AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

Efficient Policy Adaptation with Contrastive Prompt Ensemble for Embodied Agents

Wonje Choi, Woo Kyung Kim, SeungHyun Kim, Honguk Woo

NeurIPS 2023 · Reinforcement learning › Visual and representation RL · approach: Robust visual encoders for RL · 3 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 ConPE instead of EmbCLIPEmbCLIP · AI2THOR ObjectNav unseensuccess rate: 59.0 -> 79.7Improved
use ConPE instead of EmbCLIPEmbCLIP · Egocentric-Metaworld Reach-Wall unseensuccess rate: 49.3 -> 67.3Improved
use ConPE instead of EmbCLIPEmbCLIP · CARLA Map 1 unseensum of rewards: 1415.1 -> 1933.4Improved
use ConPE instead of ATC on PointNav seenATC · AI2THOR PointNav seensuccess rate: 89.1 -> 89.7Improved
add ConPE prompt ensemble to pretrained policyPretrained policy · AI2THOR ObjectNav unseensuccess rate: 65.8 -> 72.8Improved
add ConPE prompt ensemble to pretrained policyPretrained policy · Egocentric-Metaworld Button-Press targetsuccess rate: 16.8 -> 73.7Improved
increase number of prompts from 2 to 10 (ablation)ConPE · AI2THOR ObjectNav unseensuccess rate: 43.0 -> 79.7Improved
increase number of prompts from 10 to 16 (ablation)ConPE · AI2THOR ObjectNav unseensuccess rate: 79.7 -> 77.1No clear effect
increase number of prompts from 10 to 18 (ablation)ConPE · AI2THOR ObjectNav unseensuccess rate: 79.7 -> 79.0No clear effect
replace guided attention with uniform average of prompted embeddings (ENS-UNI-AVG) (ablation)ConPE · AI2THOR ObjectNav unseensuccess rate: 65.5 -> 79.7Improved
replace guided attention with prompt-level uniform average (COM-UNI-AVG) (ablation)ConPE · AI2THOR ObjectNav unseensuccess rate: 43.1 -> 79.7Improved
remove policy prompt p_pol from adaptation of pretrained policy (ablation)ConPE · AI2THOR PointNav targetsuccess rate: 54.2 -> 84.4Improved

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