|
Work item:
|
Y.IoT-DQA-FMT
|
|
Subject/title:
|
Assessment metrics for IoT data quality in foundation model training
|
|
Status:
|
Under study
|
|
Approval process:
|
AAP
|
|
Type of work item:
|
Recommendation
|
|
Version:
|
New
|
|
Equivalent number:
|
-
|
|
Timing:
|
2027-Q3 (Medium priority)
|
|
Liaison:
|
ITU-T SG21, ISO/IEC JTC1/SC42
|
|
Supporting members:
|
Zhejiang Dahua Technology, China Information Communication Technologies Group, Electronics and Telecommunications Research Institute, China Unicom, China Telecom
|
|
Summary:
|
IoT (Internet of Things) data is the cornerstone of intelligent decision-making and large-scale machine learning model training. It provides real-time, accurate environmental and device information, playing a crucial role in enhancing intelligence levels, optimizing prediction accuracy, and improving application effectiveness. With the rapid development of AIoT (Artificial Intelligence of Things), foundational models are becoming increasingly dependent on high-quality IoT data, making it necessary to establish specialized quality assessment metrics to enhance model performance.
Therefore, the goal of this draft Recommendation is to create a scientific set of IoT data quality assessment metrics in foundational model training, ensuring that these models are trained and deployed on high-quality data, thereby improving their accuracy, stability, and applicability.
It provides an overview, metrics, and methods for IoT data quality assessment in foundation model training.
|
|
Comment:
|
-
|
|
Reference(s):
|
|
|
Historic references:
|
|
Contact(s):
|
|
| ITU-T A.5 justification(s): |
|
|
|
|
First registration in the WP:
2025-01-30 13:10:02
|
|
Last update:
2025-02-04 10:13:21
|
