Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

AAC: Admissible-by-Architecture Differentiable Landmark Compression for ALT

About

We introduce \textbf{AAC} (Architecturally Admissible Compressor), a differentiable landmark-selection module for ALT (A*, Landmarks, and Triangle inequality) shortest-path heuristics whose outputs are admissible by construction: each forward pass is a row-stochastic mixture of triangle-inequality lower bounds, so the heuristic is admissible for \emph{every} parameter setting without requiring convergence, calibration, or projection. At deployment, the module reduces to classical ALT on a learned subset, composing end-to-end with neural encoders while preserving the classical toolchain. The construction is the first differentiable instance of the compress-while-preserving-admissibility tradition in classical heuristic search. Under a matched per-vertex memory protocol, we establish that ALT with farthest-point-sampling landmarks (FPS-ALT) has provably near-optimal coverage on metric graphs, leaving at most a few percentage points of headroom for \emph{any} selector. AAC operates near this ceiling: the gap is $0.9$--$3.9$ percentage points on 9 road networks and ${\leq}1.3$ percentage points on synthetic graphs, with zero admissibility violations across $1{,}500+$ queries and all logged runs. At matched memory, AAC is also $1.2$--$1.5{\times}$ faster than FPS-ALT at the median query on DIMACS road networks, amortizing its offline cost within $170$--$1{,}924$ queries. A controlled ablation isolates the binding constraint: training-objective drift under default initialization, not architectural capacity; identity-on-first-$m$ initialization closes the expansion-count gap entirely. We release the module, a reusable matched-memory benchmarking protocol with paired two-one-sided-test (TOST) equivalence and pre-registration, and a reference compressed-differential-heuristics baseline.

An T. Le, Vien Ngo• 2026

Related benchmarks

TaskDatasetResultRank
Expansion Count ReductionModena OSMnx road network 30K
Expansion Reduction91.9
18
Expansion ReductionSBM 10k (test)
Mean Expansion Reduction97.53
15
Expansion ReductionBarabási-Albert 10k (test)
Mean Expansion Reduction97.28
15
A* Expansion ReductionDIMACS NY directed 264K
Mean Expansion Reduction92.6
15
Expansion ReductionOGB-arXiv 169k (test)
Mean Expansion Reduction96.3
15
A* Expansion ReductionSBM 10K (undirected)
Mean Expansion Reduction96.9
9
Shortest Path Query AccelerationDIMACS NY
Expansion Reduction92.3
9
Shortest Path Distance EstimationFLA graph DIMACS (test)
Performance (%)91.8
6
Shortest Path Distance EstimationModena graph OSMnx (val)
Performance (%)93
6
Shortest Path Distance EstimationModena graph OSMnx (test)
Performance Score91.7
6
Showing 10 of 30 rows

Other info

Follow for update