AI Paper Atlas

Research map / Vision-language and multimodal / CLIP and zero-shot

VITA: Zero-Shot Value Functions via Test-Time Adaptation of Vision–Language Models

Christos Ziakas, Alessandra Russo

ICLR 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: Adapting VLMs for zero-shot classification

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add test-time adaptation with sequential updates to CLIP-based value functionVITA · ms_sweep (long-horizon)VOC: 0.434 -> 0.490Improved
apply sequential per-step adaptation instead of trajectory-level update (ablation)VITA · tk_pnp validationVOC: 0.1917 -> 0.7822Improved
sequential per-step adaptation instead of memoryless reset each step (ablation)VITA · tk_pnp validationVOC: 0.1918 -> 0.7822Improved
sequential adaptation instead of explicit window memory (ablation)VITA · tk_pnp validationVOC: 0.1957 -> 0.7822Improved
use dissimilarity-based sampling instead of full-trajectory sampling (ablation)VITA · Scripted datasetsDisc@VOC: 0.20 -> 1.0Improved
use dissimilarity-based sampling instead of random sampling (ablation)VITA · Scripted datasetsDisc@VOC: 0.8 -> 1.0Improved
reduce sampled windows from k=8 to k=4 (ablation)VITA · Scripted datasetsDisc@VOC: 1.0 -> 1.0No clear effect
reduce sub-trajectory length from 8 to 4 (ablation)VITA · Scripted datasetsDisc@VOC: 1.0 -> 0.60Hurt
increase test-time steps from 1 to 10 at lr 0.1 (ablation)VITA · Validation setVOC: 0.783 -> 0.369Hurt
increase test-time steps from 1 to 10 at lr 0.01 (ablation)VITA · Validation setVOC: 0.454 -> 0.781Improved
use VITA instead of GVL autoregressive VLMVITA · tk_pnp (in-distribution)VOC: 0.269 -> 0.782Improved
use VITA instead of CLIP-GRU recurrent memoryVITA · 10 evaluation datasets VOCVOC: 6 of 10 datasets betterMixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy