2021-satellite-damage-assessment.html

---
layout: publication
authors:
  - Donghoon Shin
  - Sachin Grover
  - Kenneth Holstein
  - Adam Perer
pdf: https://arxiv.org/pdf/2111.02626.pdf
arxiv: "2111.02626"
tags:
  - Explainable AI
  - Crowdsourcing
  - Healthcare
title: "Characterizing Human Explanation Strategies to Inform the Design of Explainable AI for Building Damage Assessment"
type:
  - Workshop
venue: AI for HADR at NeurIPS
venue_tags:
  - NeurIPS
  - AI for HADR
venue_url: https://www.hadr.ai
year: 2021
---

Explainable AI (XAI) is a promising means of supporting human-AI collaborations
for high-stakes visual detection tasks, such as damage detection tasks from
satellite imageries, as fully-automated approaches are unlikely to be perfectly
safe and reliable. However, most existing XAI techniques are not informed by the
understandings of task-specific needs of humans for explanations. Thus, we took
a first step toward understanding what forms of XAI humans require in damage
detection tasks. We conducted an online crowdsourced study to understand how
people explain their own assessments, when evaluating the severity of building
damage based on satellite imagery. Through the study with 60 crowdworkers, we
surfaced six major strategies that humans utilize to explain their visual damage
assessments. We present implications of our findings for the design of XAI
methods for such visual detection contexts, and discuss opportunities for future
research.