Workshop on Data Science at Best Practices Meet 2017, Data Security Council of India
2 likes706 views
The document provides an overview of a workshop on data science and machine learning, emphasizing the rapid growth of data and the need for analytics today. It explains key concepts like machine learning, deep learning, and their applications, particularly in cybersecurity. Additionally, it outlines resources for learning data science and highlights tools for big data analytics.
Workshop on Data Science at Best Practices Meet 2017, Data Security Council of India
- 1. © Copyright 2017 NETMONASTERY Inc Thinking Beyond the Human Envelope A Quick Workshop on Data Science using Machine Learning 1 Shomiron DAS GUPTA - Founder, CEO NETMONASTERY Inc.
- 2. © Copyright 2017 NETMONASTERY Inc Agenda ■ More on Analytics, ML and DL ■ Why Analytics today? ■ Where does Analytics take us ■ Machine Learning, Simply ■ The Royal Family of ML ■ Solving with ML ■ Deep Learning - Where to NEXT ! 2 Building Context Grounds Up
- 3. © Copyright 2017 NETMONASTERY Inc More on Analytics, ML and DL Data Science has many children - 3 Let’s Get it Straight - Data Science is Big Data Analytics Machine Learning Deep Learning
- 5. © Copyright 2017 NETMONASTERY Inc Why Analytics Today? We created 90% of the world's data in the last 2 years Largest customer in 2014 was generating 4TB / Month, the same customer today is generating 69TB / Month Need to look through and analyze the recorded past - there is a clear want to reference and learn from our data, more so in the cyber security Horizontal scaling is the key - brings speed with consistency 5 Bringing History Back into Cyber Security
- 6. © Copyright 2017 NETMONASTERY Inc Where Does Analytics Take Us? What Works Well with Big Data Analytics Engines Thresholding / Multi-Dimensional Thresholding Time Series Analytics including Heuristics Baseline / Profilers Correlators 6 Different Forks Available to Process Big Data Wheels on Fire
- 8. © Copyright 2017 NETMONASTERY Inc Machine Learning, Simply 8 It’s the ability to make machines think and take experiential decisions, it’s the ability of the machine to think beyond the human envelope. Speed / Accuracy of decision making Ability to perform “logically” in unknown conditions Let’s Try to Define Machine Learning
- 9. © Copyright 2017 NETMONASTERY Inc Reinforced Learning 9
- 10. © Copyright 2017 NETMONASTERY Inc The Royal Family of ML First - Classification and Regression ■ Trees - Decision Trees ■ Forests - Random Forests ■ Support Vector Machines ■ K Nearest Neighbours (KNN) ■ Linear / Logistic Regression ■ Neural Networks 10 Describing the Tip of the Iceberg
- 11. © Copyright 2017 NETMONASTERY Inc How it Works - Simple Decision Tree The Basics Features Labels Dataset Training Testing 11 Consistent Decision Making - is something we could all use ;) The Process Selecting the right features Moderating and cleaning the dataset Splitting the dataset - train / test Building and training a model Testing for accuracy - chaining / pipelining etc.
- 12. © Copyright 2017 NETMONASTERY Inc Solving with ML ■ Email SPAM ! ■ Detecting / Generating DGN’s ■ Network Anomaly Stuff we (DNIF) does ■ Detect InBots ■ Bad traffic to Applications ■ UBM - Bad Access Attempt 12 Cyber Security Challenges that have their roots in ML
- 13. © Copyright 2017 NETMONASTERY Inc Deep Learning - Where to NEXT ! Google with Deep Learning ■ Voice API ■ Video API ■ Image API ■ Translation API ■ Speech API ■ Relearning ■ finally TensorFlow 13 What is currently available - Where are we trying to go...
- 14. © Copyright 2017 NETMONASTERY Inc Where to Begin 1. Massive resources on YouTube 2. Best if you learn - Python / R ….. python will take you the farthest 3. Work on practical challenges - seek answers 4. Big data platforms with built-in logic Splunk.com - Leader in Big Data Analytics DNIF.it - Complete toolkit. 100G Free Forever 14 NO, YOU ARE NOT LATE !
- 15. © Copyright 2017 NETMONASTERY Inc Thank You email: shom@dnif.it 15