AI Paper Atlas

Research map / Vision-language and multimodal / Robot action and navigation

Visual Grounding for Object-Level Generalization in Reinforcement Learning

Haobin Jiang, Zongqing Lu

ECCV 2024 · Vision-language and multimodal › Robot action and navigation · approach: Multimodal prompt-driven manipulation · 5 citations (OpenAlex)

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
replace MineCLIP reward with object-grounded focal rewardPPO agent · hunt a cowsuccess rate (%): 3.8±4.8 -> 71.3±9.7Improved
add denoising (negative-word zeroing and 0.2 threshold) to confidence map (ablation)Focal reward · hunt a cowsuccess rate (%): 3.8±4.8 -> 71.3±9.7Improved
use focal reward level instead of its step-to-step difference (ablation)Focal reward · hunt a cowsuccess rate (%): 3.5±3.0 -> 71.3±9.7Improved
add Gaussian kernel centered on view to focal reward (ablation)Focal reward · hunt a cow, two target animalssuccess rate (%): 14.5±21.8 -> 67.0±15.0Improved
condition policy on confidence map instead of language embeddingCOPL · Hunting domain, unseen targets (avg)success rate (%): 11.5±7.6 -> 48.1±6.6Improved
condition policy on confidence map instead of language embeddingCOPL · Harvest domain, unseen targets (avg)success rate (%): 16.7±4.3 -> 40.8±8.3Improved
condition policy on confidence map instead of language embeddingCOPL · Harvest domain, training tasks (avg)success rate (%): 73.2±5.2 -> 75.6±3.2No clear effect
condition policy on confidence map instead of language embeddingCOPL · Hunting domain, training tasks (avg)success rate (%): 16.6±13.3 -> 50.0±8.9Improved
COPL versus STEVE-1 on unseen spider targetCOPL · hunt a spider (unseen)success rate (%): 54.5±12.7 vs 73.0Hurt

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