Slide
Centre Interdisciplinaire
de Recherche et d’Innovation
en Cybersécurité et Société
de Recherche et d’Innovation
en Cybersécurité et Société
1.
Allaoui, M. L.; Allili, M. S.
LG-TargetSeg: Deep semantic fusion via language-conditioned selective scanning for targeted medical image segmentation Journal Article
In: Information Fusion, vol. 139, 2027, ISSN: 15662535 (ISSN).
Abstract | Links | BibTeX | Tags: Computerized tomography, decoding, Deep vision-language fusion, Encoder-decoder, Gaussians, Medical image processing, Medical image segmentation, Modeling languages, Natural language processing systems, Reinforced gaussian dice loss, Selective scanning, Semantic fusion, Semantic segmentation, Semantics, Shallow levels, Signal encoding, State space methods, State-space modeling, State-space models, Training time
@article{allaouiLGTargetSegDeepSemantic2027,
title = {LG-TargetSeg: Deep semantic fusion via language-conditioned selective scanning for targeted medical image segmentation},
author = {M. L. Allaoui and M. S. Allili},
url = {https://www.scopus.com/pages/publications/105051242411?origin=resultslist},
doi = {10.1016/j.inffus.2026.104797},
issn = {15662535 (ISSN)},
year = {2027},
date = {2027-01-01},
journal = {Information Fusion},
volume = {139},
publisher = {Elsevier B.V.},
abstract = {Clinical language defines the semantic intent of targeted medical image segmentation, yet most vision–language methods incorporate it only at a shallow level or through training-time conditioning, leaving spatial reasoning largely language-blind. We propose LG-TargetSeg (Language-Guided Targeted Segmentation), a cross-modal framework that embeds clinical language directly into the selective scan of a Mamba-based encoder-decoder. Natural-language prompts, encoded by a frozen MedSigLIP foundation model, shape selective scanning throughout the encoder–decoder and additionally adapt the bottleneck state transition, alongside gated modulation and cross-attention, enabling semantic guidance to continuously shape feature propagation toward the target anatomy. A Text-Driven Selective Scan 2D (TD-SS2D) module achieves prompt-aware multi-scale fusion with linear complexity in the number of spatial tokens. To improve robustness to boundary ambiguity and class imbalance, we introduce the Reinforced Gaussian Dice Loss (RGDL) with learnable weighting parameters. Evaluated under both task-specific and universal generalization settings, LG-TargetSeg achieves state-of-the-art performance. Moreover, its prompt-driven selectivity remains reliable under genuine multi-target competition: on abdominal CT images containing up to 16 annotated organs, changing only the text prompt on a frozen model produces organ masks with less than 1% pairwise overlap. Additional materials and code are available at: https://github.com/MOHAMEDLamine77/LG-TargetSeg . © 2026 Elsevier B.V.},
keywords = {Computerized tomography, decoding, Deep vision-language fusion, Encoder-decoder, Gaussians, Medical image processing, Medical image segmentation, Modeling languages, Natural language processing systems, Reinforced gaussian dice loss, Selective scanning, Semantic fusion, Semantic segmentation, Semantics, Shallow levels, Signal encoding, State space methods, State-space modeling, State-space models, Training time},
pubstate = {published},
tppubtype = {article}
}
Clinical language defines the semantic intent of targeted medical image segmentation, yet most vision–language methods incorporate it only at a shallow level or through training-time conditioning, leaving spatial reasoning largely language-blind. We propose LG-TargetSeg (Language-Guided Targeted Segmentation), a cross-modal framework that embeds clinical language directly into the selective scan of a Mamba-based encoder-decoder. Natural-language prompts, encoded by a frozen MedSigLIP foundation model, shape selective scanning throughout the encoder–decoder and additionally adapt the bottleneck state transition, alongside gated modulation and cross-attention, enabling semantic guidance to continuously shape feature propagation toward the target anatomy. A Text-Driven Selective Scan 2D (TD-SS2D) module achieves prompt-aware multi-scale fusion with linear complexity in the number of spatial tokens. To improve robustness to boundary ambiguity and class imbalance, we introduce the Reinforced Gaussian Dice Loss (RGDL) with learnable weighting parameters. Evaluated under both task-specific and universal generalization settings, LG-TargetSeg achieves state-of-the-art performance. Moreover, its prompt-driven selectivity remains reliable under genuine multi-target competition: on abdominal CT images containing up to 16 annotated organs, changing only the text prompt on a frozen model produces organ masks with less than 1% pairwise overlap. Additional materials and code are available at: https://github.com/MOHAMEDLamine77/LG-TargetSeg . © 2026 Elsevier B.V.
2.
Laamari, A.; Moudoud, H.; Houda, Z. A. El
Lightweight LLM Adaptation for Intrusion Detection via Token-Efficient Flow Representation Proceedings Article
In: IEEE Conf. Artif. Intell., CAI, pp. 2122–2127, 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), Data flow analysis, Decoder-only large language model, Decoder-only LLMs, decoding, Flow classification, Intrusion Detection, Intrusion-Detection, Language model, Large language model, LLMs, LoRA, Low-rank adaptation, Network Flow Classification, Network intrusion, Network security, Networks flows, Qwen2.5, Signal encoding, T5-Small, Token-oriented object notation, Tokenization, TOON
@inproceedings{laamariLightweightLLMAdaptation2026,
title = {Lightweight LLM Adaptation for Intrusion Detection via Token-Efficient Flow Representation},
author = {A. Laamari and H. Moudoud and Z. A. El Houda},
url = {https://www.scopus.com/pages/publications/105042133342?origin=resultslist},
doi = {10.1109/CAI68641.2026.11536533},
isbn = {979-833156039-3 (ISBN)},
year = {2026},
date = {2026-01-01},
booktitle = {IEEE Conf. Artif. Intell., CAI},
pages = {2122–2127},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {Large language models (LLMs) are emerging as a promising approach for intrusion detection using structured network flow data. However, their practical deployment is constrained by context window limitations and the excessive token overhead introduced by conventional tabular serialization formats such as JSON. Verbose data representations inflate sequence lengths, often exceeding model input limits and causing feature truncation. Additionally, it remains unclear which LLM architecture is more suitable for structured intrusion detection tasks under limited training resources. To tackle this issue, we propose a novel representation-aware intrusion detection framework based on Token-Oriented Object Notation (TOON), a compact serialization format that maximizes token efficiency while preserving schema structure. Also, we integrate a Low-Rank Adaptation (LoRA) scheme to enable parameter-efficient fine-tuning. Finally, we evaluate the proposed framework as an encoder-decoder model (T5-Small) with a decoder-only model (Qwen2.5) on three benchmark datasets, including NSL-KDD, UNSW-NB15, and CIC-IDS2018 for binary and multi-class classification scenarios. The results show that our tokenizer-aligned, representation-aware preprocessing combined with lightweight encoder-decoder adaptation provides a practical and resource-efficient foundation for LLM-based intrusion detection. © 2026 IEEE.},
note = {Journal Abbreviation: IEEE Conf. Artif. Intell., CAI},
keywords = {Classification (of information), Data flow analysis, Decoder-only large language model, Decoder-only LLMs, decoding, Flow classification, Intrusion Detection, Intrusion-Detection, Language model, Large language model, LLMs, LoRA, Low-rank adaptation, Network Flow Classification, Network intrusion, Network security, Networks flows, Qwen2.5, Signal encoding, T5-Small, Token-oriented object notation, Tokenization, TOON},
pubstate = {published},
tppubtype = {inproceedings}
}
Large language models (LLMs) are emerging as a promising approach for intrusion detection using structured network flow data. However, their practical deployment is constrained by context window limitations and the excessive token overhead introduced by conventional tabular serialization formats such as JSON. Verbose data representations inflate sequence lengths, often exceeding model input limits and causing feature truncation. Additionally, it remains unclear which LLM architecture is more suitable for structured intrusion detection tasks under limited training resources. To tackle this issue, we propose a novel representation-aware intrusion detection framework based on Token-Oriented Object Notation (TOON), a compact serialization format that maximizes token efficiency while preserving schema structure. Also, we integrate a Low-Rank Adaptation (LoRA) scheme to enable parameter-efficient fine-tuning. Finally, we evaluate the proposed framework as an encoder-decoder model (T5-Small) with a decoder-only model (Qwen2.5) on three benchmark datasets, including NSL-KDD, UNSW-NB15, and CIC-IDS2018 for binary and multi-class classification scenarios. The results show that our tokenizer-aligned, representation-aware preprocessing combined with lightweight encoder-decoder adaptation provides a practical and resource-efficient foundation for LLM-based intrusion detection. © 2026 IEEE.
3.
Saumure, C.; Plouffe-Demers, M. -P.; Fiset, D.; Cormier, S.; Zhang, Y.; Sun, D.; Feng, M.; Luo, F.; Kunz, M.; Blais, C.
In: Affective Science, vol. 4, no. 2, pp. 332–349, 2023, ISSN: 26622041 (ISSN), (Publisher: Springer Nature).
Abstract | Links | BibTeX | Tags: Cross-cultural, decoding, Intensity, Pain communication, Pain facial expressions
@article{saumure_differences_2023,
title = {Differences Between East Asians and Westerners in the Mental Representations and Visual Information Extraction Involved in the Decoding of Pain Facial Expression Intensity},
author = {C. Saumure and M. -P. Plouffe-Demers and D. Fiset and S. Cormier and Y. Zhang and D. Sun and M. Feng and F. Luo and M. Kunz and C. Blais},
url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85159314374&doi=10.1007%2fs42761-023-00186-1&partnerID=40&md5=e8b8d7632d87842fe6bae4dfd6c663af},
doi = {10.1007/s42761-023-00186-1},
issn = {26622041 (ISSN)},
year = {2023},
date = {2023-01-01},
journal = {Affective Science},
volume = {4},
number = {2},
pages = {332–349},
publisher = {Springer Nature},
abstract = {Effectively communicating pain is crucial for human beings. Facial expressions are one of the most specific forms of behavior associated with pain, but the way culture shapes expectations about the intensity with which pain is typically facially conveyed, and the visual strategies deployed to decode pain intensity in facial expressions, is poorly understood. The present study used a data-driven approach to compare two cultures, namely East Asians and Westerners, with respect to their mental representations of pain facial expressions (experiment 1},
note = {Publisher: Springer Nature},
keywords = {Cross-cultural, decoding, Intensity, Pain communication, Pain facial expressions},
pubstate = {published},
tppubtype = {article}
}
Effectively communicating pain is crucial for human beings. Facial expressions are one of the most specific forms of behavior associated with pain, but the way culture shapes expectations about the intensity with which pain is typically facially conveyed, and the visual strategies deployed to decode pain intensity in facial expressions, is poorly understood. The present study used a data-driven approach to compare two cultures, namely East Asians and Westerners, with respect to their mental representations of pain facial expressions (experiment 1
4.
Saumure, C.; Plouffe-Demers, M. -P.; Fiset, D.; Cormier, S.; Zhang, Y.; Sun, D.; Feng, M.; Luo, F.; Kunz, M.; Blais, C.
In: Affective Science, vol. 4, no. 2, pp. 332–349, 2023, ISSN: 26622041 (ISSN).
Abstract | Links | BibTeX | Tags: Cross-cultural, decoding, Intensity, Pain communication, Pain facial expressions
@article{saumureDifferencesEastAsians2023,
title = {Differences Between East Asians and Westerners in the Mental Representations and Visual Information Extraction Involved in the Decoding of Pain Facial Expression Intensity},
author = {C. Saumure and M. -P. Plouffe-Demers and D. Fiset and S. Cormier and Y. Zhang and D. Sun and M. Feng and F. Luo and M. Kunz and C. Blais},
url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85159314374&doi=10.1007%2fs42761-023-00186-1&partnerID=40&md5=e8b8d7632d87842fe6bae4dfd6c663af},
doi = {10.1007/s42761-023-00186-1},
issn = {26622041 (ISSN)},
year = {2023},
date = {2023-01-01},
journal = {Affective Science},
volume = {4},
number = {2},
pages = {332–349},
publisher = {Springer Nature},
abstract = {Effectively communicating pain is crucial for human beings. Facial expressions are one of the most specific forms of behavior associated with pain, but the way culture shapes expectations about the intensity with which pain is typically facially conveyed, and the visual strategies deployed to decode pain intensity in facial expressions, is poorly understood. The present study used a data-driven approach to compare two cultures, namely East Asians and Westerners, with respect to their mental representations of pain facial expressions (experiment 1},
keywords = {Cross-cultural, decoding, Intensity, Pain communication, Pain facial expressions},
pubstate = {published},
tppubtype = {article}
}
Effectively communicating pain is crucial for human beings. Facial expressions are one of the most specific forms of behavior associated with pain, but the way culture shapes expectations about the intensity with which pain is typically facially conveyed, and the visual strategies deployed to decode pain intensity in facial expressions, is poorly understood. The present study used a data-driven approach to compare two cultures, namely East Asians and Westerners, with respect to their mental representations of pain facial expressions (experiment 1
5.
Blais, C.; Fiset, D.; Furumoto-Deshaies, H.; Kunz, M.; Seuss, D.; Cormier, S.
Facial Features Underlying the Decoding of Pain Expressions Journal Article
In: Journal of Pain, vol. 20, no. 6, pp. 728–738, 2019, ISSN: 15265900 (ISSN), (Publisher: Churchill Livingstone Inc.).
Abstract | Links | BibTeX | Tags: adult, article, attention, decoding, Eye, Facial Expression, facies, female, human, human experiment, Humans, lip, male, Memory, Nose, Pain, pain assessment, pain dimensions, pain measurement, sensory analysis, wrinkle, Young Adult
@article{blais_facial_2019,
title = {Facial Features Underlying the Decoding of Pain Expressions},
author = {C. Blais and D. Fiset and H. Furumoto-Deshaies and M. Kunz and D. Seuss and S. Cormier},
url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85060707205&doi=10.1016%2fj.jpain.2019.01.002&partnerID=40&md5=2d2dd305de430a7ce8973644f57a4996},
doi = {10.1016/j.jpain.2019.01.002},
issn = {15265900 (ISSN)},
year = {2019},
date = {2019-01-01},
journal = {Journal of Pain},
volume = {20},
number = {6},
pages = {728–738},
publisher = {Churchill Livingstone Inc.},
abstract = {Previous research has revealed that the face is a finely tuned medium for pain communication. Studies assessing the decoding of facial expressions of pain have revealed an interesting discrepancy, namely that, despite eyes narrowing being the most frequent facial expression accompanying pain, individuals mostly rely on brow lowering and nose wrinkling/upper lip raising to evaluate pain. The present study verifies if this discrepancy may reflect an interaction between the features coding pain expressions and the features used by observers and stored in their mental representations. Experiment 1 shows that more weight is allocated to the brow lowering and nose wrinkling/upper lip raising, supporting the idea that these features are allocated more importance when mental representations of pain expressions are stored in memory. These 2 features have been associated with negative valence and with the affective dimension of pain, whereas the eyes narrowing feature has been associated more closely with the sensory dimension of pain. However, experiment 2 shows that these 2 features remain more salient than eyes narrowing, even when attention is specifically directed toward the sensory dimension of pain. Together, these results suggest that the features most saliently coded in the mental representation of facial expressions of pain may reflect a bias toward allocating more weight to the affective information encoded in the face. Perspective: This work reveals the relative importance of 3 facial features representing the core of pain expressions during pain decoding. The results show that 2 features are over-represented; this finding may potentially be linked with the estimation biases occurring when clinicians and lay persons evaluate pain based on facial appearance. © 2019 the American Pain Society},
note = {Publisher: Churchill Livingstone Inc.},
keywords = {adult, article, attention, decoding, Eye, Facial Expression, facies, female, human, human experiment, Humans, lip, male, Memory, Nose, Pain, pain assessment, pain dimensions, pain measurement, sensory analysis, wrinkle, Young Adult},
pubstate = {published},
tppubtype = {article}
}
Previous research has revealed that the face is a finely tuned medium for pain communication. Studies assessing the decoding of facial expressions of pain have revealed an interesting discrepancy, namely that, despite eyes narrowing being the most frequent facial expression accompanying pain, individuals mostly rely on brow lowering and nose wrinkling/upper lip raising to evaluate pain. The present study verifies if this discrepancy may reflect an interaction between the features coding pain expressions and the features used by observers and stored in their mental representations. Experiment 1 shows that more weight is allocated to the brow lowering and nose wrinkling/upper lip raising, supporting the idea that these features are allocated more importance when mental representations of pain expressions are stored in memory. These 2 features have been associated with negative valence and with the affective dimension of pain, whereas the eyes narrowing feature has been associated more closely with the sensory dimension of pain. However, experiment 2 shows that these 2 features remain more salient than eyes narrowing, even when attention is specifically directed toward the sensory dimension of pain. Together, these results suggest that the features most saliently coded in the mental representation of facial expressions of pain may reflect a bias toward allocating more weight to the affective information encoded in the face. Perspective: This work reveals the relative importance of 3 facial features representing the core of pain expressions during pain decoding. The results show that 2 features are over-represented; this finding may potentially be linked with the estimation biases occurring when clinicians and lay persons evaluate pain based on facial appearance. © 2019 the American Pain Society
6.
Blais, C.; Fiset, D.; Furumoto-Deshaies, H.; Kunz, M.; Seuss, D.; Cormier, S.
Facial Features Underlying the Decoding of Pain Expressions Journal Article
In: Journal of Pain, vol. 20, no. 6, pp. 728–738, 2019, ISSN: 15265900 (ISSN).
Abstract | Links | BibTeX | Tags: adult, article, attention, decoding, Eye, Facial Expression, facies, female, human, human experiment, Humans, lip, male, Memory, Nose, Pain, pain assessment, pain dimensions, pain measurement, sensory analysis, wrinkle, Young Adult
@article{blaisFacialFeaturesUnderlying2019,
title = {Facial Features Underlying the Decoding of Pain Expressions},
author = {C. Blais and D. Fiset and H. Furumoto-Deshaies and M. Kunz and D. Seuss and S. Cormier},
url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85060707205&doi=10.1016%2fj.jpain.2019.01.002&partnerID=40&md5=2d2dd305de430a7ce8973644f57a4996},
doi = {10.1016/j.jpain.2019.01.002},
issn = {15265900 (ISSN)},
year = {2019},
date = {2019-01-01},
journal = {Journal of Pain},
volume = {20},
number = {6},
pages = {728–738},
publisher = {Churchill Livingstone Inc.},
abstract = {Previous research has revealed that the face is a finely tuned medium for pain communication. Studies assessing the decoding of facial expressions of pain have revealed an interesting discrepancy, namely that, despite eyes narrowing being the most frequent facial expression accompanying pain, individuals mostly rely on brow lowering and nose wrinkling/upper lip raising to evaluate pain. The present study verifies if this discrepancy may reflect an interaction between the features coding pain expressions and the features used by observers and stored in their mental representations. Experiment 1 shows that more weight is allocated to the brow lowering and nose wrinkling/upper lip raising, supporting the idea that these features are allocated more importance when mental representations of pain expressions are stored in memory. These 2 features have been associated with negative valence and with the affective dimension of pain, whereas the eyes narrowing feature has been associated more closely with the sensory dimension of pain. However, experiment 2 shows that these 2 features remain more salient than eyes narrowing, even when attention is specifically directed toward the sensory dimension of pain. Together, these results suggest that the features most saliently coded in the mental representation of facial expressions of pain may reflect a bias toward allocating more weight to the affective information encoded in the face. Perspective: This work reveals the relative importance of 3 facial features representing the core of pain expressions during pain decoding. The results show that 2 features are over-represented; this finding may potentially be linked with the estimation biases occurring when clinicians and lay persons evaluate pain based on facial appearance. © 2019 the American Pain Society},
keywords = {adult, article, attention, decoding, Eye, Facial Expression, facies, female, human, human experiment, Humans, lip, male, Memory, Nose, Pain, pain assessment, pain dimensions, pain measurement, sensory analysis, wrinkle, Young Adult},
pubstate = {published},
tppubtype = {article}
}
Previous research has revealed that the face is a finely tuned medium for pain communication. Studies assessing the decoding of facial expressions of pain have revealed an interesting discrepancy, namely that, despite eyes narrowing being the most frequent facial expression accompanying pain, individuals mostly rely on brow lowering and nose wrinkling/upper lip raising to evaluate pain. The present study verifies if this discrepancy may reflect an interaction between the features coding pain expressions and the features used by observers and stored in their mental representations. Experiment 1 shows that more weight is allocated to the brow lowering and nose wrinkling/upper lip raising, supporting the idea that these features are allocated more importance when mental representations of pain expressions are stored in memory. These 2 features have been associated with negative valence and with the affective dimension of pain, whereas the eyes narrowing feature has been associated more closely with the sensory dimension of pain. However, experiment 2 shows that these 2 features remain more salient than eyes narrowing, even when attention is specifically directed toward the sensory dimension of pain. Together, these results suggest that the features most saliently coded in the mental representation of facial expressions of pain may reflect a bias toward allocating more weight to the affective information encoded in the face. Perspective: This work reveals the relative importance of 3 facial features representing the core of pain expressions during pain decoding. The results show that 2 features are over-represented; this finding may potentially be linked with the estimation biases occurring when clinicians and lay persons evaluate pain based on facial appearance. © 2019 the American Pain Society