Washington D.C. – At its annual .conf conference, Splunk Inc. announced that it has extended its machine learning capabilities throughout its product portfolio.
The goal with this expansion is to make machine learning as accessible and transparent as possible. Splunk envisions a future where using machine learning algorithms to pull information from big data sets is so simple that the user doesn’t even necessarily know it is machine learning doing the work.
“We had all this power in the platform that no one really understood what to do with, so we created a machine learning toolkit to actually show you how to apply machine learning to solve various problems,” Rick Fitz, senior vice president of IT markets at Splunk, told ITWC. “We understand that not everyone is a data scientist, and therefore not everyone can apply these algorithms to solve problems.”
For example, with the machine learning toolkit Splunk can tell a user that this is the algorithm that it uses, and here is the training data set you would want to require, and here are the different sources, and then this is how the system will respond. The company has now extended those machine learning toolkit capabilities to the rest of its platform.
“You can give someone an algorithm and say use this to solve a problem, but most customers aren’t data scientists or have a PhD in statistics, but they have a rudimentary understanding of their own environment. The real power is being able to take this on their own and apply it in a way that is transparent in what you are doing so that the user can understand how you are using the math without actually understanding the math,” said Fitz.
Splunk Enterprise 7.0, Splunk IT Service Intelligence (ITSI) 3.0, Splunk User Behavior Analytics (UBA) 4.0, and updates to the company’s cloud service all focus on machine learning with the goal to make the technology more accessible to Splunk solution users. This is then centered around a machine learning toolkit, which recently has received its third update.
Here are the big machine learning updates to the Splunk portfolio:
- Splunk ITSI 3.0 – The latest version of the event monitoring solution combines service context with machine learning to help identify existing and potential issues while prioritizing restoration of business-critical services and delivering analytics driven IT operations.
- Splunk UBA 4.0 – Newest version allows data scientists to write and load their own machine learning algorithms to create custom anomalies and threats via Spunk’s new software developers kit (SDK).
- Splunk Machine Learning Toolkit – The free solution’s latest updates focus around a visual interface for creating and managing models, as well as public APIs for custom algorithms. It also features Spark (MLLiB) support for algorithmic training.
Fitz points to alert fatigue in Splunk ITSI, and how machine learning in 3.0 alleviates that pain point.
“There is a lot of value in this data, and you just need to be able to group it in a way that you can take action on it,” said Fitz. “We are doing [machine learning] in a way that allows the user to configure the tools, pick the features, and tune it, but we do it in a way so they don’t know it’s machine learning.”
Users look at the interface, change a slider here and there, and then they can see the data. And because users can configure it, they can then see the results of the algorithm and tune it in a way that makes the most sense for their business.
“They configure the machine learning without having to know the math. That’s the experience we want to strive for,” said Fitz.
