Slide
Centre Interdisciplinaire
de Recherche et d’Innovation
en Cybersécurité et Société
de Recherche et d’Innovation
en Cybersécurité et Société
1.
Moradi, A.; Zhu, Y.; Falk, T. H.
Towards Lightweight On-Device Audio Deepfake Detection Using Squeezeformers Proceedings Article
In: K., Adi; O., Nguena Timo; N., Boulahia-Cuppens; D., Espes; N., Stakhanova; M., Omar (Ed.): Lect. Notes Comput. Sci., pp. 376–389, Springer Science and Business Media Deutschland GmbH, 2026, ISBN: 03029743 (ISSN); 978-303220731-9 (ISBN), (Journal Abbreviation: Lect. Notes Comput. Sci.).
Abstract | Links | BibTeX | Tags: Audio DeepFake Detection, Detection mechanism, Edge Computing, Edge detection, Foundation models, High-accuracy, Large scale systems, Large-scale systems, Lightweight, Memory footprint, Performance, Real- time
@inproceedings{moradiLightweightOnDeviceAudio2026,
title = {Towards Lightweight On-Device Audio Deepfake Detection Using Squeezeformers},
author = {A. Moradi and Y. Zhu and T. H. Falk},
editor = {Adi K. and Nguena Timo O. and Boulahia-Cuppens N. and Espes D. and Stakhanova N. and Omar M.},
url = {https://www.scopus.com/pages/publications/105046137533?origin=resultslist},
doi = {10.1007/978-3-032-20732-6_24},
isbn = {03029743 (ISSN); 978-303220731-9 (ISBN)},
year = {2026},
date = {2026-01-01},
booktitle = {Lect. Notes Comput. Sci.},
volume = {16295 LNCS},
pages = {376–389},
publisher = {Springer Science and Business Media Deutschland GmbH},
abstract = {The increasing threat of audio deepfakes necessitates detection mechanisms that can operate in real-time on resource-constrained edge devices. While large-scale systems, such as detectors based on speech foundation models, have demonstrated high accuracy, their computational and memory footprints make them ill-suited for on-device applications. This paper addresses this critical gap by investigating the key factors that influence the performance of lightweight deepfake detection models. We conduct a systematic comparison of model architectures, input feature choices, and data augmentation techniques, evaluating both deepfake detection accuracy and computational complexity across three datasets. Our findings show that with proper modeling choices, a lightweight model can achieve performance comparable to that of a much larger model while being approximately 100× smaller in size. This work provides actionable insights for developing efficient and effective audio deepfake detectors tailored for the constraints of edge computing. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2026.},
note = {Journal Abbreviation: Lect. Notes Comput. Sci.},
keywords = {Audio DeepFake Detection, Detection mechanism, Edge Computing, Edge detection, Foundation models, High-accuracy, Large scale systems, Large-scale systems, Lightweight, Memory footprint, Performance, Real- time},
pubstate = {published},
tppubtype = {inproceedings}
}
The increasing threat of audio deepfakes necessitates detection mechanisms that can operate in real-time on resource-constrained edge devices. While large-scale systems, such as detectors based on speech foundation models, have demonstrated high accuracy, their computational and memory footprints make them ill-suited for on-device applications. This paper addresses this critical gap by investigating the key factors that influence the performance of lightweight deepfake detection models. We conduct a systematic comparison of model architectures, input feature choices, and data augmentation techniques, evaluating both deepfake detection accuracy and computational complexity across three datasets. Our findings show that with proper modeling choices, a lightweight model can achieve performance comparable to that of a much larger model while being approximately 100× smaller in size. This work provides actionable insights for developing efficient and effective audio deepfake detectors tailored for the constraints of edge computing. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2026.
2.
Moradi, A.; Falk, T. H.
Benchmarking Foundation Models for Cross-Domain Speaker Profiling Proceedings Article
In: IEEE Conf. Artif. Intell., CAI, pp. 78–84, Institute of Electrical and Electronics Engineers Inc., 2026, ISBN: 979-833156039-3 (ISBN), (Journal Abbreviation: IEEE Conf. Artif. Intell., CAI).
Abstract | Links | BibTeX | Tags: Benchmarking, Continuous speech recognition, Cross-domain, Forecasting, Formant frequency, Foundation models, Learning systems, Linguistics, Multi-attributes, Multi-task learning, Paralinguistic, Performance, Speaker identification, Speaker verification, Speech communication, State of the art, Verification task
@inproceedings{moradiBenchmarkingFoundationModels2026,
title = {Benchmarking Foundation Models for Cross-Domain Speaker Profiling},
author = {A. Moradi and T. H. Falk},
url = {https://www.scopus.com/pages/publications/105042046048?origin=resultslist},
doi = {10.1109/CAI68641.2026.11536565},
isbn = {979-833156039-3 (ISBN)},
year = {2026},
date = {2026-01-01},
booktitle = {IEEE Conf. Artif. Intell., CAI},
pages = {78–84},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {Speech conveys both linguistic and paralinguistic content. While pre-trained speech foundation models have been widely explored for linguistic tasks, such as speech recognition, and for speaker identification and verification tasks, very limited work has been done to test their usefulness for multi-attribute speaker profiling, i.e., simultaneous prediction of biological sex, age, and height from speech. In this paper, we aim to benchmark the performance of four state-of-the-art self-supervised foundation models, namely, WavLM, Wav2vec2, HuBERT, and XLSR-53, under both within- and cross-domain conditions. Each model is employed as a frozen pre-trained feature extractor, with lightweight task-specific heads trained jointly in a multi-task learning framework, and their feature extraction latency is empirically analyzed to assess practical deployment considerations. Experiments on two datasets show that WavLM achieves consistently strong within- and cross-domain performance for biological sex prediction, while Wav2vec2 and XLSR-53 exhibit more consistent performance for the age and height regression tasks in cross-domain settings. Interpretability analysis based on correlations with key acoustic features show (1) WavLM internal representations correlating highly with pitch and formant frequencies, corroborating the improved performance on biological sex prediction, and (2) XLSR-53 correlating highly with the second formant frequency, corroborating the results obtained for age and height. Overall, our analysis shows that pre-trained speech foundation models could serve as useful tools for cross-domain speaker profiling tasks. While no model stood out as a clear winner across all tested physical traits, future work could explore the use of ensemble methods for improved generalizability. © 2026 IEEE.},
note = {Journal Abbreviation: IEEE Conf. Artif. Intell., CAI},
keywords = {Benchmarking, Continuous speech recognition, Cross-domain, Forecasting, Formant frequency, Foundation models, Learning systems, Linguistics, Multi-attributes, Multi-task learning, Paralinguistic, Performance, Speaker identification, Speaker verification, Speech communication, State of the art, Verification task},
pubstate = {published},
tppubtype = {inproceedings}
}
Speech conveys both linguistic and paralinguistic content. While pre-trained speech foundation models have been widely explored for linguistic tasks, such as speech recognition, and for speaker identification and verification tasks, very limited work has been done to test their usefulness for multi-attribute speaker profiling, i.e., simultaneous prediction of biological sex, age, and height from speech. In this paper, we aim to benchmark the performance of four state-of-the-art self-supervised foundation models, namely, WavLM, Wav2vec2, HuBERT, and XLSR-53, under both within- and cross-domain conditions. Each model is employed as a frozen pre-trained feature extractor, with lightweight task-specific heads trained jointly in a multi-task learning framework, and their feature extraction latency is empirically analyzed to assess practical deployment considerations. Experiments on two datasets show that WavLM achieves consistently strong within- and cross-domain performance for biological sex prediction, while Wav2vec2 and XLSR-53 exhibit more consistent performance for the age and height regression tasks in cross-domain settings. Interpretability analysis based on correlations with key acoustic features show (1) WavLM internal representations correlating highly with pitch and formant frequencies, corroborating the improved performance on biological sex prediction, and (2) XLSR-53 correlating highly with the second formant frequency, corroborating the results obtained for age and height. Overall, our analysis shows that pre-trained speech foundation models could serve as useful tools for cross-domain speaker profiling tasks. While no model stood out as a clear winner across all tested physical traits, future work could explore the use of ensemble methods for improved generalizability. © 2026 IEEE.
3.
Pimentel, A.; Zhu, Y.; Falk, T. H.
Partial Audio Deepfake Detection: Are We Really Detecting Synthetic Media or Just Dataset Content Biases? Proceedings Article
In: IEEE Conf. Artif. Intell., CAI, pp. 1846–1851, Institute of Electrical and Electronics Engineers Inc., 2026, ISBN: 979-833156039-3 (ISBN), (Journal Abbreviation: IEEE Conf. Artif. Intell., CAI).
Abstract | Links | BibTeX | Tags: Audio signal processing, Computational linguistics, Condition, Detection models, Information integrity, Language model, Large datasets, Learning models, Optimistics, Performance, Self-supervised learning, Speech communication, Speech models, Speech recognition, Speech signals, Synthetic media, Transcription
@inproceedings{pimentelPartialAudioDeepfake2026,
title = {Partial Audio Deepfake Detection: Are We Really Detecting Synthetic Media or Just Dataset Content Biases?},
author = {A. Pimentel and Y. Zhu and T. H. Falk},
url = {https://www.scopus.com/pages/publications/105042045476?origin=resultslist},
doi = {10.1109/CAI68641.2026.11536622},
isbn = {979-833156039-3 (ISBN)},
year = {2026},
date = {2026-01-01},
booktitle = {IEEE Conf. Artif. Intell., CAI},
pages = {1846–1851},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {In recent years, the generation of highly realistic audio deepfakes has raised significant concerns regarding privacy and information integrity. While most research has focused on fully bonafide or spoofed speech, partial deep-fakes, where only segments of an utterance are manipulated, remain less explored. Given the nature of the task, existing datasets are relying on large language models to manipulate bonafide speech signals into partial deepfakes by altering, deleting, or replacing segments with synthetic content. These manipulations may alter the semantics and sentiment of the generated content, creating biases that can be captured by deepfake detection models, leading to overly optimistic performance and poor generalizability. In this work, we investigate the presence of such biases in two popular partial deepfake datasets, namely AV-Deepfake1M and PartialEdit. We explore four self-supervised learning models, two relying only on text transcriptions (RoBERTa and RoBERTa-sentiment) and two relying on universal speech representations (WavLM Large and Wav2vec2-XLSR). Our experiments show that, while speech models achieve higher in-domain performance, they do not generalize to out-of-domain conditions. In turn, models trained on only transcribed speech can effectively distinguish manipulated content, achieving up to 97.8% AUC in-domain and nearly 70% AUC out-of-domain. These results suggest that linguistic patterns and dataset confounds may indeed be biasing partial deepfake detection models, leading to poor generalizability. © 2026 IEEE.},
note = {Journal Abbreviation: IEEE Conf. Artif. Intell., CAI},
keywords = {Audio signal processing, Computational linguistics, Condition, Detection models, Information integrity, Language model, Large datasets, Learning models, Optimistics, Performance, Self-supervised learning, Speech communication, Speech models, Speech recognition, Speech signals, Synthetic media, Transcription},
pubstate = {published},
tppubtype = {inproceedings}
}
In recent years, the generation of highly realistic audio deepfakes has raised significant concerns regarding privacy and information integrity. While most research has focused on fully bonafide or spoofed speech, partial deep-fakes, where only segments of an utterance are manipulated, remain less explored. Given the nature of the task, existing datasets are relying on large language models to manipulate bonafide speech signals into partial deepfakes by altering, deleting, or replacing segments with synthetic content. These manipulations may alter the semantics and sentiment of the generated content, creating biases that can be captured by deepfake detection models, leading to overly optimistic performance and poor generalizability. In this work, we investigate the presence of such biases in two popular partial deepfake datasets, namely AV-Deepfake1M and PartialEdit. We explore four self-supervised learning models, two relying only on text transcriptions (RoBERTa and RoBERTa-sentiment) and two relying on universal speech representations (WavLM Large and Wav2vec2-XLSR). Our experiments show that, while speech models achieve higher in-domain performance, they do not generalize to out-of-domain conditions. In turn, models trained on only transcribed speech can effectively distinguish manipulated content, achieving up to 97.8% AUC in-domain and nearly 70% AUC out-of-domain. These results suggest that linguistic patterns and dataset confounds may indeed be biasing partial deepfake detection models, leading to poor generalizability. © 2026 IEEE.
4.
Hamadene, A.; Allili, M. S.
Cross-Model Deepfake Detection Through Contourlet-Based Inter-Channel Spectral Analysis Journal Article
In: IEEE Transactions on Biometrics, Behavior, and Identity Science, 2026, ISSN: 26376407 (ISSN).
Abstract | Links | BibTeX | Tags: Attribution, Computer forensics, Computer vision, Contourlet transform, Contourlets, Cross model, Cross-Model Generalization, Deep learning, Deepfake detection, Forensic engineering, Image Enhancement, Inter-channel contourlet feature, Inter-Channel Contourlet Features, Model generalization, Performance, Public trust, Spectral analyze, Spectrum analysis
@article{hamadeneCrossModelDeepfakeDetection2026,
title = {Cross-Model Deepfake Detection Through Contourlet-Based Inter-Channel Spectral Analysis},
author = {A. Hamadene and M. S. Allili},
url = {https://www.scopus.com/pages/publications/105038690534?origin=resultslist},
doi = {10.1109/TBIOM.2026.3687469},
issn = {26376407 (ISSN)},
year = {2026},
date = {2026-01-01},
journal = {IEEE Transactions on Biometrics, Behavior, and Identity Science},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {The rapid advancement of generative AI has fueled the proliferation of highly realistic deepfakes, raising growing concerns for public trust, digital forensics, and societal stability. While deep learning-based detectors have shown promising results in controlled settings, their effectiveness diminishes significantly across different datasets or unseen deepfakes. This performance drop reveals a core limitation: overfitting to dataset-specific artifacts and a failure to learn robust, content-agnostic forensic cues. In this work, we study a novel deepfake characterization that targets inter-channel spectral inconsistencies, a subtle but critical artifact of AI-generated imagery. Departing from conventional approaches that focus on spatial irregularities or within-channel frequency analysis, our method leverages the Contourlet Transform (CT) to extract high-frequency, multiscale, and multidirectional features across color channels. These features reveal latent spectral distortions and structural misalignments that persist across different generative models, including both GANs and Diffusion Models. Extensive experiments show that our approach generalizes well across datasets, achieving strong performance in both model-specific and cross-model detection tasks. Our findings highlight the promise of inter-channel spectral analysis for advancing the robustness and generalizability of deepfake detection systems. © 2019 IEEE.},
keywords = {Attribution, Computer forensics, Computer vision, Contourlet transform, Contourlets, Cross model, Cross-Model Generalization, Deep learning, Deepfake detection, Forensic engineering, Image Enhancement, Inter-channel contourlet feature, Inter-Channel Contourlet Features, Model generalization, Performance, Public trust, Spectral analyze, Spectrum analysis},
pubstate = {published},
tppubtype = {article}
}
The rapid advancement of generative AI has fueled the proliferation of highly realistic deepfakes, raising growing concerns for public trust, digital forensics, and societal stability. While deep learning-based detectors have shown promising results in controlled settings, their effectiveness diminishes significantly across different datasets or unseen deepfakes. This performance drop reveals a core limitation: overfitting to dataset-specific artifacts and a failure to learn robust, content-agnostic forensic cues. In this work, we study a novel deepfake characterization that targets inter-channel spectral inconsistencies, a subtle but critical artifact of AI-generated imagery. Departing from conventional approaches that focus on spatial irregularities or within-channel frequency analysis, our method leverages the Contourlet Transform (CT) to extract high-frequency, multiscale, and multidirectional features across color channels. These features reveal latent spectral distortions and structural misalignments that persist across different generative models, including both GANs and Diffusion Models. Extensive experiments show that our approach generalizes well across datasets, achieving strong performance in both model-specific and cross-model detection tasks. Our findings highlight the promise of inter-channel spectral analysis for advancing the robustness and generalizability of deepfake detection systems. © 2019 IEEE.
5.
Sadfi, Y.; Amamou, H.; Davoust, A.; Avila, A. R.
Comparative Analysis of Machine Learning, LLMs, and RAG for Fake News Detection Proceedings Article
In: IEEE Conf. Artif. Intell., CAI, pp. 2128–2133, Institute of Electrical and Electronics Engineers Inc., 2026, ISBN: 979-833156039-3 (ISBN), (Journal Abbreviation: IEEE Conf. Artif. Intell., CAI).
Abstract | Links | BibTeX | Tags: Classification (of information), Comparative analyzes, Comprehensive assessment, decision making, Decision-making process, Fake detection, Generalisation, High-accuracy, Language model, Learning algorithms, Learning systems, Machine learning, Machine learning algorithms, Machine-learning, Network platforms, Performance
@inproceedings{sadfiComparativeAnalysisMachine2026,
title = {Comparative Analysis of Machine Learning, LLMs, and RAG for Fake News Detection},
author = {Y. Sadfi and H. Amamou and A. Davoust and A. R. Avila},
url = {https://www.scopus.com/pages/publications/105042025800?origin=resultslist},
doi = {10.1109/CAI68641.2026.11536343},
isbn = {979-833156039-3 (ISBN)},
year = {2026},
date = {2026-01-01},
booktitle = {IEEE Conf. Artif. Intell., CAI},
pages = {2128–2133},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {The rapid spread of misinformation on social network platforms poses a serious threat to the decision-making processes in democratic societies. To mitigate the problem and cope with new misinformation scenarios several datasets have been proposed, followed by the emergence of new methods for detecting fake news. While several comparative analyses have been proposed to benchmark these solutions, the ever-growing development of new methods requires up-to-date analysis. For instance, recent approaches based on retrieval-augmented generation (RAG) have yet to be assessed and compared to previous methods used for fake news detection. In this study, we aim to fill this gap by presenting a comprehensive assessment of classical machine learning (ML) algorithms, Transformer-based models, such as BERT, large language models (LLMs), and RAG-based classifiers. We adopt four traditional fake news datasets, including a Portuguese one. Results indicate that classical ML and BERT deliver strong performance on in-domain tasks, demonstrating efficiency and high accuracy. RAG-augmented LLMs, on the other hand, offer enhanced generalization and robustness in cross-domain and cross-dataset settings, but are outperformed by traditional approaches for in-domain evaluation. © 2026 IEEE.},
note = {Journal Abbreviation: IEEE Conf. Artif. Intell., CAI},
keywords = {Classification (of information), Comparative analyzes, Comprehensive assessment, decision making, Decision-making process, Fake detection, Generalisation, High-accuracy, Language model, Learning algorithms, Learning systems, Machine learning, Machine learning algorithms, Machine-learning, Network platforms, Performance},
pubstate = {published},
tppubtype = {inproceedings}
}
The rapid spread of misinformation on social network platforms poses a serious threat to the decision-making processes in democratic societies. To mitigate the problem and cope with new misinformation scenarios several datasets have been proposed, followed by the emergence of new methods for detecting fake news. While several comparative analyses have been proposed to benchmark these solutions, the ever-growing development of new methods requires up-to-date analysis. For instance, recent approaches based on retrieval-augmented generation (RAG) have yet to be assessed and compared to previous methods used for fake news detection. In this study, we aim to fill this gap by presenting a comprehensive assessment of classical machine learning (ML) algorithms, Transformer-based models, such as BERT, large language models (LLMs), and RAG-based classifiers. We adopt four traditional fake news datasets, including a Portuguese one. Results indicate that classical ML and BERT deliver strong performance on in-domain tasks, demonstrating efficiency and high accuracy. RAG-augmented LLMs, on the other hand, offer enhanced generalization and robustness in cross-domain and cross-dataset settings, but are outperformed by traditional approaches for in-domain evaluation. © 2026 IEEE.
6.
Tiwari, A.; Arrabito, R.; Davoust, A.; Falk, T. H.
Quantifying Biological Sex Leakage in Electroencephalography-Based Mental Workload Measurement and Its Impact on Model Performance Proceedings Article
In: IEEE Int. Conf. Hum.-Mach. Syst., ICHMS, pp. 392–397, Institute of Electrical and Electronics Engineers Inc., 2026, ISBN: 979-833154511-6 (ISBN), (Journal Abbreviation: IEEE Int. Conf. Hum.-Mach. Syst., ICHMS).
Abstract | Links | BibTeX | Tags: Biomedical signal processing, Demographic information, Designing systems, Economic and social effects, Electroencephalography, Electrophysiology, Human operator, Mental workload, Modeling performance, Optimal performance, Performance, Population statistics, Real- time, Well being, Workload measurements
@inproceedings{tiwariQuantifyingBiologicalSex2026,
title = {Quantifying Biological Sex Leakage in Electroencephalography-Based Mental Workload Measurement and Its Impact on Model Performance},
author = {A. Tiwari and R. Arrabito and A. Davoust and T. H. Falk},
url = {https://www.scopus.com/pages/publications/105045575255?origin=resultslist},
doi = {10.1109/ICHMS69701.2026.11602273},
isbn = {979-833154511-6 (ISBN)},
year = {2026},
date = {2026-01-01},
booktitle = {IEEE Int. Conf. Hum.-Mach. Syst., ICHMS},
pages = {392–397},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {Detection of mental workload is essential for designing systems that can monitor and adapt to the needs of human operators, thereby ensuring their optimal performance and well-being. Features derived from neurophysiological signals, particularly electroencephalograms (EEG), provide a realtime, unobtrusive, and objective indicator of mental workload. However, EEG signals are often correlated with demographic variables, including biological sex and age, which can introduce systematic performance differences across demographic groups, while unintentionally encoding demographic information into models trained on EEG data. In this work, we investigate and quantify the influence of user demographics on mental workload detection models using fairness analysis. Furthermore, we propose domain-adversarial training, in which an adversary penalizes the model for retaining sex-related information, to mitigate leakage of private demographic information (biological sex). The proposed approach simultaneously increases the odds ratio for sex from 0.83 to 0.90 (values closer to 1 indicate reduced demographic disparity) while reducing statistical significance. This approach, however, also leads to a reduction in mental workload detection performance, with a 12.7% decrease in balanced accuracy (BACC), suggesting an inherent trade-off between performance and privacy protection. © 2026 IEEE.},
note = {Journal Abbreviation: IEEE Int. Conf. Hum.-Mach. Syst., ICHMS},
keywords = {Biomedical signal processing, Demographic information, Designing systems, Economic and social effects, Electroencephalography, Electrophysiology, Human operator, Mental workload, Modeling performance, Optimal performance, Performance, Population statistics, Real- time, Well being, Workload measurements},
pubstate = {published},
tppubtype = {inproceedings}
}
Detection of mental workload is essential for designing systems that can monitor and adapt to the needs of human operators, thereby ensuring their optimal performance and well-being. Features derived from neurophysiological signals, particularly electroencephalograms (EEG), provide a realtime, unobtrusive, and objective indicator of mental workload. However, EEG signals are often correlated with demographic variables, including biological sex and age, which can introduce systematic performance differences across demographic groups, while unintentionally encoding demographic information into models trained on EEG data. In this work, we investigate and quantify the influence of user demographics on mental workload detection models using fairness analysis. Furthermore, we propose domain-adversarial training, in which an adversary penalizes the model for retaining sex-related information, to mitigate leakage of private demographic information (biological sex). The proposed approach simultaneously increases the odds ratio for sex from 0.83 to 0.90 (values closer to 1 indicate reduced demographic disparity) while reducing statistical significance. This approach, however, also leads to a reduction in mental workload detection performance, with a 12.7% decrease in balanced accuracy (BACC), suggesting an inherent trade-off between performance and privacy protection. © 2026 IEEE.
7.
Lapointe, J. -F.; Allili, M. S.; Hammouche, N.
Field Trials of an AI-AR-Based System for Remote Bridge Inspection by Drone Proceedings Article
In: D., Harris; W.-C., Li; H., Krömker (Ed.): Lect. Notes Comput. Sci., pp. 278–287, Springer Science and Business Media Deutschland GmbH, 2025, ISBN: 03029743 (ISSN); 978-303176823-1 (ISBN), (Journal Abbreviation: Lect. Notes Comput. Sci.).
Abstract | Links | BibTeX | Tags: Advanced systems, Air navigation, Artificial intelligence, artificial intelligence (AI), augmented reality, augmented reality (AR), Bridge inspection, Concrete bridges, Drone, Drones, Field trial, HIgh speed networks, High-speed Networks, Network links, Performance, Remote guidance, Transportation infrastructures, UAV
@inproceedings{lapointe_field_2025,
title = {Field Trials of an AI-AR-Based System for Remote Bridge Inspection by Drone},
author = {J. -F. Lapointe and M. S. Allili and N. Hammouche},
editor = {Harris D. and Li W.-C. and Krömker H.},
url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85213387549&doi=10.1007%2f978-3-031-76824-8_20&partnerID=40&md5=565ae5dded9cfdf27632e79e702c7718},
doi = {10.1007/978-3-031-76824-8_20},
isbn = {03029743 (ISSN); 978-303176823-1 (ISBN)},
year = {2025},
date = {2025-01-01},
booktitle = {Lect. Notes Comput. Sci.},
volume = {15381 LNCS},
pages = {278–287},
publisher = {Springer Science and Business Media Deutschland GmbH},
abstract = {Bridge inspections are important to ensure the safety of users of these critical transportation infrastructures and avoid tragedies that could be caused by the collapse of these infrastructures. This paper describes the results of field trials of an advanced system for remotely guided inspection of bridges by a drone, which relies on artificial intelligence and augmented reality to achieve it. Results indicate that a high speed network link is critical to achieve good performance. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.},
note = {Journal Abbreviation: Lect. Notes Comput. Sci.},
keywords = {Advanced systems, Air navigation, Artificial intelligence, artificial intelligence (AI), augmented reality, augmented reality (AR), Bridge inspection, Concrete bridges, Drone, Drones, Field trial, HIgh speed networks, High-speed Networks, Network links, Performance, Remote guidance, Transportation infrastructures, UAV},
pubstate = {published},
tppubtype = {inproceedings}
}
Bridge inspections are important to ensure the safety of users of these critical transportation infrastructures and avoid tragedies that could be caused by the collapse of these infrastructures. This paper describes the results of field trials of an advanced system for remotely guided inspection of bridges by a drone, which relies on artificial intelligence and augmented reality to achieve it. Results indicate that a high speed network link is critical to achieve good performance. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.
8.
Allaoui, M.; Hedjam, R.; Bouanane, K.; Allili, M. S.; Kherfi, M. L.; Belhaouari, S. B.
Exploring non-negativity for improved manifold embedding: Application to t-SNE Journal Article
In: Knowledge-Based Systems, vol. 330, 2025, ISSN: 09507051 (ISSN).
Abstract | Links | BibTeX | Tags: Dimensionality reduction, Embedding technique, Embeddings, Gradient methods, Gradient-descent, Manifold embedding, Matrix algebra, Non-negative matrix factorization, Non-negativity, Nonnegative matrix factorization, Nonnegativity constraints, Performance, T-SNE
@article{allaoui_exploring_2025,
title = {Exploring non-negativity for improved manifold embedding: Application to t-SNE},
author = {M. Allaoui and R. Hedjam and K. Bouanane and M. S. Allili and M. L. Kherfi and S. B. Belhaouari},
url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-105018098090&doi=10.1016%2Fj.knosys.2025.114547&partnerID=40&md5=237540c38a928146d589b96cd6888547},
doi = {10.1016/j.knosys.2025.114547},
issn = {09507051 (ISSN)},
year = {2025},
date = {2025-01-01},
journal = {Knowledge-Based Systems},
volume = {330},
abstract = {Drawing inspiration from Non-negative Matrix Factorization (NMF), this paper explores the potential of incorporating non-negativity constraints into embedding techniques, with a focus on t-SNE as an application. Specifically, we investigate the following questions: Can enforcing non-negativity in the embedding space enhance interpretability and improve the quality of embedded data? By prioritizing non-negativity, can embedding methods achieve better performance and more meaningful representations? Additionally, does enforcing non-negativity in the embedded space help preserve both the local and global structure of data in the manifold, leading to more accurate and interpretable embeddings? In this work, we could show both objectively and subjectively how enforcing t-SNE to leverage the non-negativity of the data addresses the raised questions. To achieve this, we introduced a novel approach to transforming the additive update rule of the gradient descent used by t-SNE to a multiplicative counterpart to enforce the non-negativity in the embedded space. However, grappling with full non-negativity in the gradient descent formula presents challenges, prompting our focus solely on the (yi−yj) term, resulting in a semi-non-negative t-SNE algorithm, shortly named SN-tSNE. Nevertheless, experimental findings substantiate the significant impact of the proposed update rule on the performance and efficacy of the SN-tSNE algorithm. Furthermore, additional experiments are performed to compare SN-tSNE with its precursor t-SNE, as well as the competitive embedding technique UMAP, alongside other relevant embedding and dimensionality reduction models like NMF. The source code of SN-tSNE is available on GitHub (https://github.com/M-Allaoui/SN-tSNE.git). © 2025},
keywords = {Dimensionality reduction, Embedding technique, Embeddings, Gradient methods, Gradient-descent, Manifold embedding, Matrix algebra, Non-negative matrix factorization, Non-negativity, Nonnegative matrix factorization, Nonnegativity constraints, Performance, T-SNE},
pubstate = {published},
tppubtype = {article}
}
Drawing inspiration from Non-negative Matrix Factorization (NMF), this paper explores the potential of incorporating non-negativity constraints into embedding techniques, with a focus on t-SNE as an application. Specifically, we investigate the following questions: Can enforcing non-negativity in the embedding space enhance interpretability and improve the quality of embedded data? By prioritizing non-negativity, can embedding methods achieve better performance and more meaningful representations? Additionally, does enforcing non-negativity in the embedded space help preserve both the local and global structure of data in the manifold, leading to more accurate and interpretable embeddings? In this work, we could show both objectively and subjectively how enforcing t-SNE to leverage the non-negativity of the data addresses the raised questions. To achieve this, we introduced a novel approach to transforming the additive update rule of the gradient descent used by t-SNE to a multiplicative counterpart to enforce the non-negativity in the embedded space. However, grappling with full non-negativity in the gradient descent formula presents challenges, prompting our focus solely on the (yi−yj) term, resulting in a semi-non-negative t-SNE algorithm, shortly named SN-tSNE. Nevertheless, experimental findings substantiate the significant impact of the proposed update rule on the performance and efficacy of the SN-tSNE algorithm. Furthermore, additional experiments are performed to compare SN-tSNE with its precursor t-SNE, as well as the competitive embedding technique UMAP, alongside other relevant embedding and dimensionality reduction models like NMF. The source code of SN-tSNE is available on GitHub (https://github.com/M-Allaoui/SN-tSNE.git). © 2025
9.
Lapointe, J. -F.; Allili, M. S.; Hammouche, N.
Field Trials of an AI-AR-Based System for Remote Bridge Inspection by Drone Proceedings Article
In: D., Harris; W.-C., Li; H., Krömker (Ed.): Lect. Notes Comput. Sci., pp. 278–287, Springer Science and Business Media Deutschland GmbH, 2025, ISBN: 03029743 (ISSN); 978-303176823-1 (ISBN), (Journal Abbreviation: Lect. Notes Comput. Sci.).
Abstract | Links | BibTeX | Tags: Advanced systems, Air navigation, Artificial intelligence, artificial intelligence (AI), augmented reality, augmented reality (AR), Bridge inspection, Concrete bridges, Drone, Drones, Field trial, HIgh speed networks, High-speed Networks, Network links, Performance, Remote guidance, Transportation infrastructures, UAV
@inproceedings{lapointeFieldTrialsAIARBased2025a,
title = {Field Trials of an AI-AR-Based System for Remote Bridge Inspection by Drone},
author = {J. -F. Lapointe and M. S. Allili and N. Hammouche},
editor = {Harris D. and Li W.-C. and Krömker H.},
url = {https://www.scopus.com/pages/publications/85213387549?origin=resultslist},
doi = {10.1007/978-3-031-76824-8_20},
isbn = {03029743 (ISSN); 978-303176823-1 (ISBN)},
year = {2025},
date = {2025-01-01},
booktitle = {Lect. Notes Comput. Sci.},
volume = {15381 LNCS},
pages = {278–287},
publisher = {Springer Science and Business Media Deutschland GmbH},
abstract = {Bridge inspections are important to ensure the safety of users of these critical transportation infrastructures and avoid tragedies that could be caused by the collapse of these infrastructures. This paper describes the results of field trials of an advanced system for remotely guided inspection of bridges by a drone, which relies on artificial intelligence and augmented reality to achieve it. Results indicate that a high speed network link is critical to achieve good performance. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.},
note = {Journal Abbreviation: Lect. Notes Comput. Sci.},
keywords = {Advanced systems, Air navigation, Artificial intelligence, artificial intelligence (AI), augmented reality, augmented reality (AR), Bridge inspection, Concrete bridges, Drone, Drones, Field trial, HIgh speed networks, High-speed Networks, Network links, Performance, Remote guidance, Transportation infrastructures, UAV},
pubstate = {published},
tppubtype = {inproceedings}
}
Bridge inspections are important to ensure the safety of users of these critical transportation infrastructures and avoid tragedies that could be caused by the collapse of these infrastructures. This paper describes the results of field trials of an advanced system for remotely guided inspection of bridges by a drone, which relies on artificial intelligence and augmented reality to achieve it. Results indicate that a high speed network link is critical to achieve good performance. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.
10.
Guimaraes, H. R.; Abdollahi, M.; Zhu, Y.; Maucourt, S.; Coallier, N.; Giovenazzo, P.; Falk, T. H.
Benchmarking Self-Supervised Audio Representations for IoT-Enabled Acoustic Beehive Monitoring Journal Article
In: IEEE Internet of Things Journal, vol. 12, no. 21, pp. 45000–45010, 2025, ISSN: 23274662 (ISSN).
Abstract | Links | BibTeX | Tags: Acoustics, Audio acoustics, Audio representation, Beehive monitoring, Benchmarking, Bioacoustics, Computer vision applications, Deep learning, Honeybee, honeybees, Internet of Things (IoT), IoT, Labeled data, Performance, Real time systems, Self-supervised learning, self-supervised learning (SSL), Societal benefits, Speech applications, Speech recognition, Supervised learning, Universal feature extractors
@article{guimaraesBenchmarkingSelfSupervisedAudio2025,
title = {Benchmarking Self-Supervised Audio Representations for IoT-Enabled Acoustic Beehive Monitoring},
author = {H. R. Guimaraes and M. Abdollahi and Y. Zhu and S. Maucourt and N. Coallier and P. Giovenazzo and T. H. Falk},
url = {https://www.scopus.com/pages/publications/105013592730?origin=resultslist},
doi = {10.1109/JIOT.2025.3599483},
issn = {23274662 (ISSN)},
year = {2025},
date = {2025-01-01},
journal = {IEEE Internet of Things Journal},
volume = {12},
number = {21},
pages = {45000–45010},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {Self-supervised learning (SSL) has enabled the development of universal feature extractors that have redefined the performance envelope of computer vision and speech applications. Recent works have started to explore SSL in other domains, including bioacoustics, which could have significant societal benefits. Honeybees (Apis mellifera), for example, are crucial pollinators contributing to one-third of global food production. However, massive colony losses in recent years have raised concerns. Traditional hive monitoring methods rely on intrusive visual inspections by beekeepers, which can further disrupt colony dynamics. As such, Internet of Things (IoT)-based automated monitoring systems have emerged, integrating environmental and bioacoustic sensing to enable real-time, noninvasive hive assessment. In this work, we introduce a comprehensive evaluation and benchmarking of general-purpose and bioacoustic audio representations that generalize across various tasks in IoT-enabled acoustic beehive monitoring, even with limited labeled data. Herein, fourteen models are evaluated across four critical tasks: beehive state detection, beehive strength assessment, buzzing identification, and beekeeper voice activity detection. Reported results demonstrate the strong generalizability of existing representations, paving the way for advanced, scalable honeybee colony monitoring and preservation. © 2014 IEEE.},
keywords = {Acoustics, Audio acoustics, Audio representation, Beehive monitoring, Benchmarking, Bioacoustics, Computer vision applications, Deep learning, Honeybee, honeybees, Internet of Things (IoT), IoT, Labeled data, Performance, Real time systems, Self-supervised learning, self-supervised learning (SSL), Societal benefits, Speech applications, Speech recognition, Supervised learning, Universal feature extractors},
pubstate = {published},
tppubtype = {article}
}
Self-supervised learning (SSL) has enabled the development of universal feature extractors that have redefined the performance envelope of computer vision and speech applications. Recent works have started to explore SSL in other domains, including bioacoustics, which could have significant societal benefits. Honeybees (Apis mellifera), for example, are crucial pollinators contributing to one-third of global food production. However, massive colony losses in recent years have raised concerns. Traditional hive monitoring methods rely on intrusive visual inspections by beekeepers, which can further disrupt colony dynamics. As such, Internet of Things (IoT)-based automated monitoring systems have emerged, integrating environmental and bioacoustic sensing to enable real-time, noninvasive hive assessment. In this work, we introduce a comprehensive evaluation and benchmarking of general-purpose and bioacoustic audio representations that generalize across various tasks in IoT-enabled acoustic beehive monitoring, even with limited labeled data. Herein, fourteen models are evaluated across four critical tasks: beehive state detection, beehive strength assessment, buzzing identification, and beekeeper voice activity detection. Reported results demonstrate the strong generalizability of existing representations, paving the way for advanced, scalable honeybee colony monitoring and preservation. © 2014 IEEE.