Bibliographic record
Abstract
Large real-time search problems such as path-finding in com-puter games and robotics limit the applicability of complete search methods such as A*. As a result, real-time heuris-tic methods are becoming more wide-spread in practice. These algorithms typically conduct a limited-depth looka-head search and evaluate the states at the frontier using a heuristic. Actions selected by such methods can be subop-timal due to the incompleteness of their search and inaccu-racies in the heuristic. Lookahead pathologies occur when a deeper search decreases the chances of selecting a better action. Over the last two decades research on lookahead pathologies has focused on minimax search and small syn-thetic examples in single-agent search. As real-time search methods gain ground in applications, the importance of un-derstanding and remedying lookahead pathologies increases. This paper, for the first time, conducts a large scale inves-tigation of lookahead pathologies in the domain of real-time path-finding. We use maps from commercial computer games to show that deeper search often not only consumes addi-tional in-game CPU cycles but also decreases path quality. As a second contribution, we suggest three explanations for such pathologies and support them empirically. Finally, we propose a remedy to lookahead pathologies via a method for dynamic lookahead depth selection. This method substan-tially improves on-line performance and, as an added benefit, spares the user from having to tune a control parameter.
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.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.007 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.011 | 0.003 |
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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".