Australia’s Approach to AI Governance in Security and Defence

On 27th February 2020, a fully autonomous Australian robot with no human on board used a pre-programmed route with remote supervision to undertake and complete its mission. Australia also won the Silver medal at the 2021 Robot Olympics six months later. In this event, the Australian robot used AI to autonomously explore, map, and discover models representing lost or injured people, suspicious backpacks, or phones while navigating harsh conditions. Clearly, Australia is at the forefront of robotics, autonomous systems, and AI research and development. Australia intends to use AI for security and defense applications and is also developing governance structures in line with Australian values, standards, ethical and legal frameworks. 

Kate Devitt & Damian Copeland have discussed this in their research paper titled “Australia’s Approach to AI Governance in Security and Defence” which forms the basis of the following text.

Importance of this Research

AI has enormous potential in Security and Defence applications. AI applications could threaten human rights in warfare and pose ethical and legal risks. This research paper outlines the approach of Australia to make use of AI in Security and Defense. 

The research paper also outlines that Australia’s Department of Defense recognizes AI as a priority for future developments. The research paper also outlined the definition of AI and explained the need for Artificial Intelligence for Australian Defense. 

Use of AI in Defense

As per the Australian government, Australia focuses on AI to build robotics capability, autonomous systems, precision-guided munitions, hypersonic weapons, and integrated air and missile defense systems, space and information warfare, and cyber capabilities. 

AI Governance in Australia  


AI Ethics Principles 


The research paper also discusses in detail about

  • Standards Australia’s Artificial Intelligence Standards Roadmap: Making Australias Voice Heard, Human Rights, and Australia’s action plan for AI,
  • AI governance in Defense and Ethical AI statements across the Australian Navy, Army & Airforce,
  • Human governance in Defense,
  • Ethics in Australian Cybersecurity and Intelligence; and the Framework for Ethical AI in Defence, and
  • Defense Data Strategy

Concerning the use of AI, it is helpful to establish 

  • Responsibility: Who is responsible for AI?
  • Governance – how is AI-controlled?
  • Trust – how can AI be trusted?
  • Law: How can AI be used lawfully?
  • Traceability: How are the actions of AI recorded?


AI has tremendous potential for security and defense applications. Nations can use it to counter-terrorism, for border conflicts, and help them save armed personnel’s lives. Australia has identified the criticality of AI in defense and tried to build a framework around its application

In the words of the researchers

Australia is a leading AI nation with strong allies and partnerships. Australia has prioritised the development of robotics, AI, and autonomous systems to develop sovereign capability for the military. Australia commits to Article 36 reviews of all new means and method of warfare to ensure weapons and weapons systems are operated within acceptable systems of control. Additionally, Australia has undergone significant reviews of the risks of AI to human rights and within intelligence organisations and has committed to producing ethics guidelines and frameworks in Security and Defence (Department of Defence 2021a; Attorney-Generals Department 2020). Australia is committed to OECDs values-based principles for the responsible stewardship of trustworthy AI as well as adopting a set of National AI ethics principles. While Australia has not adopted an AI governance framework specifically for Defence; Defence Science has published A Method for Ethical AI in Defence(MEAID) technical report which includes a framework and pragmatic tools for managing ethical and legal risks for military applications of AI.

Source: Kate Devitt & Damian Copeland’s “Australia’s Approach to AI Governance in Security and Defence” 

Taxable Stock Trading with Deep Reinforcement Learning

Can machines beat humans in stock market trading? 

Machine Learning has been an area of great interest for traders because of the insane amount of money that could be made from the stock market. Arguably, Deep Reinforcement Learning can outperform human traders. However, most of the Deep Reinforcement Learning models built to date only consider the stock’s selling price minus the buying price to maximize profit. In a real-world scenario, transaction costs for buying, selling, and taxes affect the actual profit made by the investor. 

Shan Huang has discussed this approach in his research paper titled “Taxable Stock Trading with Deep Reinforcement Learning,” which forms the basis of the following text.

Importance of this research 

The researchers have demonstrated that tax ignorance could induce more than a 62% loss on the average portfolio returns. This research paper by Shan Huang aims to maximize net portfolio returns using Deep Reinforcement Learning while considering transaction and tax calculations. 

About the research 

The research paper mentions the mathematical details of the model built for Deep Reinforcement Learning. 


  • The data set includes SPY’s (Exchange Traded Fund consisting of S&P’s top 500 companies) daily closed price and volumes from 13th November 2008 to 13th November 2018.
  • Minimum time step dt = 1 representing one trading day, and hence the total trading days per year is 252.
  • The tax regime is as mentioned below (in the US)
    • When a stock is held for more than one year, a 15% tax rate is applicable for capital gains.
    • When a stock is held for less than one year, a 25% tax rate is applicable for capital gains.
    • Investors can get a tax rebate to offset their capital gain for losses.
  • The length of the trading period is set to be five years.
  • A transaction cost of 0.1% is also included for more realistic calculations.

Research Data

The researchers have created a new OpenAI Gym environment where the observation in each timestep is SPY’s daily closed price, trading volume, averaged-basis, and average holding period. 

About the Model

In the words of the researchers

To represent the policy, we use the same default neural network architecture as PPO with fixed-length trajectory segments, which was a fully-connected MLP with two hidden layers of 64 and 64 tanh units respectively. The final output layer has a linear activation. policy and value function are estimated through separated network. The number of steps of interaction (state-action pairs) for the agent and the environment in each epoch is 5000 and the number of epochs is 50. The hyperparameter for clipping in the policy objective is chosen to be 0.2 and the GAE-Lambda is 0.97. The learning rate for policy and value function optimizer is 0.001 and 0.0003 respectively. If tax is not included in the model, the average expected return is 0.44 which seems quite promising. This considerable return is the result of exploiting price trending and frequently adjusting holding positions correspondingly, similar as the results of other AI platforms. However, this is not compelling since tax is heavily charged in a taxable year. Rather than ignoring taxes, the learning of stock trading should consider the effect of tax costs. We use PPO to train the stock trading policy in the environment with tax costs. The optimal stock trading policy in the model with taxes can achieve 0.13 average returns. To illustrate the suboptimality of the policy trained in the model without considering taxes, we apply this trained policy in the environment with tax costs, the average expected return drops to only 0.05.

Research Result

  • The optimal stock trading policy with taxes can achieve 13% average returns in the model.
  • Trained policy in the environment with tax costs, the average expected return drops to only 5%.
  • This reduction is equal to 62%, which means that the returns shrunk by 62% upon consideration of taxes. 



Investment in public markets could be a good way to invest in the world’s top companies. Public markets could be a fantastic way to invest the money as the return are, on average higher than the rates offered by the banks and other investment options. Trading stocks by reinforcement learning can guide and help agents increase their portfolio returns. Deep Reinforcement Learning Models built for stock market investment sometimes neglect the tax rates, which can massively impact the overall returns. This research paper by Shan Huang is an attempt to integrate tax calculations in the net return calculations with Deep Reinforcement Learning. The objective is to maximize net returns (after subtracting taxes) for stock investments which are the actual returns for an investor.

Source: Shan Huang’s “Taxable Stock Trading with Deep Reinforcement Learning 

The Study of Machine Learning Models in Predicting the Intention of Adolescents to Smoke Cigarettes

Four hundred eighty thousand deaths are recorded every year in the US alone due to smoking. Also, as per research, life expectancy for smokers is atleast ten years shorter than for non-smokers. In recent years, the use of e-cigarettes has been increasing rapidly for adolescents. Sometimes, e-cigarettes act as a gateway to cigarette use, which could be severe. This makes it important to predict the probability of adolescents smoking cigarettes in the future. 

Seung Joon Nam, Han Min Kim, Thomas Kang, and Cheol Young Park have discussed this in their research paper titled “The Study of Machine Learning Models in Predicting the Intention of Adolescents to Smoke Cigarettes”, which forms the basis of the following text.

Importance of this research

Consuming nicotine can cause cancer (lung cancer), cardiovascular and metabolic diseases, respiratory diseases, and perinatal diseases. These conditions are fatal for individuals. If an ML algorithm can predict the probability of individuals smoking cigarettes, efforts can be made to educate these individuals about the ill effects of smoking. This counseling could help adolescents avoid the ill effects of cigarette smoking. Thus, prediction can directly help individuals stay away from the path of smoking cigarettes and help save lives!

Research Objective

The main aim of this research is to

  • Find the best-fitting model to predict smoking intention for individuals
  • Create a website to help adolescents prevent e-cigarette.

About the Research 

The researchers evaluated different models such as Decision Tree, Gaussian NB, Logistic Regression, Random Forest, and Gradient Boosting to predict the accuracy of ML models for accurately predicting the intention of Adolescents to smoke cigarettes. 

Research Result

Based on the experiments done by the researchers, they found the Gradient Boosting to be the most accurate way to predict smoking tendency in the future. The researchers have also published for an anti-smoking campaign for teenagers. 


While healthcare has made a lot of progress in recent years, the number of medical conditions affecting teenagers has only increased. Often, these diseases are of our own making caused by our lifestyle choices. This research paper attempts to identify adolescents who are more likely to smoke and help adolescents make informed and healthier lifestyle choices. In the words of the researchers

E-cigarette use has increased among adolescents. This is a worldwide problem, because it has been stated in many researches mentioned in the introduction that e-cigarette use can cause future use of cigarettes. Since e-cigarette is a recent rising issue, there is little research done on this topic, compared to smoking cigarettes. Even among the researches done, there is a lack of researches implementing prediction models, which are more practical in preventing adolescents from using (e-)cigarettes. Thus, we researched using the 2018 NYTS data and developed multiple prediction models to predict a adolescents intention to smoke cigarette. The most accurate prediction model was Gradient Boosting Classifier with an overall accuracy of 93%. This model was applied in the website we designed to allow the public to input their information in respect to tobacco products, including e-cigarette, cigarette, and cigar. With this information, the algorithm can predict the respondees probability of future of smoking. This will help the public become more aware about certain factors in their lives and be attentive about their drug use or how their environment can affect their intention to smoke cigarettes. Further research could include a wider range of ages, since our research is mainly focused on adolescents rather than adults. In order to improve the accuracy of the prediction model, it is essential to increase the amount of data or choose better, more fitting, variables.

Source: Seung Joon Nam, Han Min Kim, Thomas Kang and Cheol Young Park’s “The Study of Machine Learning Models in Predicting the Intention of Adolescents to Smoke Cigarettes”