vision-language

Definition

The combining form for systems that handle images and text together — vision-language models take pixels as input and produce or condition on natural language. The usual construction pairs a visual encoder with a language model through a projection layer, so the language side does the reasoning and the vision side supplies grounded features. The term is defined against vision-only and language-only baselines, which is how ablations in these papers are framed.

Explainers using this term

• Paper walkthrough: Qwen-Drive-1.0 — bolting 3D perception and planning onto a VLM without touching its architectureQwen-Drive-1.0: An Initial Step towards a Vision-Language Foundation Model for Autonomous Driving
• Paper Explained: Beyond Data Scaling — Why the Backbone, Not the Trajectory Count, Decides Your VLA (VLAct)Beyond Data Scaling: Representation-Centric Continued Pre-training for Vision-Language-Action Models
• Paper Walkthrough: TurboVLA — Kick the LLM Out of the Loop and Run a Robot Policy at 32 Hz on an RTX 4090 with Under 1 GB of VRAMTurboVLA: Real-Time Vision-Language-Action Model at 32 Hz on an RTX 4090 with <1 GB VRAM