arXiv subjects

Jerry Lacmou Zeutouo

Publications and source records attributed to Jerry Lacmou Zeutouo.

3 recordsLinked to original sources

arXiv2026

MIFR: A Modality-Invariant and Fair Representation Framework for Skin Disease Classification

Skin diseases represent a major global public health burden, yet machine learning tools developed to assist in their diagnosis suffer from two critical limitations: reliance on only one modality for diagnosis and systematic performance disparities across skin tones. While existing approaches address each challenge separately, this work proposes a modality-invariant framework with fair representation (MIFR) for skin disease classification. The architecture pairs clinical photographs with dermoscopic images using ViT-based encoders, projecting each input into a high-dimensional embedding space via modality-specific projection heads. The resulting model is trained with a five-component multi-objective loss including weighted cross-entropy for classification, confusion and skin-type classification losses for fairness, per-modality supervised contrastive loss for class alignment, and a modality-invariance loss for clinical and dermoscopic modality alignment. Experiments on the HIBA+Derm7pt paired dataset and the external PAD-UFES-20 and ISIC 2019 datasets showed that modality-invariant representation learning provides competitive predictive performance compare to relevant baseline models and competitive fairness on the internal dataset. t-SNE visualizations confirmed that clinical and dermoscopic embeddings of the same disease are geometrically aligned, validating the joint objectives.

cs.CV↗

arXiv2025

Transformer-Based Model for Cold Start Mitigation in FaaS Architecture

Serverless architectures, particularly the Function as a Service (FaaS) model, have become a cornerstone of modern cloud computing due to their ability to simplify resource management and enhance application deployment agility. However, a significant challenge remains: the cold start problem. This phenomenon occurs when an idle FaaS function is invoked, requiring a full initialization process, which increases latency and degrades user experience. Existing solutions for cold start mitigation are limited in terms of invocation pattern generalization and implementation complexity. In this study, we propose an innovative approach leveraging Transformer models to mitigate the impact of cold starts in FaaS architectures. Our solution excels in accurately modeling function initialization delays and optimizing serverless system performance. Experimental evaluation using a public dataset provided by Azure demonstrates a significant reduction in cold start times, reaching up to 79\% compared to conventional methods.

cs.DC↗

arXiv2022

Four-splitting based coarse-grained multicomputer parallel algorithm for the optimal binary search tree problem

This paper presents a parallel solution based on the coarse-grained multicomputer (CGM) model using the four-splitting technique to solve the optimal binary search tree problem. The well-known sequential algorithm of Knuth solves this problem in $\mathcal{O}\left(n^2\right)$ time and space, where $n$ is the number of keys used to build the optimal binary search tree. To parallelize this algorithm on the CGM model, the irregular partitioning technique, consisting in subdividing the dependency graph into subgraphs (or blocks) of variable size, has been proposed to tackle the trade-off of minimizing the number of communication rounds and balancing the load of processors. This technique however induces a high latency time of processors (which accounts for most of the global communication time) because varying the blocks' sizes does not enable them to start evaluating some blocks as soon as the data they need are available. The four-splitting technique proposed in this paper solves this shortcoming by evaluating a block as a sequence of computation and communication steps of four subblocks. This CGM-based parallel solution requires $\mathcal{O}\left(n^2/\sqrt{p} \right)$ execution time with $\mathcal{O}\left( k \sqrt{p}\right)$ communication rounds, where $p$ is the number of processors and $k$ is the number of times the size of blocks is subdivided. An experimental study conducted to evaluate the performance of this CGM-based parallel solution showed that compared to the solution based on the irregular partitioning technique where the speedup factor is up to $\times$10.39 on one hundred and twenty-eight processors with 40960 keys when $k = 2$, the speedup factor of this solution is up to $\times$13.12 and rises up to $\times$14.93 when $k = 5$.

cs.DC↗