Altruistic Bandit Learning For One-to-Many Matching Markets
Bibliographic record
Abstract
The decision-making of the agents in a matching is led by their preferences. Two-sided matching with known preferences of the agents has been studied for a long time. Present-day matching applications require considering stochastic environments where the agents learn their preferences iteratively through explorations. Competition arises when multiple agents attempt to pair with an agent. Such competitions can be resolved when the preferences of the agents being competed for are known. A recent study of modelling multi-arm bandit learning as two-sided matching with known preferences for the agents of one side has been conducted. In a real scenario, the learning agents may compete excessively for a few agents. Some learning agents may only be compatible to make a pair with those few competitive agents. The number of unmatched agents tends to be more in such cases. We propose an altruistic bandit learning model for one-to-many matching where learning agents sacrifice their choices to disburden competition from the highly desired agents. The model decreases the chances of agents with limited pairing options remaining unmatched, and the cardinality of the matching increases as a result. This model is compatible with applications in humanitarian operations where stability-cardinality trade-off exists, such as homeless shelter matching. We structure the model using the upper confidence bound algorithm. We experimented with diverse-sized and valued synthetic instances. The results indicate a satisfactory increment of matching cardinality while the reward loss of the sacrificing agents is optimistically acceptable. That justifies the model as a fair and useful two-sided matching mechanism with unknown preferences for one side.
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 | 0.004 | 0.011 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.006 | 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".