Data Science Insider: April 29th, 2022

In This Week’s SuperDataScience Newsletter: AI Weapons Necessitates Ground Rules Between the US and China. Closing Gaps in Cybersecurity Tech Stacks. Using Data Science to Combat Gender Myths. Poisoning Training Data for ML Models. Jumping Robot Could Be Used for Moon Exploration. Cheers, - The SuperDataScience Team P.S. Have friends and colleagues who could benefit from these weekly updates? Send them to [this link]( to subscribe to the Data Science Insider. --------------------------------------------------------------- [AI Weapons Necessitate Ground Rules]( brief: Here at SuperDataScience we have often covered the AI tech war between the United States and China but this opinion piece by Ryan Fedasiuk, adjunct fellow in the Technology and National Security program at the Center for a New American Security, offers a new viewpoint. In it, Fedasiuk looks at the existing relationship between the two colossi’ and argues that the pair need to introduce clear steps to mitigate the existential threats posed by AI accidents. He claims that AIs are exceptionally fragile and prone to operating in a way that’s contrary to their intended purpose, citing the Global Partnership on AI, which has logged “more than 1,200 reports of intelligent systems causing safety, fairness, or other real-world problems.” Due to this, Fedasiuk argues that the US and China need to come together, but states that there is a reluctance on both sides due to a lack of trust. Why this is important: The relationship between the two superstates appears to constantly be teetering on the edge of collapse and the competition over AI often fuels ill-will. This article highlights how important it is for the two to work together for the greater good, in order to avoid catastrophe. [Click here to sign up!]( [Closing Gaps in Cybersecurity Tech Stacks]( brief: Companies are increasingly leaving themselves vulnerable to cyberattacks due to gaps in cybersecurity tech stacks. This is primarily an issue which is brought about through failures in endpoint security and patch management. This results in increasing levels of struggle for CISOs, who are are focusing on how to drive new digital revenue strategies while reducing risk and protecting virtual workforces amidst the various threats. AI is the main driver of this hazard with engineers being recruited by cybercriminal gangs and state-funded Advanced Persistent Threat (APT) networks capable of simultaneously launching assaults across multiple attack vectors becoming increasingly common. It is due to this that the Venture Beat article argues that “Quantifying risks is now table stakes and every cybersecurity vendor in the endpoint security or adjacent markets is introducing self-healing endpoints. Cybersecurity tech stacks need AI to identify how best to thwart advanced attacks today and in the future.” Why this is important: Those with AI skills are currently the most in-demand for nefarious purposes and it may be time to fight fire with fire. This article offers a fascinating read about how AI can be used to close the gaps in cybersecurity tech stacks. [Click here to read on!]( [Using Data Science to Combat Gender Myths]( In brief: It is an unfortunate truth that the majority of us will have heard statements about women in work, such as: “Women aren't ambitious,” “women with kids lose the desire to work,” “women don't like taking risks,” and “women are uncomfortable earning more than their partners.” These are totally false and data science is being used to prove it thanks to Edwina Dunn, the co-founder of global consumer data giant, dunnhumby. Dunn has used her expertise in consumer data to found The Female Lead, a U.K.-based advocacy and education organisation that aims to tackle misinformation about women in the workplace, in part via its survey of more than 100,000 women through its Fulfillment Finder study. By collecting this volume of data, The Female Lead was able to extrapolate clear findings about female attitudes and use them to counter the negative stereotypes which exist and contribute to the gender gap. Why this is important: Women have been fighting for equality in the workplace for a long time but often struggle to prove that the kind of statements listed above are harmful and untrue. Data science can, and is, being used to challenge these negative viewpoints and ultimately make the world a fairer place. [Click here to discover more!]( [Poisoning Training Data for ML Models]( In brief: Researchers at Google, National University of Singapore, Yale-NUS College, and Oregon State University have carried out a study evaluating the risks of attacks where users have access to the ML algorithm and subsequently the data which has been used to train them through their model parameters and predictions. This is then used to “poison” a common pool of training data that is gathered by different users. One of the researchers who carried out the study, Reza Shokri, explained: "The foundation of the adversary method is an inference algorithm, known as membership inference attack, that determines the chance that any arbitrary record has been part of the training set. Inference attacks against ML is a serious data privacy threat because the adversary is a legitimate 'user' of the machine learning system and does not need to break into any system to gain access to sensitive information." Why this is important: By being able to access this data, malicious users have the ability to reconstruct and infer sensitive information included in the training dataset. Such data can be wide-reaching and this TechXplore article goes a long way towards explaining how ML models are essentially being poisoned in order to reconstruct the sensitive information hidden within. [Click here to see the full picture!]( [Jumping Robot Could Be Used for Moon Exploration]( In brief: A robot that can jump more than 100 times its own height has been developed by researchers from the University of California. Lead researcher Elliot Hawkes has suggested that it may be the perfect lunar exploration vehicle, stating: “[The robot] could hop onto the side of an inaccessible cliff or leap into the bottom of a crater, take samples and return to a wheeled rover.” The robot is powered by a carbon fibre sprung skeleton and measures 30cm tall, weighs 30g with the ability to leap to a height of 32.9m. Hawkes' claims are supported by other industry experts with Pietro Valdastri from the University of Leeds claiming that the design would allow the robot to jump to heights that have never been achieved before. He stated: “This technology has a great potential to be integrated into robots designed to rescue people after disasters like earthquakes or tsunamis.” Why this is important: Jumping robots have generally been built based on animal anatomy but their maximum jumping height is limited by the work their muscles can produce in a solitary movement. This differs from the new robot which jumps only when it has stored a large amount of energy. [Click here to find out more!]( [Super Data Science podcast]( this week's [Super Data Science Podcast](, research scientist at Meta AI, Dr. Noam Brown, joins Jon Krohn for our first live, in-person episode in front of a audience! Tune-in to hear Noam discuss his award-winning no-limit poker-playing algorithms, and the real-world implications of his game-playing AI breakthroughs. --------------------------------------------------------------- What is the Data Science Insider? This email is a briefing of the week's most disruptive, interesting, and useful resources curated by the SuperDataScience team for Data Scientists who want to take their careers to the next level. Want to take your data science skills to the next level? Check out the [SuperDataScience platform]( and sign up for membership today! Know someone who would benefit from getting The Data Science Insider? Send them [this link to sign up.]( # # If you wish to stop receiving our emails or change your subscription options, please [Manage Your Subscription]( SuperDataScience Pty Ltd (ABN 91 617 928 131), 15 Macleay Crescent, Pacific Paradise, QLD 4564, Australia