---
format: "aidr-story-markdown/v1"
id: "c4e64ec122b346fd0ca7be107526c4b2c8d16c1c3b2f73f6f604af3002d3b671"
canonical_url: "https://aidr.today/c4e64ec1?lang=en"
title: "BirdDreams: Joint Perspective and BEV Video Generation for a Plausible, Reactive Driving World Model"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-13T00:00:00.000Z"
category: "Research"
topics: ["agent"]
source_urls: ["https://huggingface.co/papers/2611.14815"]
summary: "BirdDreams is an autoregressive driving world model that jointly generates perspective video and a schematic bird's-eye-view BEV video. The BEV depicts roads and vehicles as simple shapes at a fixed metric scale. Perspective video retains scene details like traffic signals and object appearance."
---

# BirdDreams: Joint Perspective and BEV Video Generation for a Plausible, Reactive Driving World Model

> [Open the canonical story](<https://aidr.today/c4e64ec1?lang=en>)

**Published:** 2026-11-13T00:00:00.000Z
**Category:** Research
**Topics:** agent

## Summary

BirdDreams is an autoregressive driving world model that jointly generates perspective video and a schematic bird's\-eye\-view BEV video\. The BEV depicts roads and vehicles as simple shapes at a fixed metric scale\. Perspective video retains scene details like traffic signals and object appearance\.

## Sources

- [Story source](<https://huggingface.co/papers/2611.14815>)