Space and depth-related enhancements of the history-ADS strategy in game playing
Bibliographic record
Abstract
In the field of game playing, it is a well-known fact that powerful strategies, such as alpha-beta search, benefit strongly from proper move ordering. A popular metric of achieving this is the so-called “move history”, that is, prioritizing moves that have performed well, earlier in the search. The literature reports a number of techniques, such as the Killer Moves and History heuristics, that employ such a philosophy. Inspired by techniques from the field of Adaptive Data Structures (ADSs), we <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> have previously introduced the History-ADS heuristic, which uses an adaptive list to record moves, and to improve move ordering based on move history. The History-ADS heuristic has been proven to produce substantial gains in tree pruning in a wide variety of cases. However, it made use of a relatively naive application of an unbounded, single adaptive list. In this work, we attempt to refine the History-ADS heuristic, by examining its performance by constraining the length of its adaptive list, and applying multiple ADSs for each level of the tree. Our results show that the vast majority of the savings from the History-ADS heuristic remain even with a very short list, which can be applied to mitigate the drawbacks of an unbound data structure. Although results for multiple ADSs did not outperform single ADSs, we show that they provide some insight into how similar techniques may be applied in the context of the History-ADS heuristic.
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 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 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, 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".