AI Paper Atlas

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

Guide Your Agent with Adaptive Multimodal Rewards

Changyeon Kim, Younggyo Seo, Hao Liu, Lisa Lee, Jinwoo Shin, Honglak Lee, Kimin Lee

NeurIPS 2023 · Vision-language and multimodal › Robot action and navigation · approach: Pretraining and data augmentation for VLN · 1 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
use adaptive multimodal return-conditioned policy instead of InstructRLARP-DT · Procgen CoinRun, Maze I, Maze II testexpert-normalized scoreImproved
fine-tune CLIP with VIP and IDM lossesARP-DT · Procgen test environmentsexpert-normalized scoreImproved
fine-tune CLIP representationsARP-RSSM · RLBench Pick Up Cup testsuccess rate: 20.37 -> 50.93Improved
use frozen CLIP rewards instead of baselineARP-RSSM · RLBench Pick Up Cupsuccess rateNo clear effect
train multimodal encoder from scratch instead of pre-trained CLIP (ablation)ARP-DT+ · Procgen all environmentsexpert-normalized scoreHurt
add VIP loss or IDM loss when fine-tuning CLIP (ablation)ARP-DT · Procgenexpert-normalized scoreImproved
add return prediction loss (ablation)ARP-DT · Procgenexpert-normalized scoreImproved
vary return-prediction loss coefficient lambda (ablation)ARP-DT · CoinRun held-outexpert-normalized scoreNo clear effect
replace instruction text with random text (ablation)ARP-DT · Procgen train and evalexpert-normalized scoreHurt
use unseen instruction with new objectARP-DT · CoinRun-bluegemexpert-normalized scoreImproved
use multimodal reward to distinguish distractorsARP-DT · Maze IIIexpert-normalized scoreImproved
use natural language instruction instead of goal imageARP-DT vs GC-InstructRL · Procgenexpert-normalized scoreImproved

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