Two-Sided ML Assessment for CSI Feedback Error Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in efficiently managing error causes in two-sided machine learning models trained by different vendors, particularly in reducing Channel State Information (CSI) feedback reporting overhead.
Innovation Solution
A Wireless Transmit/Receive Unit (WTRU) equipped with a processor that receives configuration information for a machine learning model, activates an assessment mode, collects measurements, determines error causes, and sends reports including error indications or mitigation actions, utilizing a two-sided ML model with first and second-side models to identify out-of-distribution data and data drift.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If two-sided ML models are independently trained by different vendors to reduce CSI feedback reporting overhead, then productivity is improved, but reliability deteriorates due to interoperability errors and data distribution mismatches
Solution Approach 1:
The patent implements an error cause determination mechanism that collects measurements from the wireless communication system, compares them against reference values, and generates feedback reports identifying specific error causes (data drift, OOD data, interoperability issues). This feedback loop enables vendors to identify and correct reliability issues while maintaining independent model training for productivity.
Solution Approach 2:
The patent introduces an intermediary assessment mode that acts as a mediator between independently trained models from different vendors. This intermediary mechanism collects measurements, determines error causes, and provides diagnostic information that facilitates interoperability without requiring joint model training, thus maintaining productivity while improving reliability.
2Reliability
If assessment mode is activated to determine error causes, then reliability is improved through error identification, but loss of time increases due to measurement collection and analysis overhead
Solution Approach 1:
The patent implements preliminary action by pre-configuring reference values, measurement parameters, and assessment criteria before the wireless communication system operates. This preparation enables rapid error cause determination during operation without requiring extensive real-time analysis, thus improving reliability while minimizing time loss.
Solution Approach 2:
The patent applies partial action by selectively activating the assessment mode only when needed for error cause determination, rather than continuously monitoring all parameters. This approach obtains sufficient diagnostic information to identify error causes without the time overhead of comprehensive continuous monitoring.
3Measurement precision
If comprehensive measurements are collected for error cause determination, then measurement precision is improved, but device complexity increases due to additional processing requirements
Solution Approach 1:
The patent segments the error cause determination process into distinct measurement types (data distribution measurements, performance measurements, intermediate KPIs) and error cause categories (data drift, OOD data, interoperability errors). This segmentation allows the system to collect comprehensive measurements for precision while organizing processing tasks into manageable segments that reduce overall device complexity.
Solution Approach 2:
The patent implements a universal measurement collection framework that can handle multiple types of measurements and error causes through a single assessment mode. This multi-functional approach improves measurement precision for various error types while avoiding the need for separate complex processing systems for each measurement type.
Data Source
AI summary
An example Wireless Transmit/Receive Unit (WTRU) comprising a processor is provided. The processor is configured to receive configuration information for a machine learning (ML) model. The processor is further configured to receive a request for activating an assessment mode associated with the ML model. The processor is further configured to collect measurements for the assessment mode. The processor is further configured to determine an error cause associated with the ML model based on the measurements. The processor is further configured to send one or more reports that include at least one of an indication of the error cause, an indication of a measurement related to the error cause, or an indication of a mitigation action for the error cause.