On learning in agent-centered search
Bibliographic record
Abstract
Since the introduction of the LRTA * algorithm, real-time heuristic search algorithms have generally followed the same plan-actlearn cycle: an agent plans one or several actions based on locally available information, executes them and then updates (i.e., learns) its heuristic function. Algorithm evaluation has almost exclusively been empirical with the results often being domain-specific and incomparable across papers. Even when unification and crossalgorithm comparisons have been carried out in a single paper, there was no understanding of how efficient the learning process was with respect to a theoretical optimum. This paper addresses the problem with two primary contributions. First, we formally define a lower bound on the amount of learning any heuristic-learning algorithm needs to do. This bound is based on the notion of heuristic depressions and allows us to have a domain-independent measure of learning efficiency across different algorithms. Second, using this measure we propose to learn “costs-so-far ” (g-costs) instead of “costs-to-go ” (h-costs). This allows us to quickly identify redundant paths and dead-end states, thereby leading to asymptotic performance improvement as well as 1-2 orders of magnitude convergence speed-ups in practice.
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 | 0.005 | 0.019 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.001 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.001 | 0.003 |
| Scholarly communication | 0.002 | 0.004 |
| Open science | 0.003 | 0.003 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
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".