Self-Driving Cars Pedalled a Lie So Far?

Can't read or see images? View this email in a browser

Tesla identifies itself as an autonomous vehicle maker. But all the hype you hear around Tesla cars manoeuvring themselves is just hype – far from being truly autonomous. They are still languishing at Level 2, where the vehicle can only assist with steering or acceleration functions and allow the driver to disengage from some of their tasks. The driver must always be ready to take control of the vehicle and is still responsible for most safety-critical functions and monitoring of the environment. Meanwhile, a UK-based self-driving car company Wayve has cracked the Level 5 technology, which will require absolutely no human attention. There is no need for pedals or a steering wheel, as the autonomous vehicle system controls all critical tasks, monitoring the environment and identification of unique driving conditions, like traffic jams. In June of this year, Wayve published a technical report detailing its generative world model, GAIA-1 (Generative AI for Autonomy) which leverages video, text, and action inputs to generate realistic driving scenarios. At the core of GAIA-1 lies a world model, boasting an impressive 6.5 billion parameters. This pivotal component excels at forecasting image tokens by drawing insights from historical data. These predictions undergo a transformation into tangible images, thanks to the assistance of a video model equipped with a formidable 2.6 billion parameters. Impressively, GAIA-1, fortified by its training on a robust dataset encompassing 4,700 hours of driving experiences, harnesses a total of 9 billion parameters. This substantial computational prowess empowers GAIA-1 to generate remarkably precise and realistic visual representations. Alex Kendall, chief of Wayve, shared on X that unlike the other models out there, it is not just a simulation but a full world model that understands how the world works.  Wayve's GAIA-1 draws from their extensive research in areas such as future prediction, the concept of "dreaming" about driving scenarios, predictive modelling from a bird's eye perspective, and the development of comprehensive world models. This amalgamation of technologies positions GAIA-1 as a significant advancement in the realm of self-driving systems. Read the full story here. A Spy Shade The new Ray-Ban smart glasses have raised a big privacy concern, for these can be discreetly used to record people without their knowledge or consent. According to internet activists, the smart glasses will be a constant presence on our noses, observing our every move. Rather than solely focusing on integrating people into the metaverse, Meta appears to be interested in monitoring individuals' day-to-day activities, as mentioned during Meta Connect 2023. Why not? Meta has a history of handling people's privacy in a controversial manner over the years. Read the full story here. Eye in the Sky Bengaluru-based startup GalaxEye is working on a revolutionary satellite named the 'Drishti Mission', set to launch in collaboration with ISRO in late 2024. This pioneering micro-satellite, boasting the world's first multi-sensor capability, addresses a persistent challenge in satellite imaging – data disruption caused by clouds during multispectral or optical camera satellite imaging. GalaxEye's innovative approach combines generative AI and synthetic aperture radar (SAR) to overcome this hurdle, ensuring all-weather imaging.  The company recently secured $3.5 million in seed funding and collaborates with notable entities like ideaForge and Antaris Inc. This initiative marks a significant step forward in the space tech sector and shows the growing interest and investment in India's space tech ecosystem. Read the full story here. Robots can Do General Tasks Google DeepMind recently released resources for general-purpose robots, a significant step in robotics. This development marks a shift toward knowledge transfer between robots, eliminating the need for individualised training for specific tasks. Two notable models in this space are RT-X and RoboCat. RT-X, or the Robotics Transformer X, is a versatile model designed for general robotics research. It excels in actions, vision, and language understanding and can generalise its skills across various robots and tasks. RT-X builds upon the foundations of RT-1 and RT-2, achieving higher performance due to diverse cross-embodiment data. RoboCat is an AI agent with a remarkable ability to learn multiple tasks across different robotic systems and generate its own training data. This rapid learning capability reduces the need for extensive human-guided training. Read the full story here. TAUSIF ALAM & AMIT RAJA NAIK Monday, Oct 9, 2023 | Was this email forwarded to you? Sign up here DOWNLOAD OUR MOBILE APP Stay Connected info@analyticsindiamag.com © 2023 Analytics India Magazine

This email was sent by support@analyticsindiamag.com to alexvarboffin.abbb@blogger.com Not interested? Unsubscribe | Manage Preference | Update profile Analytics India Magazine | 280, 2nd floor, 5th Main, 15 A cross, Sector 6, HSR layout Bengaluru, Karnataka 560102