Anomaly Detection


What's new with BigQuery ML: Unsupervised anomaly detection for time series and non-time series data

When it comes to anomaly detection, one of the key challenges that many organizations face is that it can be difficult to know how to define what an anomaly is. How do you define and anticipate unusual network intrusions, manufacturing defects, or insurance fraud? If you have labeled data with known anomalies, then you can choose from a variety of supervised machine learning model types that are already supported in BigQuery ML.


Science of Network Anomalies

Today’s networks have evolved a long way since their early days and have become rather complicated systems that comprise numerous different network devices, protocols, and applications. Consequently, it is practically impossible to have a complete overview of what is happening in the network or whether everything in the network works as it should. Eventually, network problems will arise.

Rethinking Anomaly Detection

John Sipple, Staff Software Engineer in AI, at Google Cloud presents Google's story about rethinking anomaly detection. In 2019, Google Smart Buildings asked the team to develop an AI-based fault-detection solution to help find and fix problems in climate control devices in large office buildings. Technicians were dissatisfied with conventional outlier approaches because they didn’t give the necessary insight to predict, diagnose and intervene. The result was a distributed deep-learning solution that provides explanations to aid understanding, prioritizing and fixing faults. We applied it to other domains, like data center monitoring and fraud detection, and then open-sourced the MADI machine learning algorithm behind it. We’ll describe our vision of how AI will shape the future of interpretable anomaly detection.

Bridge the gap in your OSS by adding an AI brain on top

Telecom companies monitor their network using a variety of monitoring tools. There are separate fault management and performance management platforms for different areas of the network (core, RAN, etc.), and infrastructure is monitored separately. Although these solutions monitor network functions and logic – something that would seem to make sense — in practice this strategy fails to produce accurate and effective monitoring or reduce time to detection of service experience issues.

Correlation Analysis Explained

When you detect that something is off in your business, how long does it take you to find the root cause? The longer it takes, the more it can cost you. Correlation analysis identifies relationships between KPIs, which business teams use to accelerate root cause analysis (RCA) and mean time to remediation (MTTR). Doing it manually however can be tedious and limit your visibility.

Using Elastic machine learning rare analysis to hunt for the unusual

It is incredibly useful to be able to identify the most unusual data in your Elasticsearch indices. However, it can be incredibly difficult to manually find unusual content if you are collecting large volumes of data. Fortunately, Elastic machine learning can be used to easily build a model of your data and apply anomaly detection algorithms to detect what is rare/unusual in the data. And with machine learning, the larger the dataset, the better.


Only Autonomous Anomaly Detection Scales

Say you’re looking for a smart product to detect anomalies in your organization’s IT environment. A sales rep drops by and shows you all kinds of great artificial intelligence (AI) features with fancy-sounding algorithms. It sounds very impressive and seems like there is a lot of very valuable AI in the product. But, in fact, the opposite is true. This is a manual AI product wrapped in a deceiving jacket. Let me tell you more.


Introducing: Business Impact Alerts

Anodot is the only monitoring solution built from the ground up to find and fix key business incidents, as they’re happening. As opposed to most monitoring solutions, which focus on machine and system data to track performance, Anodot also monitors the more volatile and less predictable business metrics that directly impact your company’s bottom line. Now there’s an easy way to measure the business impact of every incident.