Tupl’s Network Advisor is an AI-powered assistant designed to boost your Network Performance team productivity by automating daily repetitive performance and troubleshooting tasks.Request Demo
How Network Advisor Works
The power of Machine Learning for network engineering automation. Your team can develop their own Root Cause Analysis and Action Recommendation models and automate these tasks to 80% level. Adapts to any changing environment with intuitive training and re-training when needed. Perfect for complex networks including spectrum refarming, RAN vendor swap, IoT and 5G.
Benefits of integrating Tupl to your network
- Turns your expert knowledge into digital knowledge
- Unlocks the power of AI for Network Automation
- Easy integration with SON, NOC & field services
- AI Engine: Based on open source Big Data and Machine Learning libraries, Tupl’s AI Engine is the core of Network Advisor, providing the capability to train and run multiple-stage root cause analysis and action recommendation models in a live network.
- Machine Learning UI: Tupl’s UI has a unique approach. It simplifies Machine Learning model creation, training and evaluation. Tupl ML toolkit UI uses a combination of supervised and unsupervised learning in a way that your experts can create and maintain automation ML models without prior coding knowledge.
- Feature Engineering Editor: UI based functionality supports the creation of the most suitable features as inputs for maximum ML model performance.
- Multiple source data integration: Through TuplOS data and cloud utilities, multiple data sources can be easily correlated and processed to allow for complex decision making.
How We Fit
Network Advisor is designed to fit within your organization’s daily network performance and troubleshooting process.
Network Advisor connects to any and all relevant data sources for optimization tasks, key ones being configuration, topology and network counters and KPIs. Once the models are designed with the help of Tupl ML toolkit, the AI Engine correlates data from multiple sources, such as performance, configuration, alarms or detailed call records to automate the detection, troubleshooting and action recommendations for the network performance issues.
Easy Creation of ML model
Once a problem is identified, a machine learning based root cause analysis model is applied to determine the most probable cause and subsequent actions are triggered to other systems to provide a fix. In case the problem requires specific engineering review, it is forwarded to the appropriate team.
Frequently Asked QuestionsNetwork Advisor
What type of root causes is Network Advisor addressing?
One key design principle of the system is flexibility, so as long as the proper data sources are integrated, and enough training samples are provided, any root cause that is relevant to you can be addressed.
How accurate are the root cause analysis models?
It typically depends on a combination of targeted number of root causes, number of features required to model them and the number of available samples to train the system. Some of the models created by Network Advisor have reached over 90% accuracy.
We don't keep records of solved issues to use as training data. Is this an issue?
Not at all. Network Advisor leverages unsupervised clustering techniques to detect group issues with similar patterns and helps you analyze as many problems as required to train your model in the most efficient way.
What other systems can work with Network Advisor?
Network Advisor is conceived to work in conjunction with other systems within your company, such as Ticketing, SON, CM, or Work Order management systems. Just as any engineer would do.
My organization does not have data science skills. How steep is the learning curve for this solution?
No coding skills or database modelling are required, only general concepts from data science are advisable to make the most of the system. Network Advisor relies upon TuplOS, Tupl's AI engine, which is a set of utilities designed to synthesize and digitize the complex knowledge base from your engineers and free them to focus on putting their expertise to work.
Tupl can guide you through the process via a coaching and training program from our services team.