PSUG 3 - 2024-07-15 - Splunk & AI with Philipp Drieger
tomasmoser
329 views
44 slides
Jul 17, 2024
Slide 1 of 44
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
About This Presentation
Once in a life time opportunity for Prague Splunk User Group and Splunkers in Czechia and abroad. Join us to discover Splunk AI and Machine Learning (ML) capabilities in a rare session presented by Philipp Drieger, Global Principal Machine Learning Architect at Splunk. With AI hype all over the worl...
Once in a life time opportunity for Prague Splunk User Group and Splunkers in Czechia and abroad. Join us to discover Splunk AI and Machine Learning (ML) capabilities in a rare session presented by Philipp Drieger, Global Principal Machine Learning Architect at Splunk. With AI hype all over the world these days this is a unique moment and a chance to bring together those already familiar with Splunk universal machine data platform but without any AI/ML knowledge or experience and seasoned or full time data scientists interested in Splunk and its AI/ML capabilities.
Part 1: Introduction to Splunk AI (45min)
Get to know Splunk AI first hand from Philipp, Global Principal Architect for Machine Learning at Splunk. He will share a easy to understand overview of Splunk's key AI components and also highlight some real world customer use cases.
Open Q&A
Part 2: Splunk AI demos and open AMA session (45min)
Join Philipp showing live demos including Splunk's Machine Learning Toolkit, the Splunk App for Data Science and Deep Learning and the latest Splunk AI Assistant.
Open AMA session: Ask Me Anything about Splunk AI
Embedded capabilities
within products
Customizable
ML, deep
learning, and
data science
tools
AI libraries and
APIs for
developers
Generative
AI chatbots
Guided
assistive
workflows
AI
Tools
Splunk ITSI applies machine learning to
proactively prevent outages by correlating and
reducing alerts, monitoring service health, and
streamlining incident management.
❏Clustering & aggregation to reduce alert noise
❏Adaptive (dynamic) thresholds incorporate seasonality
❏Anomaly and outlier detection
❏Actionable additional context
❏Assisted root cause investigation
❏Predict service health to prevent outages
New updates!
❏Outlier Exclusion in Adaptive Thresholds
❏ML-Assisted Thresholding (Preview)
Create Datasets
Collect data and use Splunk to parse the data and identify patterns that can be used to detect the threat
Build ML-Powered Detections
Build a model based on data in order to make predictions or decisions; enable systems to learn from data, identify patterns,
and make decisions with minimal human intervention; and craft rules or queries designed to identify specific activity associated
with threats
Test Detections
Run queries against a dataset that simulates attacker behavior to improve accuracy and reduce false positives
Release
Package detections to deliver timely and effective protections against emerging threats to Splunk customers
Designed for Splunk users at all levels
•ML-Powered Splunk Searches: Apply techniques like anomaly
detection and predictions within search to power dashboards &
insights
•Showcase and Experiments: Simple low-code experience to
guide model building, testing, and deployment
•Extensible out of the box: 80+ built-in scikit-learn algorithms,
and API support to plug in new runtimes
New updates!
•Ability to upload externally pre-trained ONNX models with a
simple UI and then use the model with your Splunk data
with no modification to your existing workflows
•Extended user anomaly detection capabilities with a new
algorithm for multivariate outlier detection
Extend Splunk to Operationalize Machine Learning Use Cases Within Search
Splunk Enterprise 9.1, Splunk Cloud Platform
Built for Data Scientists
‒35+ Code Examples: Guided model building, testing, and
deployment of data science and deep learning frameworks
‒Container Management: Models can be productionized for
scalability & optimization of resources, e.g. CPU & GPU
‒State of the art AI frameworks and tools: Jupyter Lab,
MLflow, PyTorch, TensorFlow, SpaCy, DASK, Rapids, Spark, …
‒Flexible deployments and open source: deploy on-prem,
hybrid or in the cloud. Github repository for customization.
New updates in version 5.1!
‒Two AI assistants to leverage LLMs to build and train
models for text summarization and text classification use
cases
‒Customizable for adapting to own domain specific data
●
Extension for MLTK to operationalize advanced custom AI / ML use cases
Splunk Enterprise 9.1, Splunk Cloud Platform
power and protect the AI revolution.
Networking and
compute solutions
for training AI
models at scale.
Massive breadth
and depth of data
across domains is
the foundation for
AI models.
Observability
supports
trustworthy,
performant, and
reliable AI
deployment.
AI solutions
improve efficiency,
efficacy, and
economics of
defending against
security threats.
Cisco’s Identity
Intelligence and
User Protection
and Splunk’s
analytics help
protect enterprises
in their use of AI
from within.
Infrastructure
for AI
Data
for AI
Observability
for AI
AI
for Security
Security
for AI