βπ» I am passionate about data science and making data valuable by translating it into actionable insights.
π¨π»βπ» During my professional experiences, I had the opportunity to work on several data-driven projects using cutting edge technologies on various types of data (building predictive/clustering models, creating data visualisation reports, storage and processing of data, etc.).
π¨π»βπ» I participated in the different phases of an IT project life cycle (from design to delivery of the product and support for the client in the realisation of improvements and the correction of incidents).
π I am dynamic, ambitious and passionate about the use of data and new technologies. Technical expertise is an area that attracts me, especially in its aspects of quality and reliability of the delivered products. Currently, I am looking for new technical challenges. I am happy every day to use my knowledge for the benefit of various data projects.
Research initiation project (TER) :
Deep learning methods have made significant progress in several areas such as object recognition in images, signal analysis, automated language processing, and so on. Despite their predictive power, deep neural networks are considered as black boxes, making their interpretation difficult. Over the past 20 years, many authors have proposed techniques for extracting rules from a neural network. The objective of this project is to study the different existing methods and to implement an appropriate method for rules extraction from a deep neural network.
Learned features versus engineered features for multi-concepts detection in images.
In addition to the standard low-level descriptors used in image indexing systems, other types of high-level features have emerged and yielded interesting results. This kind of descriptors are extracted either through deep learning approaches or by using detection scores of a set of semantics. As part of this work, a comparative study of the three types of features is carried out in the context of multi concepts detection in images. Experiments on the international standard corpus βPascal VOC 2012β are conducted for concepts pairs and triplets of concepts detection.
Design and realization of an application of management of the computer park for Sonelgaz.
Activities and Societies:
Activities and Societies:
In this work, we propose two approaches that consider the semantic context for multi-concepts detection in still images. We tested and evaluated our proposal on the international standard corpus Pascal VOC for the detection of concepts pairs and triplets of concepts. Our contributions have shown that context is useful and improves multi-concepts detection in images. The combination of the use of semantic context and deep learning-based features yielded much better results than those of the state of the art. This difference in performance is estimated by a relative gain on mean average precision reaching + 70% for concepts pairs and + 34% for the case of triplets of concepts.
View Project