# Explaining a machine learning decision to physicians via counterfactuals

arXiv: https://arxiv.org/abs/2306.06325 · HTML: https://mlsanity.com/paper/2306.06325/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 16 of its 42 references that are in the map.
- First posted on arXiv 10 Jun 2023.

## Builds on (most cited first, 16 of 16)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [A Unified Approach to Interpreting Model Predictions](https://mlsanity.com/paper/1705.07874/) (2017)
- [Auto-Encoding Variational Bayes](https://mlsanity.com/paper/1312.6114/) (2013)
- [Towards A Rigorous Science of Interpretable Machine Learning](https://mlsanity.com/paper/1702.08608/) (2017)
- [Counterfactual Explanations without Opening the Black Box: Automated Decisions and the GDPR](https://mlsanity.com/paper/1711.00399/) (2017)
- [Explaining Machine Learning Classifiers through Diverse Counterfactual Explanations](https://mlsanity.com/paper/1905.07697/) (2019)
- [CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning](https://mlsanity.com/paper/1711.05225/) (2017)
- [Tutorial on Variational Autoencoders](https://mlsanity.com/paper/1606.05908/) (2016)
- [A Survey on Explainable Artificial Intelligence (XAI): Towards Medical XAI](https://mlsanity.com/paper/1907.07374/) (2019)
- [Explanations based on the Missing: Towards Contrastive Explanations with Pertinent Negatives](https://mlsanity.com/paper/1802.07623/) (2018)
- [Towards Realistic Individual Recourse and Actionable Explanations in Black-Box Decision Making Systems](https://mlsanity.com/paper/1907.09615/) (2019)
- [MIMIC-Extract: A Data Extraction, Preprocessing, and Representation Pipeline for MIMIC-III](https://mlsanity.com/paper/1907.08322/) (2019)
- [Instance-based Counterfactual Explanations for Time Series Classification](https://mlsanity.com/paper/2009.13211/) (2020)
- [Counterfactual Explanations for Machine Learning on Multivariate Time Series Data](https://mlsanity.com/paper/2008.10781/) (2020)
- [Learning to Detect Sepsis with a Multitask Gaussian Process RNN Classifier](https://mlsanity.com/paper/1706.04152/) (2017)
- [RADNET: Radiologist Level Accuracy using Deep Learning for HEMORRHAGE detection in CT Scans](https://mlsanity.com/paper/1710.04934/) (2017)

## Built on by (most cited first, 4 of 4)

- [Beyond Known Reality: Exploiting Counterfactual Explanations for Medical Research](https://mlsanity.com/paper/2307.02131/) (2023)
- [WANDER: An Explainable Decision-Support Framework for HPC](https://mlsanity.com/paper/2506.04049/) (2025)
- [On the notion of missingness for path attribution explainability methods in medical settings: Guiding the selection of medically meaningful baselines](https://mlsanity.com/paper/2508.14482/) (2025)
- [CDEG: Learning Decision-Critical Evidence for Long-Horizon Diagnostic Agents](https://mlsanity.com/paper/2608.22899/) (2026)

## Abstract (arXiv)

Machine learning models perform well on several healthcare tasks and can help reduce the burden on the healthcare system. However, the lack of explainability is a major roadblock to their adoption in hospitals. textitHow can the decision of an ML model be explained to a physician? The explanations considered in this paper are counterfactuals (CFs), hypothetical scenarios that would have resulted in the opposite outcome. Specifically, time-series CFs are investigated, inspired by the way physicians converse and reason out decisions `I would have given the patient a vasopressor if their blood pressure was lower and falling'. Key properties of CFs that are particularly meaningful in clinical settings are outlined: physiological plausibility, relevance to the task and sparse perturbations. Past work on CF generation does not satisfy these properties, specifically plausibility in that realistic time-series CFs are not generated. A variational autoencoder (VAE)-based approach is proposed that captures these desired properties. The method produces CFs that improve on prior approaches quantitatively (more plausible CFs as evaluated by their likelihood w.r.t original data distribution, and 100× faster at generating CFs) and qualitatively (2× more plausible and relevant) as evaluated by three physicians.