AI Paper Atlas

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

Learning-Augmented Scalable Linear Assignment Problem Optimization via Neural Dual Warm-Starts

Ilay Yavlovich, Jad Agbaria, Muhamed Mhamed, Nir Weinberger, Jose Yallouz

ICML 2026 · Learning theory and algorithms › Combinatorial and search optimization · approach: Unsupervised neural combinatorial optimization · 0 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
warm-start LAPJV with predicted dualsLAPJV · Dense Model, N=16,384end-to-end speedup: ≈2.0× vs SciPy, ≈2.5× vs LAPImproved
warm-start LAPJV with predicted dualsLAPJV · Block-Structured Model, N=16,384end-to-end speedup: ≈2.25× vs SciPy, ≈4.0× vs LAPImproved
warm-start at small sizeLAPJV · N=512end-to-end speedupHurt
use neural warm-start on sparse MOT matricesLAPJV · Multi-object trackingend-to-end speedup: ≈1.25× vs LAP, ≈2× vs SciPyMixed
neural warm-start improves greedy matchingLAPJV · Dense/Blockgreedy match rate: ≈26% -> ≈76%Improved
add neural warm-start to reduce runtime varianceLAPJV · MOTcoefficient of variation: ≈45% -> ≈30%Improved
replace RowDualNet with linear regression (ablation)RowDualNet · Dense/Blockspeedup: below baseline for N > 4,096Hurt
replace RowDualNet with Random or Row Mean heuristic (ablation)RowDualNet · Densespeedup: speedup <1.0Hurt
replace RowDualNet with time-bounded subgradient ascent (ablation)RowDualNet · Dense/Block N=8192speedup: ≈2.5× -> ≈1.5×Hurt
subgradient initialization on sparse MOT (ablation)LAPJV · MOTspeedup: speedup < 1.0×Hurt
add positional encodings to features (d=13 -> d=21) (ablation)RowDualNet · Block-structured, N=4096speedupImproved
add positional encodings to features (ablation)RowDualNet · Dense, N=4096speedupNo 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