AI Paper Atlas

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

A Solver-free Framework for Scalable Learning in Neural ILP Architectures

Yatin Nandwani, Rishabh Ranjan, Mausam ., Parag Singla

NeurIPS 2022 · Learning theory and algorithms › Combinatorial and search optimization · approach: Semidefinite relaxations for verification · 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 RRN neural baseline with ILP-LossVisual Sudoku 9x9 · Visual Sudoku 9x9board accuracy (%): 71.1 -> 98.3Improved
replace RRN with ILP-LossSymbolic Sudoku 9x9 · Symbolic Sudoku 9x9board accuracy (%): 91.3 -> 100.0Improved
compare ILP-Loss with RRN on small boardsVisual Sudoku 4x4 · Visual Sudoku 4x4board accuracy (%): 99.8 -> 99.7No clear effect
replace SATNet reasoning layer with ILP-LossVisual Sudoku 9x9 · Visual Sudoku 9x9board accuracy (%): 17.8 -> 98.3Improved
replace CombOptNet with ILP-LossSudoku · Symbolic Sudoku 9x9board accuracy (%): 0.0 -> 100.0Improved
replace CombOptNet with ILP-LossRandom constraints Binary, 8 constraints · Random constraints Binaryvector accuracy (%): 63.4 -> 87.8Improved
replace CombOptNet with ILP-LossRandom constraints Dense, 8 constraints · Random constraints Densevector accuracy (%): 2.0 -> 41.5Improved
replace CombOptNet with ILP-LossRandom constraints Binary, 1 constraint · Random constraints Binaryvector accuracy (%): 97.6 -> 97.8No clear effect
replace CombOptNet with ILP-LossRandom constraints Dense, 8 constraints · Random constraints Densetraining time (min): 48.2 -> 20.6Improved
replace CombOptNet with ILP-LossKnapsack N=30 · Knapsack from sentencesvector accuracy (%): 0.0 -> 28.4Improved
replace CombOptNet with ILP-LossKnapsack N=10 · Knapsack from sentencestraining time (min): 41.0 -> 44.0No clear effect
replace CombOptNet with ILP-LossKnapsack N=30 · Knapsack from sentencestraining time (min): 153.0 -> 111.6Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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