Source link : https://tech365.info/breakdown-of-the-fastdrivevla-ai-led-l4-autonomous-driving-from-xpeng-peking-college/
Assist CleanTechnica’s work via a Substack subscription or on Stripe.
This week, XPENG, in collaboration with Peking College, introduced a major leap ahead on this area with the acceptance of their newest analysis paper at AAAI 2026, one of many world’s premier synthetic intelligence conferences. Titled “FastDriveVLA: Efficient End-to-End Driving via Plug-and-Play Reconstruction-based Token Pruning,” the analysis particulars a novel framework that drastically reduces the computational load of onboard AI, bringing the {industry} one step nearer to viable, scalable Stage 4 autonomy.
CleanTechnica secured a duplicate of the paper and used the summary to develop this report. (The press launch is right here.)
As context, XPENG’s 2025 Roadmap had targeted on the event of L4 autonomy early on. The breakthrough arrives at a pivotal second for the Chinese language expertise firm, serving as a power multiplier for the corporate’s upcoming {hardware} and architectural shifts scheduled for 2025.
Simply in November at XPENG’s Tech Day, the VLA 2.0 structure eliminated the standard “language translation” step, enabling direct Visible-to-Motion technology. FastDriveVLA seems to be a vital optimization layer for this new pipeline.
And even earlier within the 2nd quarter of 2024, it’s proprietary “Turing” AI chip was set for mass manufacturing. The effectivity positive factors from FastDriveVLA may permit the brand new silicon—already able to 2,200+ TOPS—to…
—-
Author : tech365
Publish date : 2025-12-30 10:46:00
Copyright for syndicated content belongs to the linked Source.
—-
1 – 2 – 3 – 4 – 5 – 6 – 7 – 8