UAS Multi-Temporal Disaster Change Reasoning QA Dataset

#Deductive reasoning #Temporal comparison #Change classification #Evidence-grounded question answering #Disaster response mapping #Pre/post-disaster change analysis #Damage assessment #UAS inspection

5750
0.4GB
JSON
CATL
MOBIUSI INC

Updated:2026-11-03

AI Analysis & Value Prop

This dataset presents pre- and post-disaster UAS mapping observations as question-answer examples, with explicit rules for distinguishing newly observed changes, pre-existing damage, and cases where evidence is insufficient. Each example includes a question, answer, change classification, reasoning, and evidence summary, supporting reasoning SFT and disaster change attribution evaluation while helping reduce false attribution of existing conditions as new damage. The dataset size is not specified in the available information.

Dataset Insights

Sample Examples

{}

Technical Specifications

Field	Type	Description
file_name	string	File name
question	string	The question to answer by comparing the pre- and post-disaster survey records and applying the inference rules.
reasoning	string	A brief explanation of how the pre- and post-disaster records and stated rules support the classification.
answer_text	string	A concise natural-language answer grounded in the temporal survey evidence and inference rules.
inference_rules	string	Explicit rules for distinguishing newly observed changes, pre-existing damage, and cases where the evidence is insufficient.
evidence_summary	string	A concise summary of the observations supporting the conclusion, including relevant temporal comparisons or missing evidence.
pre_disaster_survey	string	Textualized UAS mapping observations recorded before the disaster, including any damage or conditions already present.
post_disaster_survey	string	Textualized UAS mapping observations recorded after the disaster.
change_classification	string	The inferred temporal status of the relevant condition.

Compliance Statement

Authorization Type	Proprietary - Commercial AI Training License (No Redistribution)
Commercial Use	Requires exclusive subscription or authorization contract (monthly or per-invocation charging)
Privacy and Anonymization	No PII, no real company names, simulated scenarios follow industry standards
Compliance System	Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs

Frequently Asked Questions

What does the dataset contain?: Examples include pre- and post-disaster UAS mapping records, inference rules, questions and answers, as well as change classifications, reasoning, and evidence summaries.

How does it distinguish new damage from pre-existing damage?: Compare the pre- and post-disaster records over time and apply the supplied inference rules to classify a condition as a new change, pre-existing damage, or a case where the available evidence is insufficient.

What are the dataset size and annotation quality?: The available information does not specify the number of examples or quantitative quality metrics. The structured rules, classifications, and evidence summaries make it easier to verify whether conclusions are supported by temporal evidence.

Which tasks is it suitable for, and how can I access it and check its license?: It is suitable for reasoning SFT, disaster change attribution, change classification, and evidence-grounded question-answering evaluation. Refer to the dataset listing or license agreement for access details and permitted use.

Can't find the data you need?

Post a request and let data providers reach out to you.

Get this Dataset

Verified for Enterprise Use

Cite this Work

@dataset{Mobiusi2026,
  title={UAS Multi-Temporal Disaster Change Reasoning QA Dataset},
  author={MOBIUSI INC},
  year={2026},
  url={https://www.mobiusi.com/datasets/d07a8f5a2d437a17dcac4196c82c572f},
  urldate={2026-10-26},
  keywords={UAS disaster dataset,pre and post-disaster change detection,disaster mapping QA,disaster damage assessment,multi-temporal reasoning data,UAS mapping data,disaster change attribution,reasoning SFT dataset},
  version={1.0}
}

Using this in research? Please cite us.

placeholder

Scene

Domain

Modality

Task