---
format: "aidr-story-markdown/v1"
id: "fa6f5139b53e08df2b5d026ca738a6b9fefadbff11ca17a88af6a5729282c1a1"
canonical_url: "https://aidr.today/fa6f5139?lang=en"
title: "MDN-VJEPA: Multimodal Probabilistic Joint-Embedding World Models"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-10-14T04:00:00.000Z"
category: "Research"
topics: ["inference"]
source_urls: ["https://arxiv.org/abs/2610.13956"]
summary: "MDN-VJEPA is a multimodal probabilistic JEPA that uses a mixture density network to represent future states with a Gaussian mixture encoding alternative futures, their probabilities, and within-mode uncertainty."
---

# MDN\-VJEPA: Multimodal Probabilistic Joint\-Embedding World Models

> [Open the canonical story](<https://aidr.today/fa6f5139?lang=en>)

**Published:** 2026-10-14T04:00:00.000Z
**Category:** Research
**Topics:** inference

## Summary

MDN\-VJEPA is a multimodal probabilistic JEPA that uses a mixture density network to represent future states with a Gaussian mixture encoding alternative futures, their probabilities, and within\-mode uncertainty\.

## Sources

- [Story source](<https://arxiv.org/abs/2610.13956>)