Asma Atamna's Home Page
About Me
I'm a postdoctoral researcher in the group of Prof. Tobias Glasmachers at the Ruhr-University Bochum. My research focuses on facilitating machine learning (ML) use by non-ML experts.
Education
I received my Ph.D. in Computer Science from University of Paris-Saclay (formerly Paris-Sud 11) in January 2017, under the supervision of Nikolaus Hansen and Anne Auger. During my Ph.D. thesis, I studied and designed various Evolution Strategies—a class of adaptive stochastic algorithms for black-box continuous optimization—for constrained optimization in particular.
Before my Ph.D., I completed a Master's degree in Artificial Intelligence at University of Paris-Sud 11 (2013) and an engineer's degree in Computer Science at the National School of Computer Science, Algiers (2011).
Research Interests
Deep learning, reinforcement learning, AutoML, neural architecture design
Real-world machine learning applications
Continuous optimization, derivative-free optimization, constrained optimization
Evolutionary algorithms, convergence analysis, Markov chains
Benchmarking
Past Experience
Postdoctoral researcher at the LTCI, Télécom Paris. Worked on user engagement analysis in Human-Robot Interaction with Deep Learning approaches, in particular Recurrent Neural Networks (RNNs).
Postdoctoral researcher at the ICMPE (CNRS—UPEC). Worked on the generation of metal hydrides for hydrogen storage using Machine Learning approaches, in particular Generative Models (GANs), as well as on Graph Neural Networks (GNNs).
Postdoctoral researcher at the CMAP, RandOpt team. Worked on benchmarking algorithms for black-box (derivative-free) continuous constrained optimization.
Contact
E-mail: asma [dot] atamna [at] ini [dot] rub [dot] de
Publications
Conference Papers
A. Pendyala, J. Dettmer, T. Glasmachers, A. Atamna (2023). ContainerGym: A Real-World Reinforcement Learning Benchmark for Resource Allocation. In Conference on Machine Learning, Optimization, and Data Science (LOD). [PDF]
A. Atamna, C. Clavel (2020). HRI-RNN: A User-Robot Dynamics-Oriented RNN for Engagement Decrease Detection. In INTERSPEECH 2020. [PDF]
A. Atamna, N. Sokolovska, J.-C. Crivello (2020). A Principled Approach to Analyze Expressiveness and Accuracy of Graph Neural Networks. In Symposium on Intelligent Data Analysis (IDA). [PDF]
A. Atamna, A. Auger, N. Hansen (2017). Linearly Convergent Evolution Strategies via Augmented Lagrangian Constraint Handling. In Proceedings of the 14th ACM/SIGEVO Conference on Foundations of Genetic Algorithms (FOGA), pp. 149—161. [PDF]
A. Atamna, A. Auger, N. Hansen (2016). Augmented Lagrangian Constraint Handling for CMA-ES—Case of a Single Linear Constraint. In Parallel Problem Solving from Nature (PPSN), pp. 181—191. [PDF]
A. Atamna, A. Auger, N. Hansen (2016). Analysis of Linear Convergence of a (1+1)-ES with Augmented Lagrangian Constraint Handling. In Genetic and Evolutionary Computation Conference (GECCO), pp. 213—220. Nominated for Best Paper Award. [PDF]
N. Hansen, A. Atamna, A. Auger (2014). How to Assess Step-Size Adaptive Mechanisms in Randomised Search. In Parallel Problem Solving from Nature (PPSN), pp. 60—69. [PDF]
Journal Papers
R. Akrour, A. Atamna, J. Peters (2021). Convex Optimization with an Interpolation-based Projection and its Application to Deep Learning. In Machine Leaning (MACH). [PDF]
A. Atamna, A. Auger, N. Hansen (2018). On Invariance and Linear Convergence of Evolution Strategies with Augmented Lagrangian Constraint Handling. In Theoretical Computer Science (TCS). [PDF]
Workshop Papers
P. Dufossé, A. Atamna (2022). Benchmarking several strategies to update the penalty parameters in AL-CMA-ES on the bbob-constrained testbed. In Proceedings of the Genetic and Evolutionary Computation Conference (GECCO) Companion, pp. 1691–1699. [PDF]
A. Atamna (2015). Benchmarking IPOP-CMA-ES-TPA and IPOP-CMA-ES-MSR on the BBOB Noiseless Testbed. In Proceedings of the Genetic and Evolutionary Computation Conference (GECCO) Companion, pp. 1135—1142. [PDF]
Preprints
A. Atamna, N. Sokolovska, J.-C. Crivello (2019). SPI-GCN: A Simple Permutation-Invariant Graph Convolutional Network. [PDF]
Thesis
A. Atamna (2017). Analysis of Randomized Adaptive Algorithms for Black-Box Continuous Constrained Optimization (title in French: Analyse d'algorithmes stochastiques adaptatifs pour l'optimisation numérique boîte-noire avec contraintes). University of Paris-Saclay, University of Paris-Sud, France. [PDF]
Teaching
I worked as a teaching assistant at University of Paris-Saclay during my Ph.D. thesis. I gave tutorials and practical courses to first and second year computer science undergraduate students (192 hours in total). Below is the list of my teaching activities.
Algorithmics tutorial. 1st year students (2013—2016)
Advanced algorithmics tutorial. 2nd year students (2014—2016)
C++ practical course. 1st year students (2014—2016)
OOP practical course in Java. 1st year students (2013—2014)
Supervision of a robotics project. 2nd year students (2013—2014)