Workshop on Neural Architecture Search
Scope
Neural Architecture Search (NAS) is a powerful machine learning technique that automates the design of neural network architectures instead of manually defining the network's structure, such as layers, number of neurons, and activation functions. NAS can discover innovative architectures that surpass those designed manually, often resulting in more accurate, efficient, or straightforward models. NAS involves searching through a predefined space of potential architectures to find the most effective one for a specific task, such as image classification or natural language processing. NAS aims to optimize aspects like accuracy, efficiency, and model size, often using techniques such as:
Reinforcement Learning: Agents explore different architectures and learn to improve designs based on performance feedback.
Evolutionary Algorithms: Inspired by natural selection, these algorithms evolve architectures over successive generations.
Gradient-Based Methods: Utilize differentiable architecture representations to optimize structures using gradient descent.
The goal of NAS is to reduce the time and expertise required to design high-performing neural networks, making the process more efficient and accessible.
Topics of Interest
The workshop will explore evolutionary NAS's latest advancements and methodologies, focusing on its application across various domains such as computer vision, natural language processing, and reinforcement learning. It will cover theoretical foundations, search strategies, evaluation techniques, and practical implementations, providing participants with a comprehensive understanding of the NAS landscape. A broad range of topics will be discussed, including but not limited to:
Representation and Encoding of Neural Networks
Development of Objective Functions
Multi-Objective NAS
Bilevel NAS
Parallel and Distributed search algorithms for NAS
Assessment Methodologies
Interpretability and Explainability
Theoretical Foundations of NAS
Real-World Applications and Case Studies
Submissions
We invite submissions of the following types of papers: regular research papers (up to 6 pages) and short position papers (up to 2 pages). Accepted workshop papers are published in the conference proceedings in the workshop section, available online via IEEE, and indexed by IEEE Xplore.
Important Dates
Workshop Paper Submission Deadline: 15 January, 2025 10 March, 2025
Workshop Paper Acceptance Notification: 1 April, 2025
The submission procedure, deadlines, and paper format follow the same guidelines as the IEEE CAI'2025 main conference. Submissions must be made via the IEEE CAI'2025 online system.
Organizers
Saúl Zapotecas-Martínez
Computer Science Department
Instituto Nacional de Astrofísica Óptica y Electrónica, Tonantzintla, Puebla, Mexico
szapotecas [at] inaoep [dot] mx
Alejandro Rosales-Pérez
Centro de Investigación en Matemáticas (CIMAT)
Monterrey Campus
alejandro.rosales [at] cimat [dot] mx
Efrén Mezura-Montes
Artificial Intelligence Research Institute
University of Veracruz, MEXICO
emezura [at] uv [dot] mx