5 images via IEEE Spectrum – AI
🤖AI Summary
General Motors is developing scalable AI systems that can train autonomous driving at 50,000x real-time speed through high-fidelity simulations. The company combines Vision Language Action models, reinforcement learning, and millions of daily simulations to handle rare 'long-tail' driving scenarios that current systems struggle with.
Key Takeaways
- →GM runs millions of high-fidelity closed-loop simulations daily, equivalent to tens of thousands of human driving days compressed into hours.
- →The company uses Vision Language Action models that can understand complex scenarios like police gestures overriding traffic lights.
- →GM employs a 'Dual Frequency VLA' approach with large models for semantic decisions and smaller models for immediate spatial control.
- →Long-tail scenarios include both rare events like burst fire hydrants and common-sense situations requiring human-like courtesy.
- →The simulation-based training allows testing of hazardous scenarios that would be impossible to safely replicate in real-world testing.
#autonomous-driving#general-motors#ai-simulation#vision-language-models#reinforcement-learning#automotive-ai#edge-cases#real-time-training
Read Original →via IEEE Spectrum – AI
Act on this with AI
Stay ahead of the market.
Connect your wallet to an AI agent. It reads balances, proposes swaps and bridges across 15 chains — you keep full control of your keys.
Related Articles