The Gradient’s Post

Completed Udemy’s Self-Driving Car Course! Learned practical deep learning skills for autonomous driving—excited to keep building. #DeepLearning #AI

Our very own Steve Lowry was interviewed in the latest article by Sharp Magazine 'Welcome to the Future'! Dive into the discussion about the transformative power of AI and the innovative trends shaping our future. Steve shares insights on how AI is seamlessly integrating into our daily lives. As he puts it, “The new DJ feature on Spotify is simple yet quite helpful and fun. It’s one less thing to think about — a good jam to put you in the right mood.” He also highlights the future of AI in driving, noting, “In perhaps five years, autonomous driving will start to become commonplace... In the long run, computer drivers will be far safer than human drivers.” Discover how AI is making life more effortless and enjoyable. Check out the full article here: https://lnkd.in/g_HpbXms

🚗 Waymo’s EMMA new research show how promising some of the Tesla and Wayve approaches could be in the long term however how does it compare to Tesla’s Vision-Centric Approach, and Mobileye’s CAIS? 🔹 Waymo EMMA: Redefining Multimodal Learning Waymo’s EMMA, powered by Google’s Gemini AI, combines sensor data and textual reasoning to interpret dynamic road scenarios. By using multimodal inputs like video, radar, and textual information, it’s improving accuracy in tasks like object detection and trajectory planning. This could set a new benchmark for safer, adaptable autonomous systems. https://lnkd.in/e-nJ32UR 🔹 Tesla’s Vision-Centric End-to-End Approach Tesla focuses solely on camera-based systems, relying on massive datasets to train its neural networks. While this simplifies system architecture, it can struggle with depth perception and the absence of sensor fusion. Tesla’s strategy is bold but faces challenges in scalability and robustness under diverse real-world conditions. 🔹 Mobileye’s CAIS: A Balanced Hybrid Model Mobileye’s Centralized Autonomous Intelligence System (CAIS) combines radar and camera data within a modular framework. This approach emphasizes reliability through redundancy, making it ideal for safety-critical decisions. Mobileye’s Q3 2024 focus on EyeQ6-based products reflects its vision for scalable and adaptable AI. 🌟 What’s Next? Waymo’s EMMA and Mobileye’s CAIS highlight a growing trend towards multimodal and sensor-fusion technologies, while Tesla’s vision-first approach pushes simplicity and scalability. With Waymo already operating in cities, it could be difficult for other companies like Wayve to be able to match the data gathering and training that Waymo will have. 💬 What do you think? Can Tesla’s pure vision-based strategy compete with Waymo’s multimodal AI or Mobileye’s hybrid approach? Is the new EMMA research a significant threat for Wayve? #Waymo #Tesla #Mobileye #Wayve #AutonomousDriving #AI #MachineLearning #Innovation

I'm excited to announce the publication of my latest paper, co-authored with Vincenzo Riccio and Paolo Tonella, titled "Focused Test Generation for Autonomous Driving Systems," in ACM Transactions on Software Engineering and Methodology (TOSEM). In this work, we introduce DeepAtash-LR, a novel approach that significantly enhances the testing process of Autonomous Driving Systems (ADSs). By integrating a surrogate model, DeepAtash-LR efficiently generates targeted and diverse tests, addressing the critical need for reliable ADS performance in complex environments. Our experiments demonstrate that DeepAtash-LR can produce up to 60 times more failure-inducing inputs compared to baseline methods, greatly improving the quality of ADS through fine-tuning. Check out the full paper and the replication package for more details: [Paper Link](https://lnkd.in/ex_XXusw) | [Replication Package](https://lnkd.in/e835ETmV). #AutonomousDriving #MachineLearning #DeepLearning #AI #Testing #Research #ADS #ACMTOSEM #SoftwareEngineering

Wayve Secures Over $1 Billion Investment Led by SoftBank to Pioneer Embodied AI Products for Automated Driving In a groundbreaking move for the UK tech landscape, Wayve, a leading AI company, has announced a staggering $1.05 billion Series C investment round. Spearheaded by SoftBank Group and featuring contributions from new investor NVIDIA and existing investor Microsoft, this substantial funding is set to accelerate Wayve's mission to redefine autonomous mobility through embodied intelligence. 🚀 But what exactly is embodied AI, and how does Wayve plan to revolutionise the autonomous driving space? 🤖 Embodied AI represents the cutting edge of AI innovation, integrating advanced AI directly into vehicles and robots to fundamentally transform how machines interact with and learn from human behaviour in real-world environments. For Wayve, this means leveraging deep learning techniques to empower vehicles to perceive, understand, and navigate any environment autonomously. 🧠🚗 Since its founding in 2017, Wayve has been a trailblazer in the field of Embodied AI for autonomous driving. Notably, they were the first to develop and test an end-to-end (e2e) AI autonomous driving system on public roads. This pioneering effort has positioned Wayve as a market leader, with their technology serving as the foundation for what they describe as a 'GPT for driving'—a comprehensive AI model capable of enabling any vehicle to see, think, and drive through diverse environments. 🛣️🌟 With this latest injection of funding, Wayve is poised to fully develop and launch the first Embodied AI products for production vehicles. These products will allow original equipment manufacturers (OEMs) to efficiently upgrade cars to higher levels of driving automation, from L2+ assisted driving to L4 automated driving, as Wayve's core AI model advances. 🚗💡 As someone entrenched in the IT hardware industry, the convergence of advanced AI algorithms with cutting-edge hardware technologies holds particular significance. The implications for industries and businesses like ours are profound, promising transformative innovation and enhanced safety standards. 💼💻 https://wayve.ai/ #Wayve #SoftBank #NVIDIA #Microsoft #AI #AutonomousDriving #EmbodiedAI #SeriesC #TechNews #UKTech

To create high-fidelity simulations, you need data. To gather that data, you have to drive in the real world. To pressure-test your real-world driving, you need to have simulations. True autonomous driving isn’t about a single approach—it’s about combining all of these elements. By combining data from multiple domains, we learn more, faster, and can drive more places than any other player in the industry. Kodiak #selfdriving #AI

Wayve Introduces the First Vision-Language-Action Model Tested on Public Roads Wayve introduces LINGO-2, the first vision-language-action model for autonomous driving, enhancing AI explainability and human-machine interaction on roads. Read more here: https://lnkd.in/eWXt72pe

Exciting developments in autonomous technology! Learn how gen AI is shaping the future of autonomous driving in our recent blog post.

From multimodal models to closed-loop simulations, generative AI is the key to robust autonomous driving systems. Learn about Applied Intuition's approach in our latest blog post.