Fast Synthesis of Algebraic Heuristic Functions for Video-game Pathfinding
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Heuristic search is widely used in games for pathfinding and general planning. High-quality heuristic functions are key to finding a low-cost solution quickly. Commonly used heuristic functions for video-game pathfinding are either manually designed and generic or pre-computed for a specific map. The former fail to take advantage of pathfinding specifics while the latter tend to have a large memory footprint, may require substantial pre-computation and are not portable to other maps or easily presentable to humans. In this work we attempt to combine the best of both approaches by automatically synthesizing well performing pathfinding-specific yet compact and human-readable heuristics. We do so by defining a space of algebraic formulae expressing heuristic functions and then conducting an automated search of the space. To make the synthesis tractable we employ a multi-tier evaluation which allows us to quickly filter out low-quality heuristics while saving time to more thoroughly evaluate better ones. Such triage of candidate heuristics enables us to synthesize compact heuristics that outperform the standard baseline on video-game pathfinding benchmarks. By then adding the synthesized heuristics back to the synthesis space we show that synthesis on new maps can be substantially sped up to merely few minutes per map.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it