Exploring the Connection Between AI and Robotics

Robotic AI’s reliance on vision is limiting its ability to interact with the physical world accurately. Vision systems dominate robotic AI because they’re cost-effective and can collect massive datasets. But this overemphasis on vision overlooks the critical role of force sensing—providing tactile data that vision simply can’t replicate. Without it, robots are limited to estimating force feedback from visuals, leading to inefficiencies in delicate tasks like assembly, gripping, or threading. As Edward Adelson, professor at Massachusetts Institute of Technology, explained in his TED Talk, “Force feedback allows robots to perform tactile tasks that vision alone cannot achieve—like folding a towel or threading a cable—by feeling their way through interactions, just as humans do.” Adelson’s work on GelSight technology highlights how tactile sensing can unlock superhuman precision for robots, enabling them to understand their environment through touch. The challenge? Force sensors are an added cost, generate less data, and are harder to integrate. But they offer essential benefits: • Reliability and Safety: For tasks where mistakes aren’t an option, force feedback provides the assurance vision alone cannot. • Deeper Learning: Force sensing enriches AI by adding layers of contact-based data for more robust decision-making. • Expanding Applications: From industrial automation to medical robotics, tactile data opens doors to tasks beyond vision’s reach. ATI Industrial Automation supports robotics through robust, precise robotic force sensors—helping to bring accuracy to robotic AI data collection. Edward Adelson’s TED Talk: https://lnkd.in/epeCvwqj #robotics

I had the opportunity to share my MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) and Liquid AI teams’ work on Liquid Networks with Communications of the ACM magazine. I described how we are working to bridge the gap between AI and robotics by developing Physical AI — intelligent systems that can understand text, images, and video, and apply that understanding to real-world machines like robots, sensors, power grids, cars, and more. At the core of this effort is a new class of physics-based neural networks, inspired by the simple but remarkably capable nervous system of C. elegans, a worm with just 302 neurons. These models bring intelligence closer to the physical world, enabling smarter, more responsive machines. https://lnkd.in/e3epwKPf

Stanford University researchers released a new AI report, partnering with the likes of Accenture, McKinsey & Company, OpenAI, and others, highlighting technical breakthroughs, trends, and market opportunities with large language models (LLMs).  Since the report is 500+ pages!!! (link in comments), sharing a handful of the insights below: 1. Rise of Multimodal AI: We're moving beyond text-only models. AI systems are becoming increasingly adept at handling diverse data types, including images, audio, and video, alongside text. This opens up possibilities for apps in areas like robotics, healthcare, and creative industries. Imagine AI systems that can understand and generate realistic 3D environments or diagnose diseases from medical scans. 2. AI for Scientific Discovery: AI is transforming scientific research. Models like GNoME are accelerating materials discovery, while others are tackling complex challenges in drug development. Expect AI to play a growing role in scientific breakthroughs, leading to new materials and more effective medicines. 3. AI and Robotics Synergy: The combination of AI and robotics is giving rise to a new generation of intelligent robots. Models like PaLM-E are enabling robots to understand and respond to complex commands, learn from their environment, and perform tasks with greater dexterity. Expect to see AI-powered robots playing a larger role in manufacturing, logistics, healthcare, and our homes. 4. AI for Personalized Experiences: AI is enabling hyper-personalization in areas like education, healthcare, and entertainment. Imagine educational platforms that adapt to your learning style, healthcare systems that provide personalized treatment plans, and entertainment experiences that cater to your unique preferences. 5. Democratization of AI: Open-source models (e.g., Llama 3 just released) and platforms like Hugging Face are empowering a wider range of developers and researchers to build and experiment with AI. This democratization of AI will foster greater innovation and lead to a more diverse range of applications.