Avatar

Sevvandi Kandanaarachchi

Senior Research Scientist

CSIRO's Data61

Hello! Thanks for visiting my site! 👋

I work on statistical machine learning and data science research. I am interested in many things. Lately, I’ve been interested in networks or graphs. There is a lot of cool mathematics related to graphs. Previously, I’ve worked on anomaly and event detection and algorithm evaluation. I still work on these topics, but somewhat less than I used to. There are more links between these topics than meets the eye. Sometimes linking different topics gives us better insights. It is fun when that happens. I also work on applications where AI is used in Science. This is a growing area, encompassing many topics and disciplines.

I am a Senior Research Scientist at Australia’s National Science Agency, CSIRO. My work website is here. At CSIRO’s Data61 I work on interesting cutting-edge scientific research. I work with other researchers who come from different backgrounds like biology and social science. It is fascinating to work with people from different disciplines and see “Science at Work”! I really enjoy these interactions and love the experience.

I am a multi-disciplinary researcher, an applied mathematician by training. To give a bit of my background my PhD was in mean curvature flow, which is in the intersection of differential geometry and partial differential equations. There is a lot of mathematics in AI and machine learning. I love learning new mathematics and computer science concepts.

Interests

  • Anomaly Detection
  • Algorithm Evaluation
  • AI in Science
  • Graphs and Networks

Education

  • Graduate Certificate in Data Mining and Applications, 2015

    Stanford University

  • PhD in Mathematics, 2011

    Monash University

  • MSc Preliminary in Mathematics, 2007

    Monash University

  • BSc Eng. in Computer Science and Engineering, 2002

    University of Moratuwa

Research themes

Anomalies and Algorithms

New algorithms for anomaly detection

Cyber Security

Detecting intrusions in computer networks

Sensor data

Are the sensors telling us something different?

Recent & Upcoming Talks

Software

netseer

Predicting the structure of dynamic graphs - an R package.

stxplore

Tools for spatio-temporal data exploration - an R package.

oddnet

Network anomaly detection - an R package.

outlierensembles

A collection of ensembles for outlier detection - an R package.

lookout

Leave-one-out kernel density estimates for outlier detection - an R package.

composits

Outliers in univariate, multivariate and compositional time series - an R package.

airt

Algorithmic IRT - an R package.

eventstream

Event detection and early classification for streaming data - an R package.

dobin

Dimension reduction for outlier detection - an R package.

Blog posts

Anomaly Detection Ensembles

What is an anomaly detection ensemble? It is a bunch of anomaly detection methods put together to get a final anomaly score/prediction. …

Testing an Outlier Detection Method

Suppose you have developed an outlier detection method. What are the ways to test it? You can generate some random data and add a …

Anomaly detection data repositories

In this post we will look at data repositories available for anomaly detection. So, can you use a standard classification dataset for …

Which nonlinear dimension reduction methods preserve outliers inside a sphere?

In this post we will look at how well nonlinear dimension reduction techniques preserve outliers that are placed inside a sphere, when …

Anomaly detection dilemmas

Finding anomalies/outliers in data is a task that is increasingly getting more attention mainly due to the variety of applications …

Podcasts

Forecasting Impact podcasts

In 2023 I conducted a couple of episodes for the Forecasting Impact podcast with with Mahdi Abolghasemi. Prof George Athanasopoulos on …

Random Sample podcasts in 2023

Industry PhDs I did two more podcasts in 2023 for the Random Sample. One podcast was on industry PhDs. Industry PhDs give a lot more …