. The central efficacy claim rests on experimental outcomes and comparative analysis, which are independent of the inputs in the sense required by the circularity criteria. No load-bearing self-citations, uniqueness theorems, or ansatzes are invoked.

Axiom & Free-Parameter Ledger

2 free parameters · 1 axioms · 0 invented entities

Only the abstract is available, so the ledger is necessarily incomplete and based on stated components.

free parameters (2)

Attention fusion weights
Learnable parameters that assign modality importance during training on IFND.
ANFIS membership function parameters
Fuzzy inference parameters tuned to produce the reliability score.

axioms (1)

domain assumption The IFND dataset constitutes a valid and representative testbed for Indian fake news detection.
All performance claims rest on evaluation against this single collection.

pith-pipeline@v0.9.0 · 5673 in / 1243 out tokens · 50463 ms · 2026-05-20T14:52:55.102305+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

ANFIS layer with 2 Gaussian membership functions... 16 fuzzy rules... weighted sum of normalized firing strengths
IndisputableMonolith/Foundation/BranchSelection.lean branch_selection unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

lightweight attention-based fusion module assigns learnable weights to each modality

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

20 extracted references · 20 canonical work pages

[1]

In: 2024 28th International Computer Science and Engineering Conference (ICSEC), pp

El-Amrany, S., Brust, M.R., Pecero, J.E., Bouvry, P.: Tri-FusionDet: Leveraging User Engagement, Textual, and Visual Features for Enhanced Fake News Detec- tion. In: 2024 28th International Computer Science and Engineering Conference (ICSEC), pp. 1–6 (2024).https://doi.org/10.1109/ICSEC62781.2024.10770746

work page doi:10.1109/icsec62781.2024.10770746 2024
[2]

indiatimes

BrandEquity Bureau, 91 Per Cent Believe Fake News Can Influence Voting Deci- sions:Report,Accessed:2025-07-17.(2024).https://brandequity.economictimes. indiatimes . com / news / research / 91 - per - cent - believe - fake - news - can - influence-voting-decisions-report/110004307?utm_source=chatgpt.com

work page arXiv 2025
[3]

(2024).https://www.communicationstoday.co.in/ by- 2025- 56- percent- new- indian- internet- users- from- rural- areas/?utm_ source=chatgpt.com

Communications Today, By 2025, 56 Percent New Indian Internet Users from Rural Areas, Accessed: 2025-07-20. (2024).https://www.communicationstoday.co.in/ by- 2025- 56- percent- new- indian- internet- users- from- rural- areas/?utm_ source=chatgpt.com

work page 2025
[4]

Accessed: 2025-07-20

Exchange4Media, Misinformation Queries Raised, 1,575 Fake News Cases Flagged Since 2022: PIB, (2025).https://www.exchange4media.com/media-others-news/ 68914 - misinformation - queries - raised - 1575 - fake - news - cases - flagged - since-2022-pib-142128.html. Accessed: 2025-07-20

work page 2022
[5]

Applied Soft Computing179, 113277 (2025)

Gedara, T.M.H., Loia, V., Tomasiello, S.: A fuzzy-based multimodal approach for interpretable fake news detection. Applied Soft Computing179, 113277 (2025). https://doi.org/10.1016/j.asoc.2025.113277.https://www.sciencedirect. com/science/article/pii/S1568494625005885

work page doi:10.1016/j.asoc.2025.113277.https://www.sciencedirect 2025
[6]

https://indianexpress.com/article/india/214- rise- in- cases- relating- to-fake-news-rumours-7511534/

Indian Express, 214% Rise in Cases Relating to Fake News, Rumours, (2021). https://indianexpress.com/article/india/214- rise- in- cases- relating- to-fake-news-rumours-7511534/. Accessed: 2025-07-20

work page 2021
[7]

Expert Systems with Applications95, 172–189 (2018).https : / / doi

Iyer,A.R.,Prasad,D.K.,Quek,C.H.:PIE-RSPOP:Abrain-inspiredpseudo-incremental ensemble rough set pseudo-outer product fuzzy neural network. Expert Systems with Applications95, 172–189 (2018).https : / / doi . org / 10 . 1016 / j . eswa . 2017 . 11 . 027.https : / / www . sciencedirect . com / science / article / pii / S0957417417307832 F2IND-IT! - Multimoda...

work page 2018
[8]

arXiv:2411.06097 [cs.AI].https://arxiv.org/abs/ 2411.06097

Jun-hao, Xu, A Multimodal Adaptive Graph-based Intelligent Classification Model for Fake News, (2024). arXiv:2411.06097 [cs.AI].https://arxiv.org/abs/ 2411.06097

work page arXiv 2024
[9]

IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics)31(6), 902–918 (2001).https://doi.org/10.1109/3477

Kasabov, N.: Evolving fuzzy neural networks for supervised/unsupervised online knowledge-based learning. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics)31(6), 902–918 (2001).https://doi.org/10.1109/3477. 969494

work page doi:10.1109/3477 2001
[10]

IEEE Transactions on Fuzzy Systems 10(2), 144–154 (2002).https://doi.org/10.1109/91.995117

Kasabov, N., Song, Q.: DENFIS: dynamic evolving neural-fuzzy inference system and its application for time-series prediction. IEEE Transactions on Fuzzy Systems 10(2), 144–154 (2002).https://doi.org/10.1109/91.995117

work page doi:10.1109/91.995117 2002
[11]

https://www.ndtv.com/india-news/nearly-half-of-the-fake-news-stories- in-india-are-political-study-7291481

NDTV, Nearly Half of the Fake News Stories in India Are Political: Study, (2024). https://www.ndtv.com/india-news/nearly-half-of-the-fake-news-stories- in-india-are-political-study-7291481. Accessed: 2025-07-20

work page 2024
[12]

Press release

Press Information Bureau, PM inaugurates, dedicates to the nation and lays the foundation stone for multiple development projects worth about Rs 13,000 crores in Varanasi,https://www.pib.gov.in/PressReleasePage.aspx?PRID=2040566 (2025).https://www.pib.gov.in/PressReleasePage.aspx?PRID=2040566. Press release

work page 2025
[13]

Com- plex & Intelligent Systems9(3), 2843–2863 (2023).https://doi.org/10.1007/ s40747-021-00552-1

Sharma, D.K., Garg, S.: IFND: a benchmark dataset for fake news detection. Com- plex & Intelligent Systems9(3), 2843–2863 (2023).https://doi.org/10.1007/ s40747-021-00552-1

work page 2023
[14]

In: ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp

Tian, K., Rao, G., Wang, X., Yu, M., Zhang, J., Zhang, L.: CMFNThinker: A Novel Cross-source Multi-modal Fake News Detection Model. In: ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 1–5 (2025).https : / / doi . org / 10 . 1109 / ICASSP49660 . 2025 . 10889602

work page 2025
[15]

Interna- tional Journal of Multimedia Information Retrieval12(2), 28 (2023).https : / / doi.org/10.1007/s13735-023-00296-3

Tufchi, S., Yadav, A., Ahmed, T.: A Comprehensive Survey of Multimodal Fake News Detection Techniques: Advances, Challenges, and Opportunities. Interna- tional Journal of Multimedia Information Retrieval12(2), 28 (2023).https : / / doi.org/10.1007/s13735-023-00296-3

work page doi:10.1007/s13735-023-00296-3 2023
[16]

Accessed: 2025-07-20

World Economic Forum, Global Risks Report 2024,https://www.weforum.org/ publications/global-risks-report-2024/(2024). Accessed: 2025-07-20

work page 2024
[17]

Yadav, A., Gaba, S., Khan, H., Budhiraja, I., Singh, A., Singh, K.K.: ETMA: Efficient Transformer-Based Multilevel Attention Framework for Multimodal Fake NewsDetection.IEEETransactionsonComputationalSocialSystems11(4),5015– 5027 (2024).https://doi.org/10.1109/TCSS.2023.3255242

work page doi:10.1109/tcss.2023.3255242 2024
[18]

https://doi.org/10.1109/ACCESS.2021.3114093

Ying, L., Yu, H., Wang, J., Ji, Y., Qian, S.: Multi-Level Multi-Modal Cross- AttentionNetworkforFakeNewsDetection.IEEEAccess9,132363–132373(2021). https://doi.org/10.1109/ACCESS.2021.3114093

work page doi:10.1109/access.2021.3114093 2021
[19]

Amortized Mixture Prior for Variational Sequence Generation

Zhang, T., Wang, D., Chen, H., Zeng, Z., Guo, W., Miao, C., Cui, L.: BDANN: BERT-Based Domain Adaptation Neural Network for Multi-Modal Fake News De- tection. In: 2020 International Joint Conference on Neural Networks (IJCNN), pp. 1–8 (2020).https://doi.org/10.1109/IJCNN48605.2020.9206973

work page doi:10.1109/ijcnn48605.2020.9206973 2020
[20]

In: 2023 IEEE International Conference on Multimedia and Expo (ICME), pp

Zhou,Y.,Yang,Y.,Ying,Q.,Qian,Z.,Zhang,X.:MultimodalFakeNewsDetection via CLIP-Guided Learning. In: 2023 IEEE International Conference on Multimedia and Expo (ICME), pp. 2825–2830 (2023).https://doi.org/10.1109/ICME55011. 2023.00480

work page doi:10.1109/icme55011 2023

[1] [1]

In: 2024 28th International Computer Science and Engineering Conference (ICSEC), pp

El-Amrany, S., Brust, M.R., Pecero, J.E., Bouvry, P.: Tri-FusionDet: Leveraging User Engagement, Textual, and Visual Features for Enhanced Fake News Detec- tion. In: 2024 28th International Computer Science and Engineering Conference (ICSEC), pp. 1–6 (2024).https://doi.org/10.1109/ICSEC62781.2024.10770746

work page doi:10.1109/icsec62781.2024.10770746 2024

[2] [2]

indiatimes

BrandEquity Bureau, 91 Per Cent Believe Fake News Can Influence Voting Deci- sions:Report,Accessed:2025-07-17.(2024).https://brandequity.economictimes. indiatimes . com / news / research / 91 - per - cent - believe - fake - news - can - influence-voting-decisions-report/110004307?utm_source=chatgpt.com

work page arXiv 2025

[3] [3]

(2024).https://www.communicationstoday.co.in/ by- 2025- 56- percent- new- indian- internet- users- from- rural- areas/?utm_ source=chatgpt.com

Communications Today, By 2025, 56 Percent New Indian Internet Users from Rural Areas, Accessed: 2025-07-20. (2024).https://www.communicationstoday.co.in/ by- 2025- 56- percent- new- indian- internet- users- from- rural- areas/?utm_ source=chatgpt.com

work page 2025

[4] [4]

Accessed: 2025-07-20

Exchange4Media, Misinformation Queries Raised, 1,575 Fake News Cases Flagged Since 2022: PIB, (2025).https://www.exchange4media.com/media-others-news/ 68914 - misinformation - queries - raised - 1575 - fake - news - cases - flagged - since-2022-pib-142128.html. Accessed: 2025-07-20

work page 2022

[5] [5]

Applied Soft Computing179, 113277 (2025)

Gedara, T.M.H., Loia, V., Tomasiello, S.: A fuzzy-based multimodal approach for interpretable fake news detection. Applied Soft Computing179, 113277 (2025). https://doi.org/10.1016/j.asoc.2025.113277.https://www.sciencedirect. com/science/article/pii/S1568494625005885

work page doi:10.1016/j.asoc.2025.113277.https://www.sciencedirect 2025

[6] [6]

https://indianexpress.com/article/india/214- rise- in- cases- relating- to-fake-news-rumours-7511534/

Indian Express, 214% Rise in Cases Relating to Fake News, Rumours, (2021). https://indianexpress.com/article/india/214- rise- in- cases- relating- to-fake-news-rumours-7511534/. Accessed: 2025-07-20

work page 2021

[7] [7]

Expert Systems with Applications95, 172–189 (2018).https : / / doi

Iyer,A.R.,Prasad,D.K.,Quek,C.H.:PIE-RSPOP:Abrain-inspiredpseudo-incremental ensemble rough set pseudo-outer product fuzzy neural network. Expert Systems with Applications95, 172–189 (2018).https : / / doi . org / 10 . 1016 / j . eswa . 2017 . 11 . 027.https : / / www . sciencedirect . com / science / article / pii / S0957417417307832 F2IND-IT! - Multimoda...

work page 2018

[8] [8]

arXiv:2411.06097 [cs.AI].https://arxiv.org/abs/ 2411.06097

Jun-hao, Xu, A Multimodal Adaptive Graph-based Intelligent Classification Model for Fake News, (2024). arXiv:2411.06097 [cs.AI].https://arxiv.org/abs/ 2411.06097

work page arXiv 2024

[9] [9]

IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics)31(6), 902–918 (2001).https://doi.org/10.1109/3477

Kasabov, N.: Evolving fuzzy neural networks for supervised/unsupervised online knowledge-based learning. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics)31(6), 902–918 (2001).https://doi.org/10.1109/3477. 969494

work page doi:10.1109/3477 2001

[10] [10]

IEEE Transactions on Fuzzy Systems 10(2), 144–154 (2002).https://doi.org/10.1109/91.995117

Kasabov, N., Song, Q.: DENFIS: dynamic evolving neural-fuzzy inference system and its application for time-series prediction. IEEE Transactions on Fuzzy Systems 10(2), 144–154 (2002).https://doi.org/10.1109/91.995117

work page doi:10.1109/91.995117 2002

[11] [11]

https://www.ndtv.com/india-news/nearly-half-of-the-fake-news-stories- in-india-are-political-study-7291481

NDTV, Nearly Half of the Fake News Stories in India Are Political: Study, (2024). https://www.ndtv.com/india-news/nearly-half-of-the-fake-news-stories- in-india-are-political-study-7291481. Accessed: 2025-07-20

work page 2024

[12] [12]

Press release

Press Information Bureau, PM inaugurates, dedicates to the nation and lays the foundation stone for multiple development projects worth about Rs 13,000 crores in Varanasi,https://www.pib.gov.in/PressReleasePage.aspx?PRID=2040566 (2025).https://www.pib.gov.in/PressReleasePage.aspx?PRID=2040566. Press release

work page 2025

[13] [13]

Com- plex & Intelligent Systems9(3), 2843–2863 (2023).https://doi.org/10.1007/ s40747-021-00552-1

Sharma, D.K., Garg, S.: IFND: a benchmark dataset for fake news detection. Com- plex & Intelligent Systems9(3), 2843–2863 (2023).https://doi.org/10.1007/ s40747-021-00552-1

work page 2023

[14] [14]

In: ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp

Tian, K., Rao, G., Wang, X., Yu, M., Zhang, J., Zhang, L.: CMFNThinker: A Novel Cross-source Multi-modal Fake News Detection Model. In: ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 1–5 (2025).https : / / doi . org / 10 . 1109 / ICASSP49660 . 2025 . 10889602

work page 2025

[15] [15]

Interna- tional Journal of Multimedia Information Retrieval12(2), 28 (2023).https : / / doi.org/10.1007/s13735-023-00296-3

Tufchi, S., Yadav, A., Ahmed, T.: A Comprehensive Survey of Multimodal Fake News Detection Techniques: Advances, Challenges, and Opportunities. Interna- tional Journal of Multimedia Information Retrieval12(2), 28 (2023).https : / / doi.org/10.1007/s13735-023-00296-3

work page doi:10.1007/s13735-023-00296-3 2023

[16] [16]

Accessed: 2025-07-20

World Economic Forum, Global Risks Report 2024,https://www.weforum.org/ publications/global-risks-report-2024/(2024). Accessed: 2025-07-20

work page 2024

[17] [17]

Yadav, A., Gaba, S., Khan, H., Budhiraja, I., Singh, A., Singh, K.K.: ETMA: Efficient Transformer-Based Multilevel Attention Framework for Multimodal Fake NewsDetection.IEEETransactionsonComputationalSocialSystems11(4),5015– 5027 (2024).https://doi.org/10.1109/TCSS.2023.3255242

work page doi:10.1109/tcss.2023.3255242 2024

[18] [18]

https://doi.org/10.1109/ACCESS.2021.3114093

Ying, L., Yu, H., Wang, J., Ji, Y., Qian, S.: Multi-Level Multi-Modal Cross- AttentionNetworkforFakeNewsDetection.IEEEAccess9,132363–132373(2021). https://doi.org/10.1109/ACCESS.2021.3114093

work page doi:10.1109/access.2021.3114093 2021

[19] [19]

Amortized Mixture Prior for Variational Sequence Generation

Zhang, T., Wang, D., Chen, H., Zeng, Z., Guo, W., Miao, C., Cui, L.: BDANN: BERT-Based Domain Adaptation Neural Network for Multi-Modal Fake News De- tection. In: 2020 International Joint Conference on Neural Networks (IJCNN), pp. 1–8 (2020).https://doi.org/10.1109/IJCNN48605.2020.9206973

work page doi:10.1109/ijcnn48605.2020.9206973 2020

[20] [20]

In: 2023 IEEE International Conference on Multimedia and Expo (ICME), pp

Zhou,Y.,Yang,Y.,Ying,Q.,Qian,Z.,Zhang,X.:MultimodalFakeNewsDetection via CLIP-Guided Learning. In: 2023 IEEE International Conference on Multimedia and Expo (ICME), pp. 2825–2830 (2023).https://doi.org/10.1109/ICME55011. 2023.00480

work page doi:10.1109/icme55011 2023