Machine Learning without Coding

Introduction

Have you ever wondered about the possibilities of Machine Learning without Coding? Step into the world of innovation with RAVSim, a groundbreaking tool designed for efficient and flexible computations in resource-intensive tasks, especially computer vision! This tool explored the domain of neuromorphic intelligence, using analog logic to simulate brain-like capabilities. Imagine simplifying model training and testing, making machine learning accessible even to those with no programming skills.

Let’s Explore How It’s Possible

Can you imagine a tool that introduces a quick image classification method based on spiking neural networks (SNNs) without the need for additional coding? RAVSim (Runtime Analysis and Visualization Simulator) does just that, inviting you to explore how it enables real-time interaction, visualization, and analysis of SNNs while addressing challenges in their modeling. Not only that, but it also supports various RGB picture dataset preprocessing, which means from raw data to preprocessed datasets, all can be done by using the RAVSim tool. RAVSim also offers to verify which mathematical models are better for designing SNNs and we identify the LIF model as optimal for real-time multi-object applications.

Now, picture the application based on the LIF model in a real-time scenario, combining neuromorphic computing with event-based cameras for human detection, that is what we are designing. This multi-core architecture-based application combines object detection, transfer learning with SNNs, human recognition, location and tracking, feature extraction, and real-time analysis. Let’s think about how this powerful architecture ensures efficient real-time operation, making it suitable for applications like autonomous driving and robotic systems.

Conclusion

The world of Machine Learning without Coding is now at your fingertips with RAVSim. Have you ever considered creating a user-friendly interface for people with little programming expe- rience? Consider using spiking neural networks and different neural models to do efficient and flexible analysis in resource-intensive activities such as computer vision, then experience the unlimited possibilities that RAVSim v1.1 provides to the world of technology and innovation, and explore the future of machine learning without the coding effort.

Additional resources

TBA

Cooperation

Industrial Internet of Things Group

Prof. Dr.-Ing. habil. Thorsten Jungeblut

PhD student: Sanaullah

Cognitronics and Sensor Systems Group

Prof. Dr.-Ing. Ulrich Rückert

PhD student: Shamini Koravuna

Project Publications

  • Koravuna, Shamini, Sanaullah, Thorsten Jungeblut, and Ulrich Rückert (2023). ‘‘”Digit Recognition Using Spiking Neural Networks on FPGA”’’. In: Interna- tional Work-Conference on Artificial Neural Networks (IWANN) Conference. Cham: Springer International Publishing, pp. 406–417.
  • Pennino, Federico, Shamini Koravuna, Christoph Ostrau, and Ulrich Rückert (2022). ‘‘N-MNIST object recognition with Spiking Neural Networks’’. In: Dataninja Spring School, Poster Presentation.
  • Sanaullah, Amanullah, Kaushik Roy, Jeong-A Lee, Son Chul-Jun, and Thorsten Jungeblut (2023). ‘‘A Hybrid Spiking-Convolutional Neural Network Ap- proach for Advancing High-Quality Image Inpainting’’. In: IEEE Conference on Computer Vision (ICCV) at Workshop PerDream.
  • Sanaullah and Thorsten Jungeblut (2023). ‘‘”Analysis of MR Images for Early and Accurate Detection of Brain Tumor Using Resource Efficient Simulator Brain Analysis”’’. In: International Conference on Machine Learning and Data Mining (MLDM) Conference.
  • Sanaullah, Shamini Koravuna, Ulrich Rückert, and Thorsten Jungeblut (2022a). ‘‘”Real-Time Resource Efficient Simulator for SNNs-based Model Experimentation”’’. In: Dataninja Spring School Poster Presentation.
  • Sanaullah, Shamini Koravuna, Ulrich Rückert, and Thorsten Jungeblut (2022b). ‘‘”SNNs Model Analyzing and Visualizing Experimentation Using RAVSim”’’. In: Engineering Applications of Neural Networks. Cham: Springer International Publishing, pp. 40–51.
  • Sanaullah, Shamini Koravuna, Ulrich Rückert, and Thorsten Jungeblut (2022c). ‘‘Real-time resource efficient simulator for snns-based model experimentation’’. In: Dataninja Spring School, Poster Presentation.
  • Sanaullah, Shamini Koravuna, Ulrich Rückert, and Thorsten Jungeblut (2022d). ‘‘SNNs Model Analyzing and Visualizing Experimentation Using RAVSim’’. In: Engineering Applications of Neural Networks. (EANN). Vol. 161. Springer, pp. 40–51.
  • Sanaullah, Shamini Koravuna, Ulrich Rückert, and Thorsten Jungeblut (2023a). ‘‘”Design-Space Exploration of SNN Models using Application-Specific Multi- Core Architectures”’’. In: Neuro-Inspired Computing Elements (NICE) Con- ference.
  • Sanaullah, Shamini Koravuna, Ulrich Rückert, and Thorsten Jungeblut (2023b). ‘‘”Evaluating Spiking Neural Network Models: A Comparative Performance Analysis”’’. In: Dataninja Spring School Poster Presentation.
  • Sanaullah, Shamini Koravuna, Ulrich Rückert, and Thorsten Jungeblut (2023c). ‘‘”Streamlined Training of GCN for Node Classification with Automatic Loss Function and Optimizer Selection”’’. In: Engineering Applications of Neural Networks (EANN) Conference. Cham: Springer International Publishing, pp. 191–202.