Prasannavenkatesh Balaji’s Post

New issue available from SAE International Journal of Connected and Automated Vehicles. Browse the titles, authors, and abstracts below, and find the full issue in SAE Mobilus® here: https://lnkd.in/ejsppgbk. Topics include: platooning, driver style recognition, generating virtual test scenarios, and more! There's an Open Access article available in the issue too: "Deep Learning Algorithms for Longitudinal Driving Behavior Prediction: A Comparative Analysis of Convolutional Neural Network and Long–Short-Term Memory Models," https://lnkd.in/e4ijCnsB. Editors-in-Chief: Daniel Watzenig and Terry Fruehling #connectedcars #transportation #autonomousvehicles

"𝗔𝗿𝘁𝗶𝗳𝗶𝗰𝗶𝗮𝗹" 𝗼𝗿 "𝗖𝗼𝗺𝗽𝘂𝘁𝗮𝘁𝗶𝗼𝗻𝗮𝗹" 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲? At TU Kaiserslautern there is a course called "Methods of Soft Computing". During my PhD studies, my supervisor Prof. Georg Frey handed me over the responsibility to supervise the tutorials. The topic was a bit new to me, but felt really eager to learn how Fuzzy Logic or Genetic Algorithm or Neural network can complement or supplement traditional control theory. While understanding the topic, one special argument caught my eyes where the authors argued that the "Artificial" aspect (the opposite of Humane possibly) is not that important in AI. The most important part of it is the Computational aspect. It is not simply anything from anywhere without any logical or mathematical basis. Thus, it shall be rather called "Computational Intelligence". Especially, when it is applied in deterministic applications like autonomous driving.

Today, we celebrate a monumental achievement as Geoffrey Hinton, often called the "Godfather of AI," has been awarded the Nobel Prize in Physics. His groundbreaking work, which revolutionized the field of machine learning, now finds itself recognized not just in the realm of computer science, but in physics—a testament to the far-reaching impact of his contributions. Hinton’s development of neural networks has reshaped how we understand and approach complex problems. His pioneering work in backpropagation and deep learning laid the foundation for many of today’s artificial intelligence (AI) systems, which are now integral in everything from healthcare to autonomous vehicles. It is no surprise that his brilliance is now seen as essential to physics, a field that, like AI, thrives on uncovering and predicting the unseen. In recent years, machine learning has played a transformative role in scientific research, and nowhere is this clearer than in physics. Whether it’s detecting gravitational waves, modeling quantum mechanics, or predicting particle behavior at the quantum level, AI has become an indispensable tool for physicists. Hinton’s vision of using neural networks to model complex systems has bridged the gap between abstract mathematics and real-world applications. His Nobel Prize marks a significant moment in history, not just for his well-deserved recognition, but for what it signifies. This award underscores that machine learning is no longer a niche technology—it's an essential part of scientific exploration, capable of solving some of the greatest mysteries of our universe. Congratulations, Geoffrey Hinton, on this extraordinary and well-earned achievement! Your work has forever changed the landscape of technology and science, and this Nobel Prize is a testament to the profound influence machine learning will continue to have in every aspect of our lives, including the deepest corners of physics.

Convolutional Neural Networks (CNNs) are key to solving complex computer vision tasks like image classification, object detection, and segmentation. With their unique architecture, CNNs extract intricate features from images, making them indispensable in fields like healthcare and autonomous driving 🖼️🔍 Learn more about the building blocks of CNNs in my latest article! #AI #MachineLearning #ComputerVision #LabelYourData

🚀 Thrilled to announce the completion of our professional personal project: "Visual Servoing of a Line Following Robot" My teammate, Mohamed Karim Ben Boubaker, and I, under the guidance of Mrs. Amani AYEB, have successfully designed and implemented an autonomous line-following robot as part of our Industrial IT and Automation Engineering program at the National Institute of Applied Science and Technology. 🔧 Our project integrates computer vision and deep learning techniques to enhance Lane Detection and Lane-Keeping systems. We employed various algorithms and image processing techniques such as thresholding, warping, and pixel summation to achieve these functionalities. 💡 Inspired by the ALVINN neural network architecture , which calculates the optimal steering direction using laser rangefinder and camera inputs, we aim to develop a convolutional neural network model to predict the steering angle from images ,as artificial neural networks have proven to be highly effective and flexible in dynamic contexts with noise and frequent changes #Self_driving cars OpenCV Deep Learning Computer Vision #LaneDetection 🚗 #ComputerVision 👁️ #DeepLearning 🧠 #ADAS 🚙 #AutonomousVehicles 🤖 #AI 💻 #OpenCV #EngineeringExcellence

I'm delighted to share that our recent paper, titled "Enabling Uncertainty Estimation in Iterative Neural Networks," has been accepted at #ICML2024 in Vienna! Estimating the uncertainty of model predictions is crucial for a wide range of applications across various fields, such as self-driving cars, robotics, manufacturing, medical applications, and many others. However, many contemporary methods for uncertainty estimation face significant challenges. They often produce inaccurate uncertainties, incur prohibitive computational costs, are difficult to integrate into existing workflows, or are limited to specific tasks and model architectures. These limitations make their deployment in real-world scenarios challenging, if not impossible. In this work, we introduce a novel method for estimating uncertainty in iterative neural networks. Our approach is simple, easy to implement, and task-agnostic, generating uncertainties based on the variance of predictions across multiple iterations. We demonstrate that our method provides reliable estimates comparable to state-of-the-art ensembles while also improving the models' accuracy—all with minimal additional computational cost. Come see our poster on July 25th in Vienna! 𝗽𝗿𝗼𝗷𝗲𝗰𝘁 𝗽𝗮𝗴𝗲: https://lnkd.in/djab4-JV 𝗮𝗿𝗫𝗶𝘃: https://lnkd.in/dMqS4jEg 𝗰𝗼𝗱𝗲: https://lnkd.in/d3fEAyyA 𝗽𝗼𝘀𝘁𝗲𝗿: https://lnkd.in/dgUkvQhV

🌐 Predictive Maintenance AI Certification I'm proud to have acquired the Applications of AI for Predictive Maintenance certification from NVIDIA in collaboration with NOVATION CITY. 🚀 This specialist course focuses on advanced approaches, such as  📊 GPU XGBoost Models with RAPIDS for time-series analysis. 🤖 LSTM Models with Keras and TensorFlow for predictive modeling. 🔍 Autoencoders for anomaly detection and proactive maintenance. I am grateful for the chance to expand my knowledge and am prepared to use these AI-driven solutions to improve operational efficiency and dependability. #AI #PredictiveMaintenance #NVIDIA #XGBoost #TensorFlow #AnomalyDetection #DataScience #Innovation #SmartMaintenance

Dear All, This is a gentle reminder about tomorrow's talk by Prof. Marta D'Elia from Stanford University, USA, as part of the Param-Intelligence (π) Seminar Series. The seminar will take place on December 5, 2024, from 12:00 to 1:00 PM ET. Please mark your calendars and join us for this session. The Zoom link is provided below. Details of the talk are as follows: Title: On the use of Graph and Point networks in scientific applications Abstract: In the context of scientific and industrial applications, one often has to deal with unstructured space-time data obtained from numerical simulations. The data can be either in the form of a mesh or a point cloud. In this context, graph neural networks (GNNs) have proved to be effective tools to reproduce the behavior of simulated data; however, depending on the physical nature of the datasets, variations of vanilla GNNs have to be considered to ensure accurate results. Furthermore, when only a point cloud is available, one can also consider a graph-free approach by building a "point network" that doesn't require connectivity information. In this presentation we focus on particle-accelerator simulations; a computationally demanding class of problems for which rapid design and real-time control are challenging. We propose a machine learning-based surrogate model that leverages both graph and point networks to predict particle-accelerator behavior across different machine settings. Our model is trained on high-fidelity simulations of electron beam acceleration, capturing complex, nonlinear interactions among macroparticles distributed across several initial state dimensions and machine parameters. Our initial results show the model’s capacity for accurate, one-shot tracking of electron beams at downstream observation points, outperforming baseline graph convolutional networks. This framework accommodates key symmetries inherent in particle distributions, enhancing stability and interpretability. Speaker's Biography: Marta D'Elia is the Director of AI and ModSim at Atomic Machines and an Adjunct Professor at Stanford University, ICME. She previously worked at Pasteur Labs, Meta, and Sandia National Laboratories as a Principal Scientist and Tech Lead. She holds a PhD in Applied Mathematics and master's and bachelor's degrees in Mathematical Engineering. Her work deals with design and analysis of machine-learning models and optimal design and control for complex industrial applications. She is an expert in nonlocal modeling and simulation, optimal control, and scientific machine learning. She is an Associate Editor of SIAM and Nature journals, a member of the SIAM industry committee, the Vice Chair of the SIAM Northern California section, and a member of the NVIDIA advisory board for scientific machine learning. #SciML #DeepLearning Zoom link is available here:

5 Emerging AI Technologies that Future of Machine Learning 1. Advanced Neural Architectures and Generative Models 2. Quantum Machine Learning 3. Neuromorphic Computing 4. Federated Learning and Privacy-Preserving ML 5. Automated Machine Learning (AutoML) and Neural Architecture Search

Dear All, A gentle reminder about today's talk by Dr. Chris Rackauckas from Massachusetts Institute of Technology as part of our Param-Intelligence (π) Seminar Series at 12 p.m. ET. The Zoom link is provided below. Details of the talk are as follows: Title:  Enabling Industrially-Robust AI for Engineering through Scientific Machine Learning Abstract: The combination of scientific models into deep learning structures, commonly referred to as scientific machine learning (SciML), has made great strides in the last few years in incorporating models such as ODEs and PDEs into deep learning through differentiable simulation. Such SciML methods have been gaining steam due to accelerating the development of high-fidelity models for improving industrial simulation and design. However, many of the methods from the machine learning world lack the robustness required for scaling to industrial tasks. What needs to change about AI in order to allow for methods which can guarantee accuracy and quantify uncertainty? In this talk we will go through the details of how one can enable robustness in building and training SciML models. Numerical robustness of algorithms for handling neural networks with stiff dynamics, continuous machine learning methods with certifiably globally-optimal training, alternative loss functions to mitigating local minima, integration of Bayesian estimation with model discovery, and tools for validating the correctness of surrogate models will be discussed to demonstrate a next generation of SciML methods for industrial use. Demonstrations of these methods in applications such as two-phase flow HVAC systems, modeling of sensors in Formula One cars, and lithium-ion battery packs will be used to showcase the improved robustness of these approaches over standard (scientific) machine learning. Speaker's Bio.: Dr. Chris Rackauckas is the VP of Modeling and Simulation at JuliaHub, the Director of Scientific Research at Pumas-AI, Co-PI of the Julia Lab at MIT, and the lead developer of the SciML Open Source Software Organization. For his work in mechanistic machine learning, his work is credited for the 15,000x acceleration of NASA Launch Services simulations and recently demonstrated a 60x-570x acceleration over Modelica tools in HVAC simulation, earning Chris the US Air Force Artificial Intelligence Accelerator Scientific Excellence Award. He is the lead developer of the Pumas project and has received a top presentation award at every ACoP in the last 3 years for improving methods for uncertainty quantification, automated GPU acceleration of nonlinear mixed effects modeling (NLME), and machine learning assisted construction of NLME models with DeepNLME. For these achievements, Chris received the Emerging Scientist award from ISoP. Zoom Link is available here: https://lnkd.in/dzdRiFeX All are welcome!