Lyes BADIS

Dr. Lyes BADIS

orcid

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
25 September 2024 in Informatics and Automation
Verified
Journal Article
Open access
PDF file
AIAbuzer Hussein Ibrahim
Abuzer Hussein Ibrahim
MBMohamed Ahmed Boudref
Mohamed Ahmed Boudref
LBLyes BADIS
Lyes BADIS
A network is an aggregation of nodes joined by edges, representing entities and their relationships. In social network clustering, nodes are organized into clusters according to their connectivity patterns, with the goal of community detection. The detection of community structures in networks is essential. However, existing techniques for community detection have not yet utilized the potential of the Fish School Search (FSS) algorithm and modularity principles. We have proposed a novel method, clustering with the Fish School Search algorithm and modularity function (FSC), that enhances modularity in network clustering by iteratively partitioning the network and optimizing the modularity function using the Fish School Search Algorithm. This approach facilitates the discovery of highly modular community structures, improving the resolution and effectiveness of network clustering. We tested FSC on well-known and unknown network structures. Also, we tested it on a network generated using the LFR model to test its performance on networks with different community structures. Our methodology demonstrates strong performance in identifying community structures, indicating its effectiveness in capturing cohesive communities and accurately identifying actual community structures.

1 June 2019 in Applied Mechanics and Materials
Verified
Journal Article
Closed access
PDF file
MAMourad Amad
Mourad Amad
ABAbdelmalek Boudries
Abdelmalek Boudries
LBLyes BADIS
Lyes BADIS
Application Layer Multicast (ALM) is considered as an attractive approach for implementing wide area multicast services. In ALM, multicast functionality is implemented at the edge instead of the core network (routers). As opposed to network-layer multicast, application layer multicast requires no infrastructure support and can be easily deployed in the Internet. In this paper, we propose a new efficient and scalable model for optimizing application layer multicast using HPM architecture (HPM: A novel hierarchical Peer-to-Peer model for lookup acceleration with provision of physical proximity). This approach benefits from P2P properties and characteristics. In this contribution, we consider our optimized tree construction algorithm simultaneously for each ring of HPM. The global tree construction algorithm is composed of two steps. In the first step, we construct a sub-tree for each ring; the second step is to build a global tree using sub sets of adjacent rings in HPM architecture. The proposed model inherits from main P2P attributes such as: scalability, fault tolerance characterized HPM. Preliminarily performance evaluations show that results are globally satisfactory, the depth of the resulting multicast tree is optimized.


10 October 2026 in Electronics
Verified
Journal Article
Open access
PDF file
SBSarra Bouchareb
Sarra Bouchareb
AAAkli Abbas
Akli Abbas
LBLyes BADIS
Lyes BADIS
Deepfake detectors are rarely evaluated on Arabic-media content. We introduce Arab-DF, an Arabic-media visual deepfake dataset of 1066 authentic and 11,240 manipulated videos covering four manipulation families. Thirteen public detector checkpoints are evaluated zero-shot under a common protocol, and each must first reproduce its DeepfakeBench Celeb-DF-v2 frame-level AUC within ±0.02; eight meet this criterion. On the fixed 824-video subset, their mean video-level AUC falls from 0.820 on Celeb-DF-v2 to 0.552 on Arab-DF (ΔAUC = −0.267), and five have bootstrap 95% confidence intervals above 0.5. When the authentic sample is enlarged from 75 to 392 while retaining 749 manipulated videos, the mean is 0.508 (95% CI [0.478, 0.537]); only EfficientNet-B4 retains a 95% confidence interval entirely above 0.5. Lip synchronization has the lowest family-level point estimate, although intervals overlap. Seven file-level variables separate the classes (out-of-fold AUC 0.895), yet normalization on a 94-video subset leaves mean detector AUC essentially unchanged (ΔAUC = −0.0008; 95% CI [−0.054, 0.042]). Effort reaches 0.695 on the fixed subset but fails the control and is reported separately. Overall, reproducible public checkpoints transfer poorly to this fixed Arabic-media evaluation domain.