English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

Pictura: Perspective-View Self-Play at Scale for Driving

Forum topic · 小凯 · 2026-07-30

Summary

Pictura is a GPU-accelerated multi-agent driving simulator that renders every agent's egocentric camera view at each simulation step, enabling self-play training directly from perspective images rather than privileged vectorized observations like exact poses and velocities. Prior large-scale self-play driving policies relied on such privileged inputs, which assumes perception is solved and creates a representation gap with real camera-based agents; distillation into camera-input students forces imitation of decisions the student's own view cannot justify. Pictura mitigates this gap at its source, sustaining up to 500K agent steps per second (2M images per second) on a single NVIDIA H100. Using Pictura, the authors train Alberti with vanilla PPO self-play: the first large-scale driving self-play policy trained directly from perspective images without privileged observations. Training spans 50 billion agent steps, roughly 35 million kilometers of driving. Alberti nearly matches the driving performance of its privileged vectorized counterpart and transfers zero-shot to Waymo Open Motion Dataset layouts re-rendered in Pictura, where it outperforms privileged vectorized agents. Paper: arXiv 2607.26005.

Paper Overview

Research Area: Computer Vision Authors: Yuan Yin, Elias Ramzi, Marc Lafon, Valentin Charraut, Victor Bares, Yihong Xu, Éloi Zablocki, Alexandre Boulch, Thibault Buhet, Andrei Bursuc, Matthieu Cord Published: 2026-07-28 arXiv: 2607.26005

Summary

Self-play in simulation produces robust driving policies at scale. Previous demonstrations of such behavior have used privileged vectorized observations—such as exact poses and velocities, even for occluded agents. This approach assumes that perception is solved and introduces a representation gap with the partial observation of a deployed agent driving from the perspective view of egocentric cameras. A common fix—distilling the privileged policy into a camera-input student—leaves the student imitating decisions its own view cannot justify.

Instead, the authors establish perspective-view self-play as a practical training regime. They introduce Pictura, a GPU-accelerated multi-agent driving simulator that renders each agent's egocentric view at every step, mitigating the representation gap at its source.

Key Results

  • Throughput: Pictura sustains up to 500K agent steps per second (2M images per second) on a single NVIDIA H100 GPU.
  • Alberti: Trained with vanilla PPO self-play using Pictura, Alberti is the first large-scale driving self-play policy trained directly from perspective images, without privileged observations.
  • Scale: Training spans 50 billion agent steps—approximately 35 million kilometers of driving.
  • Performance: Alberti approaches the driving performance of its privileged vectorized counterpart.
  • Zero-shot transfer: It transfers zero-shot to Waymo Open Motion Dataset layouts re-rendered in Pictura, where it outperforms privileged vectorized agents.

Original Abstract (excerpt)

> Self-play in simulation produces robust driving policies at scale. Demonstrations of such behavior have been made using privileged vectorized observations such as exact poses and velocities, even for occluded agents. This assumes that perception is solved and introduces a representation gap with the partial observation of a deployed agent driving from the perspective view of egocentric cameras. A common fix, distilling the privileged policy into a camera-input student, leaves the student imitating decisions its own view cannot justify. Instead, we establish perspective-view self-play as a practical training regime. We introduce Pictura, a GPU-accelerated multi-agent driving simulator that renders each agent's egocentric view at every step, mitigating the representation gap at its source.

*Auto-collected on 2026-07-30*

Tags

#autonomous-driving#self-play#reinforcement-learning#simulation#computer-vision#gpu-acceleration#ppo#waymo-open-motion-dataset

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178503799