Title: Building Data-Centric AI and Signal Processing Applications with MATLAB & Simulink for Industrial Innovation

Date/Time: 16:10-17:40, Tuesday, 16 April 2024
Location: Auditorium, 3F 

Abstract:
In the rapidly evolving landscape of artificial intelligence (AI), certain fields such as autonomous driving, speech recognition, and machine translation have seen exponential growth due to the availability of large datasets and extensive research. Improvements in these areas are driven primarily by investments in ever more complex machine learning and deep learning models. However, the scenario is markedly different in the realm of industrial signal processing applications, where challenges around scarcity and noise levels of available data, rarity of tailored AI models, and gaps in AI expertise abound.

This workshop delves into the transformative potential of data-centric AI workflows, particularly when combined with domain-specific signal processing knowledge. Using code-based examples we will highlight how such approaches can not only enhance model performance but also pave the way for the broader adoption of AI in real-world industrial applications. Attendees will gain insights into:

• strategies for improving dataset quality, labelling, data augmentation and data synthesis.
• methods for enhancing model performance through preprocessing and feature extraction.
• development of custom deep learning architectures for signal processing within MATLAB.
• leveraging state-of-the-art AI models and transfer learning for signal processing.
• utilizing cloud-based and local GPUs for speeding up feature extraction and training.
• generating embedded prototypes that integrate signal processing and deep learning.

Through an array of practical examples with MATLAB & Simulink, this presentation aims to equip professionals with the tools and knowledge needed to harness data-centric AI for signal processing, ensuring they are well-positioned to tackle the unique challenges of industrial applications.

Some familiarity with MATLAB & Simulink will be desirable but not strictly required.

Speaker Bio:
Praful Pai, PhD
Manager, ECE, Signal Processing and Communications | MathWorks

Praful Pai is a Manager, Electrical and Computer Engineering – Signal Processing and Communications at MathWorks. He engages with researchers at universities worldwide to accelerate research and development in signal processing and wireless communications with MATLAB & Simulink. He received his Ph.D. from IIT Kharagpur, India, and worked extensively in multi-disciplinary teams across domains focusing on mathematical modeling, instrumentation, signal and image processing, as well as machine learning prior to joining MathWorks. He is also passionate about making engineering education engaging and collaborates with educators worldwide to utilize computational thinking and model-based design in teaching courses and labs.