Artem Lenskiy

Senior Research Fellow

Picture of Artem Lenskiy

Hanna Neumann Building 145


+61 405555553

Data Science

Google Scholar

Artem Lenskiy an interdisciplinary expert with over 15 years of experience in math. modeling, finance, cryptography, software development, and delivering projects in the health, defence, and industry domains.

Employment History

  • Senior Research Fellow, Australian National University, Australia, Sep 2019 - Sep 2022
  • Assistant Professor, Korea University of Technology \& Education, S. Korea. Mar 2011 - Sep 2019
  • Research Intern, RIKEN Brain Science Institute, Japan, June 2011 - Aug 2011


  • MS (completed coursework), Applied Mathematics, Johns Hopkins Univ., 2017- , U.S.A.
  • Ph.D., Electrical Engineering(Computer Vision), Univ. of Ulsan, S. Korea
  • MS, Digital Signal Processing, Novosibirsk State Technical Univ., R.F.
  • BS, Computer Engineering, Novosibirsk State Technical Univ., R.F.

Funded Projects / Contracts

  • Chief Investigator: Developing machine/deep learning models for detecting improvised explosive devices ($500,000). Funded by National Intelligence and Security Discovery Research Grants, Department of Defence (2022-), Australia.
  • Principal Investigator: Investigated graph-based machine learning algorithms for discovering causal relations in COVID-19 literature. Funded by The Defence Science and Technology Group, 2021, Department of Defence, Australia.
  • Chief Investigator: Performed privacy analysis of Immunisation Register and Pharmaceutical Benefits Scheme data sets (in collaboration with Software Innovation Institute). Funded by Department of Health, 2020, Australia.
  • Principal Investigator: Developed ROS-based educational robots, The National Human Resources Development Institute, 2016, S. Korea.
  • Chief Investigator: (2012). Autonomous Vehicle Competition. Funded by The Agency for Defense Development (ADD), S. Korea.
  • Chief Investigator: (2007-2010). A terrain identification system for navigating military robots. Funded by ADD, S. Korea.


Selected Publications

Privacy/Computer networks

Machine Learning / Recommender Systems


Computer vision/Robotics

Natural Language Processing

Quantitative finance


Curriculum development and Teaching

Computer engineering/science

  • Programming as Problem Solving (in Haskell), 2020.
  • Networked Info. Systems. (several invited lectures incl. blockchain 1, 2, 3), 2020, 2021.

Courses taught prior joining the ANU

  • Introduction to programming, 2011 - 2019.
  • C language, 2011 - 2019. The curriculum for both Introduction to Programming and C language were developed and video lectures were recorded for flip-learning.
  • Data communication, 2012 - 2016
  • Computer networks, 2012 - 2017
  • Multimedia communication, 2014 - 2017.
  • Digital logic, 2011
  • Computer architecture, 2011


Courses taught prior joining the ANU

  • Probability theory and random processes, 2011 - 2019.
  • Linear models for time-series analysis, 2017 - 2019.
  • Matrix theory and applications, 2018 - 2019.
  • Machine learning, 2018 - 2019.
  • Differential equations, 2018.
  • Real analysis, 2019.

Canberra Computer Science Enrichment (a program for gifted high-school students)

  • Secure computation (2022)
  • Introduction to machine learning (2022)
  • Blockchain technologies and cryptocurrencies (2021)
  • Network attacks (2020)
  • Basics of computer networks (2020)

Academic supervision

  • Currently 5 PhD, 3 have graduated
  • Currently 5 MS, 13 have graduated
  • Currently 3 undergraduate, 23 have graduated

Recent talks (overall over 20)

  • Privacy preserving computing, Global Cybersecurity Camp, Taiwan, 2022
  • Introduction to Homomorphic encryption and machine learning, Cybersecurity workshop, Northeastern Illinois University, USA, 2020
  • The latest thinking on what machine learning can do and how to do it, Government Economist Conference, Canberra, Australia, 2019
  • Text mining in social media, University of Technology Sydney, Sydney, Australia, 2019


Available Projects

Privacy and Cryptography

  • Homomorphic encryption: Developing a method for approximating any function by a polynomial under the constraints defined by the acceptable error or the degree of a polynomial. Evaluating this function using Ring Learning With Error (implemented by Seal) scheme to solve real world problems that require privacy.
  • Differential privacy: Employ models of machine learning over differentially private data to make decisions w/o leaking information 1, 2.
  • Quantifying privacy leaks in recommender systems: reconstructing users’ demographic profiles and other user information by analysing users’ preferences (e.g. items purchased, music listened, movies watched, etc) with two goals in mind: (1) looking for privacy leaks and possible crosslinks between databases, (2) improving recommendation by taking into account reconstructed demographics.

Mathematical physics

  • Solve (model) non-linear physical systems by means of neural networks. Some problems include double pendulum and 3-body problems. ( 1, 2, 3, 4 ).
  • Modeling cardiovascular system using Neural Networks: A high complexity of the cardiovascular system makes it hard for ordinary differential equations (ODEs) to model the system especially given many unobservable parameters. In fact, most models simplify the cardiovascular structure to varying degrees, for example, according to previous study, T.G. Myers et al. designed a four compartments model of the cardiovascular system based on Ottesen’s approach that relies on the ODEs. In order to overcome the limitation of unobservability and complexity of cardiovascular system, neural models are proposed as a replacement for standard ODEs. According to Tian Qi Chen et al., neural ODEs are better suitable for grasping the dynamics of complex systems. The core step of the proposed project is to improve the EKG accuracy in modeling healthy EKGs and EKGs with heart diseases.


  • Generating fractal images e.g. Mandelbrot or Julia sets with various initial conditions and search for meaningful patterns using convolutional neural network pretrained on ImageNet or CIFAR-100. The goal is to find pictures similar to Burning Ship fractal.


  • Employ machine learning to analyse financial quarterly reports published by the U.S Security and Exchange Commission (SEC) to make investment decisions. The investments strategies should be tested on historic data.

Algorithmic trading

  • Trading engine (cosupervised by Dr Vitalii Pruks): Develop a trading engine using Robot Operating System 2 (ROS2) and test it on an arbitrage trading strategy. There is a basic engine developed by previous students that needs to be extended.
  • Market making: develop a marking making agent using reinforcement learning or short-term orderbook prediction using matrix autoregressive models or other types of regression.
  • Modelling market regimes: Simulating market regimes by the multi-agent discrete event time market simulator (ABIDES by JP Morgan). The model should account for trending markets as well as for “flat” regimes i.e. when the price stays in a channel. Should this model be successful, the size and the number of stop-losses will decrease, resulting in higher profits.
  • Classification of market regimes: The idea of this project is to employ dynamic time warping to learn a vocabulary of price patterns and a transition probability matrix (Markov Chain) from the sequences of the extracted patterns conditioning on market regimes.


  • Parkinson prediction: Employ convolutional neural networks to EEG signals for Parkinson disease prediction. The EEG are transformed onto a map that represents spatial electrode locations and electrical potential activity, similar to *.
  • Learning score systems: The severity of many diseases is assessed by score systems designed by medical experts. For instance, Parkinson disorder is assessed by UPDRS and severity of Stroke by SSS or NIHSS. In this project a (neural) symbolic regression approach is used to find a better score that represents a formula that takes into account known patients’ factors.
  • A smartphone app for self-diagnosing: Development of a smartphone app that uses the phone’s camera to extract photoplethysmogram and then devise such user’s biomedical parameters such as biological age, systolic/diastolic blood-pressure, and perhaps weight, height, gender. The dataset is available here.

Hardware project

  • Design glasses (3D printed) with an electrooculography (EOG) sensor to measure eye movements and blink detection. The data is then analysed on an onboard computer (Raspberry Pi / Arduino) for blink detection and eye-movement analysis.

Blockchain and smart contracts

  • Develop a smart contract on Ethereum blockchain to match organ donors and recipients. The goal is to remove third parties from organ donation and make the protocol decentralised while preserving privacy of donors and recipients.

Bioinformatics and biosensing:

  • RNA decoding: Develop a deep neural network for RNA/DNA decoding from the signals captured by nanopore sequencing devices. (an example of a model and data can be found here).
  • Biosensor design: Machine learning guided search for biosensors. The goal is to design sensor with the materials selected by a machine learning model *.
  • Flow cytometry data analysis: Analyse cell-distributions in order to detect anomalies and pathologies using machine learning.

Machine learning

  • Expanding the feature space into a superposition of polynomials is possible for lower dimensional spaces, however quickly becomes infeasible because of the curse of dimensionality. Consider MNIST digits image dataset, every image in the dataset is represented by a 784-dimensional vector (28x28 greyscale images), the data is extremely sparse, and vectors reside closely to each other. This space can be adequately represented by as little as 2 dimensions. Usually, such compact feature space representations are discovered by finding a low-dimensional manifold often with a corresponding probability function using some form of an autoencoder (e.g. variational autoencoder). The goal of this project is to compactly estimate a low-dimensional representation probability density function using Hermite, Chebyshev, Legendre polynomials or using any other basis. The authors of this paper explain good performance of neural networks by symmetry, locality, compositionality, and polynomial log-probability of our universe.


  • Combinatorics and number theory: Counting integer partitions is a long (almost) standing problem in number theory. There is a fascinating movie devoted to a famous Indian mathematical genius who worked on the problem of counting integer partitions. This video quite well introduces this problem. Recently a solution to this problem was introduced by Ken Ono, however the formula is cumbersome to use. This project aims at tackling this problem from a geometric perspective, with the hope of obtaining either a closed form formula or a “good” approximation that’s easy to compute.

You are on Aboriginal land.

The Australian National University acknowledges, celebrates and pays our respects to the Ngunnawal and Ngambri people of the Canberra region and to all First Nations Australians on whose traditional lands we meet and work, and whose cultures are among the oldest continuing cultures in human history.

bars search times arrow-up