# Discretizing Logged Interaction Data Biases Learning for Decision-Making

arXiv: https://arxiv.org/abs/1810.03025 · HTML: https://mlsanity.com/paper/1810.03025/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 6 Oct 2018.

## Built on by (most cited first, 3 of 3)

- [Exploring Time-Step Size in Reinforcement Learning for Sepsis Treatment](https://mlsanity.com/paper/2511.20913/) (2025)
- [The hidden risks of temporal resampling in clinical reinforcement learning](https://mlsanity.com/paper/2602.06603/) (2026)
- [Insulin4RL: Real-Time Insulin Management in the Intensive Care Unit for Offline Reinforcement Learning](https://mlsanity.com/paper/2606.19481/) (2026)

## Abstract (arXiv)

Time series data that are not measured at regular intervals are commonly discretized as a preprocessing step. For example, data about customer arrival times might be simplified by summing the number of arrivals within hourly intervals, which produces a discrete-time time series that is easier to model. In this abstract, we show that discretization introduces a bias that affects models trained for decision-making. We refer to this phenomenon as discretization bias, and show that we can avoid it by using continuous-time models instead.