AI Paper Atlas

Research map / Learning theory and algorithms / Combinatorial and search optimization

Data-Driven Approximations to NP-Hard Problems

Anton Milan, S. Rezatofighi, Ravi Garg, Anthony Dick, Ian Reid

AAAI 2017 · Learning theory and algorithms › Combinatorial and search optimization · approach: Unsupervised neural combinatorial optimization · 56 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
train with objective-based gradient replacement instead of loss-only training (ablation)LSTM · Synthetic trackingOSPA-T: 0.42 -> 0.37Improved
train with objective-based gradient replacement instead of loss-only training (ablation)LSTM · Synthetic trackingID switches: 0.70 -> 0.60Improved
train with objective-based gradient replacement instead of loss-only training (ablation)LSTM · Feature point matching (motorbikes)accuracy: 0.68 -> 0.76Improved
train with objective-based gradient replacement instead of loss-only training (ablation)LSTM · Feature point matching (motorbikes)objective: 9.54 -> 10.52Improved
add objective-based gradient replacement to Pointer-Net training (ablation)Pointer-Net · TSP 20 nodes, Qatar maptour length: 3.80 -> 3.45Improved
train LSTM on approximate JPDA10 labels onlyLSTM · Synthetic trackingOSPA-T: 0.43 -> 0.42Improved
use LSTM (obj.) instead of exact JPDALSTM · Synthetic trackingID switches: 0.40 -> 0.60Hurt

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