# Align-Deform-Subtract: An Interventional Framework for Explaining Object Differences

arXiv: https://arxiv.org/abs/2203.04694 · HTML: https://mlsanity.com/paper/2203.04694/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 7 of its 20 references that are in the map.
- First posted on arXiv 9 Mar 2022.

## Builds on (most cited first, 7 of 7)

- [Mask R-CNN](https://mlsanity.com/paper/1703.06870/) (2017)
- [Self-Supervised Learning with Data Augmentations Provably Isolates Content from Style](https://mlsanity.com/paper/2106.04619/) (2021)
- [Representation Learning via Invariant Causal Mechanisms](https://mlsanity.com/paper/2010.07922/) (2020)
- [Convolutional neural network architecture for geometric matching](https://mlsanity.com/paper/1703.05593/) (2017)
- [Deep Convolutional Inverse Graphics Network](https://mlsanity.com/paper/1503.03167/) (2015)
- [End-to-end weakly-supervised semantic alignment](https://mlsanity.com/paper/1712.06861/) (2017)
- [Proposal Flow](https://mlsanity.com/paper/1511.05065/) (2015)

## Built on by (most cited first, 1 of 1)

- [Towards Trustworthy and Aligned Machine Learning: A Data-centric Survey with Causality Perspectives](https://mlsanity.com/paper/2307.16851/) (2023)

## Abstract (arXiv)

Given two object images, how can we explain their differences in terms of the underlying object properties? To address this question, we propose Align-Deform-Subtract (ADS) -- an interventional framework for explaining object differences. By leveraging semantic alignments in image-space as counterfactual interventions on the underlying object properties, ADS iteratively quantifies and removes differences in object properties. The result is a set of "disentangled" error measures which explain object differences in terms of the underlying properties. Experiments on real and synthetic data illustrate the efficacy of the framework.