AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

Reward Finetuning for Faster and More Accurate Unsupervised Object Discovery

Katie Z Luo, Zhenzhen Liu, Xiangyu Chen, Yurong You, Sagie Benaim, Cheng Perng Phoo, Mark Campbell, Wen Sun, Bharath Hariharan, Kilian Q Weinberger

NeurIPS 2023 · Detection and segmentation › Open-set and few-shot detection · approach: Hand-object and articulated interaction · 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
replace repeated self-training with reward-ranked local exploration and finetuningPointRCNN object discovery · Lyft test, 0-80mmAP IoU 0.5: 39.6 -> 45.3Improved
replace repeated self-training with reward-ranked local exploration and finetuningPointRCNN object discovery · Lyft test, 0-80mmAP IoU 0.7: 18.8 -> 29.6Improved
replace repeated self-training with reward-ranked finetuningPointRCNN object discovery · Ithaca365 test, 0-80mmAP: 27.1 -> 35.1Improved
add exploration by locally sampling boxes around predictions (ablation)DRIFT · Lyft test, 0-80mmAP IoU 0.5: 0.0 -> 41.8Improved
add the shape prior reward (ablation)DRIFT with filtering and alignment rewards · Lyft test, 0-80mmAP IoU 0.5: 22.4 -> 38.3Improved
add alignment reward to filtering and shape prior (ablation)DRIFT · Lyft test, 0-80mmAP IoU 0.5: 0.0 -> 38.3Improved
increase exploration sample count from 50 to 200 at noise scale 0.3 (ablation)DRIFT · Lyft test, 0-80mmAP IoU 0.5: 32.9 -> 38.3Improved
increase exploration noise scale from 0.3 to 0.6 with 200 samples (ablation)DRIFT · Lyft test, 0-80mmAP IoU 0.5: 38.3 -> 0.6Hurt
add sampling near dynamic points to prediction-based exploration (ablation)DRIFT · Lyft test, 0-80mmAP IoU 0.5: 41.8 -> 39.3Hurt
add sampling from initial seed labels to exploration (ablation)DRIFT · Lyft test, 0-80mmAP IoU 0.5: 41.8 -> 39.5Hurt
change alignment reward μ_scale from 0.8 to 0.9 (ablation)DRIFT · Lyft test, 0-80mmAP IoU 0.5: 38.3 -> 38.1No clear effect
change alignment reward σ_scale from 0.2 to 0.1 (ablation)DRIFT · Lyft test, 0-80mmAP IoU 0.5: 38.3 -> 34.2Hurt

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