Our platform leverages many forms of data science as well as artificial intelligence.
Our Modeler records every interaction taken by every user in your system
By combining different models we achieve higher testing accuracy
We use long short-term memory models to predict next steps with 85% accuracy
Functionize AI Platform
Reach Test Perfmance as if you have a army of QA professionals. By introducing the statistical models your test case becomes more robust and error prone.
Define Test Objective
Reinforcement Learning Markov Decision Process Q Learning
Unsupervised Learning Expectation Maximization Local Sensitivity Hashing LSTM Model
English Test Workflow
Word Embeddings Element Embeddings Neural Networks
Talk Test Steps
Speech to Text Time Segmentation NLP Modeling
Markov Decision Process
Computer Vision Template Recognition
Probability Mass Functions
Robust Test Model
Multinomial Logistic Regression
Convolutional Neural Networks
Recurrent Neural Networks
Runtime System in Cloud
- Visual Anomaly detection
- Element Anomaly detection
- Timing Anomaly detection
Root Cause Analysis
- Decision Tree
- Neural Networks
ML Modeling Benefits
Understand real users
Real users never behave as you would predict. By tagging and recording every interaction on your UI we are able to give you deeper insights into how users actually interact with your application. This can help define new tests, but can also assist your product team to identify UX problems or unused flows.
Identifying unique user journeys
Having identified user flows, we are then able to predict next steps for users with high accuracy. This is a powerful tool, as it also allows us to improve our understanding of how your application works.
By itself, AI is pretty dumb. To work properly, it needs as much data as possible. Using data science approaches helps us to extract additional data from your application and thus acts to further boost our AI models.
Coping with sparse data
When you start creating tests for a new application, our system has no past history to go on. So, we have to use any techniques we can to enhance the information available. Often, this means we turn to classic data science for assistance.
Enabling template recognition
Our template recognition approach relies on knowledge of how UI elements are typically grouped together. However, the link between elements isn’t rigid e.g. dates are sometimes MM/DD/YY, sometimes DD/MM/YYYY, etc. So, we identify these relationships by using data science techniques like Akaike Information Criteria.
For machine learning and AI to reach its full potential, other elements of supervised learning are often required to “boost” model performance and optimize results. Boosting is a term where weak models are made into strong models. Adaptive Boosting is a machine learning meta-algorithm that can be used in conjunction with many other types of learning algorithms to improve performance.
Nowadays, boosting techniques are used to help solve a wide range of AI problems. For instance here at Functionize, we use an autonomous intelligent test agent to run all your automated tests. This test agent uses multiple forms of artificial intelligence. Many of these rely in turn on boosting. For instance, our Adaptive Event Analysis™ engine uses computer vision as one way to identify and select elements on the screen.
Sign Up Today
The Functionize platform is powered by our Adaptive Event Analysis™ technology which incorporates self-learning algorithms and machine learning in a cloud-based solution.