cs.LGNov 9, 2026

DeltaForge: A Post-hoc Mixture-of-Experts VLA for Multi-Domain Autonomous Driving

Authors: Bingzhang Wang, Mo Poorsartep, Yasaman Esfandiari

Organizations: NVIDIA Corporation, Santa Clara, CA, USA. · University of Washington, Seattle, WA, USA.

Abstract

Fleets of driving vision-language-action (VLA) models are developed one domain at a time: each operator finetunes a shared base model on private data, producing specialized experts that are costly to consolidate into a single model. Weight merging loses accuracy as heterogeneous domains accumulate, and dense mixture-of-experts routing is prohibitively expensive to combine them. We present DeltaForge, to our knowledge the first post-hoc mixture-of-experts VLA for driving, which consolidates N independently finetuned per-domain models into a single multi-domain model from their weights and a small proxy set alone. When different operators finetune the same base model, their weight updates end up sharing a lot of structure. We exploit this by decomposing all the updates together using singular value decomposition (SVD), splitting each one into a shared part (the same across all domains) and a tiny domain-specific part. The router then only needs to work with those tiny domain-specific pieces, not the full expert weights, which is what keeps the approach efficient. Then two cooperating clip-level blind routers, one over the VLM reasoning experts and one over the flow-matching dynamics experts, are trained from soft labels on the proxy dataset. Across six diverse domains spanning urban, off-road, and highway driving on cars, trucks, and all-terrain vehicles (ATVs), DeltaForge approaches the accuracy of the per-domain experts and surpasses every merging, routing, centralized, and federated baseline at close to single-model parameter and compute cost. Adding a new domain requires only projecting its weight update onto the shared basis with no end-to-end retraining needed. The resulting model also generalizes to domain combinations it was never explicitly trained on.

Figures & tables

Explore similar work

CardsList
  1. FlashDrive: Flash Vision-Language-Action Inference for Autonomous Driving

    Aug 13, 2026Zekai Li, Yihao Liang, Hongfei Zhang +3Efficient VLA Model InferenceAutonomous Driving

  2. Fast-dDrive: Efficient Block-Diffusion VLM for Autonomous Driving

    May 22, 2026Kewei Zhang, Jin Wang, Sensen Gao +9VLMs for Autonomous DrivingEnd-to-End Autonomous Driving

  3. DriveFine: Refining-Augmented Masked Diffusion VLA for Precise and Robust Driving

    Feb 16, 2026Chenxu Dang, Sining Ang, Yongkang Li +7VLMs for Autonomous DrivingDiffusion-Based Planning