AI Paper Atlas

Research map / Multi-agent systems and games / Human-AI and goal-conditioned

Housekeep: Tidying Virtual Households Using Commonsense Reasoning

Yash Kant, Arun Ramachandran, Sriram Yenamandra, Igor Gilitschenski, Dhruv Batra, Andrew Szot, Harsh Agrawal

ECCV 2022 · Multi-agent systems and games › Human-AI and goal-conditioned · approach: Embodied agent learning · 57 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
fine-tune embeddings with contrastive matching (CM) instead of zero-shot MLMRoBERTa ranker · Housekeep ORR test-unseenmAP: 0.42 -> 0.81Improved
fine-tune embeddings with contrastive matching (CM) instead of zero-shot MLMRoBERTa ranker · Housekeep OR test-unseenmAP: 0.52 -> 0.65Improved
use RoBERTa instead of GloVe embeddings with CMCM ranker · Housekeep ORR test-unseenmAP: 0.76 -> 0.81Improved
use RoBERTa instead of GloVe embeddings with CMCM ranker · Housekeep OR test-unseenmAP: 0.66 -> 0.65No clear effect
zero-shot MLM ranking instead of random rankingLM ranker · Housekeep OR test-unseenmAP: 0.59 -> 0.52Hurt
replace oracle ranker with LLM rankerOracle exploration agent · Housekeep test-seenObject Success (OS): 1.00 -> 0.44Hurt
replace oracle ranker with LLM rankerOracle exploration agent · Housekeep test-seenEpisode Success (ES): 1.00 -> 0.04Hurt
replace oracle exploration with frontier explorationOracle ranker agent · Housekeep test-seenObject Success (OS): 1.00 -> 0.64Hurt
replace oracle exploration with frontier explorationOracle ranker agent · Housekeep test-seenEpisode Success (ES): 1.00 -> 0.35Hurt
replace oracle exploration with frontier explorationLM ranker agent · Housekeep test-seenObject Success (OS): 0.44 -> 0.30Hurt
evaluate on unseen instead of seen object categoriesFull LM + frontier agent · HousekeepObject Success (OS): 0.30 -> 0.23Hurt
evaluate on unseen instead of seen object categoriesFull LM + frontier agent · HousekeepSoft Object Success (SOS): 0.39 -> 0.36No clear effect

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