Agilent AI Peak Integration for MassHunter

AI is a technique which enables machines to mimic human behavior. It uses logic, if-then rules, decision trees, and machine learning.

Machine Learning (ML) is a subset of AI technique which uses statistical methods to enable machines to improve with experience. It uses algorithms to incorporate intelligence into machines by automatically learning from data.

Deep Learning (DL) is a subset of ML which makes the computation of multi-layer neural network feasible. It uses algorithms that permit software to train itself to perform tasks, like image recognition, by exposing multi-layered neutral networks to vast amount of data.

A machine learning architecture, designed for GC/MS data processing software allows continuous “training” of lab-specific methods for various testing services. The tool operates initially in passive mode, using image recognition in DL to learn how a chemist manually integrates peaks during their data analysis workflow. The data analysis activity is continuously monitored and fed into a training pipeline to generate a specific model from the dataset using the DL neural network.

The first release of the AI software is targeting phthalates in consumer products. Additional applications are in development.  

Many automated peak integrators have a ‘one size fits all’ model that may not deliver the best results and will still require manual integration. The ML technology copies how you manually integrate specific compounds, eventually eliminating the need for any subjective manual integration. 

• Reproducible results
• Global consistency of analytical results  
• Ease-of-use for expert and new lab members 
• Reduced turnaround time
• Improved overall lab efficiency  

The software is a cloud-based with an AI algorithm  that learns in the background while your lab operators perform their work. Once the work is learned, the data analysis software automates manual integration from your method. 

One of the main benefits of this AI tool is it eliminates the need for manual integration which leads to automated integration of peaks with little error. As a result, newer analysts are not required to perform work they’re not accustomed to. 

Model performance is monitored within the GC/MS data processing software once thresholds are met, allowing reproducible results across labs and improved accuracy.