Loading
Life sciences · Preprint
arXiv · September 8, 2026
Early or partial results. Treat as a signal, not a conclusion.
This preprint proposes BAFF, a parameterized filtering strategy to mitigate training data interference in RTB A/B tests caused by control and treatment models sharing a serving log. In offline simulation and a single live DSP deployment, filter-based variants preserved business metrics (CPC, CTR) closer to an interference-free reference model than log-sharing or log-splitting baselines, with performance dependent on setting-specific parameter choice.
Methodological study with offline simulation and single live deployment. Online A/B tests in real-time bidding; control and treatment models trained on shared serving logs in a DSP. Intervention: BAFF (k,l)-parameterized hard filters applied to training data. Compared with: Log-sharing (all data, unaddressed bias) and log-splitting (no bias, reduced data). Live deployment on a demand-side platform (DSP); specific location not stated.
BAFF (k,l)-parameterized hard filters control tolerance to ad-ranking and bid-pricing disagreement independently Offline simulation identified operating points with smaller deviation from interference-free reference than both log-sharing and log-splitting Live RTB deployment showed filter-based variants preserved business metrics (CPC, CTR) more closely than both baselines
Safety was not reported in the material analysed. Check the source before drawing any conclusion about harm.
The source did not state who this applies to in practice.
A novel methodological approach to A/B testing bias in RTB systems, demonstrated in simulation and one live deployment, but lacking peer review, formal statistical inference, and independent replication.
As stated by the source record.
Quoted from the source exactly as published.
Graded across the dimensions that decide whether you should act, each from what the source actually supports. There is no single score, and where a dimension was not assessed it says so.
In online A/B tests for real-time bidding (RTB), control and treatment models are typically trained on a shared serving log that includes data generated by the counterpart model. This shared-log training biases each model's training data through two channels: the counterpart model may have selected a different ad from the ad-candidate pool (ad-ranking disagreement) and may have bid a different price (bid-pricing disagreement), potentially distorting the A/B test outcome. Log-splitting eliminates the bias but sacrifices training data; log-sharing retains all data but leaves the bias unaddressed. We formalize the Bid-Aware Filter Family (BAFF), a class of (k,l)-parameterized hard filters that controls tolerance to each channel independently, providing a structured search space between these two extremes. We further propose a three-stage online measurement protocol that enables evaluating data-sharing strategies by their deviation from an interference-free reference model in production. In offline simulation, a (k,l) sweep surfaces operating points with smaller deviation from the interference-free reference model than both log-sharing and log-splitting. In a live RTB deployment on a demand-side platform (DSP), filter-based variants preserve the reference model's business metrics (e.g., CPC, CTR) more closely than both baselines. The best operating point is setting-dependent, underscoring the practical value of the search space itself.
Taken from the source record, never inferred. Follow any of these and new work involving them reaches your briefing.