Data Science Insider: December 24th, 2021

In This Week’s SuperDataScience Newsletter: Predictions for AI In 2022. New Theory of Consciousness in Humans, Animals and AI. Face Detection in Untrained Deep Neural Networks Detected. A Call for Algorithmic Reparation. AI Art App Offers a Glimpse at the Future of Synthetic Media. 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. --------------------------------------------------------------- [Predictions for AI In 2022]( brief: AI is a game of predictions. Every day, ML models in nearly every industry make important predictions to inform everything from actuarial decisions to zinc futures. With that in mind, it’s only fitting that the product leaders of a company that helps humans better understand AI go on the record with some predictions for the year ahead, which is where these fifth annual Neural predictions for 2022 come into play. If the events of 2021 make anything clear, it’s that we live in a post-COVID world where anything can happen, and once-outlier events are becoming common. Today’s businesses must not only navigate evolving consumer tastes and record demand but also a complex supply chain where inflation, delays, shortages, and other unforeseen factors are increasingly common. It’s a tough environment for any team building and deploying ML models. It’s also a time of incredible opportunity to make an impact. Why this is important: In our industry, we must always look ahead to the future. By looking at expert predictions we can consider our own views on what 2022 may bring and think about the role we wish to play in the coming year’s developments. [Click here to sign up!]( [New Theory of Consciousness in Humans, Animals and AI]( brief: Two researchers at Ruhr-Universität Bochum (RUB) have come up with a new theory of consciousness. They have long been exploring the nature of consciousness, the question of how and where the brain generates consciousness, and whether animals also have it. The new concept describes consciousness as a state that is tied to complex cognitive operations—and not as a passive basic state that automatically prevails when we are awake. The complex cognitive operations that are associated with consciousness are applied to mental representations that are maintained and processed. Professor Armin Zlomuzica from the Behavioral and Clinical Neuroscience research group at RUB and Professor Ekrem Dere, describe their theory in the journal Behavioural Brain Research: "The hypotheses underlying our platform theory of consciousness can be tested in experimental studies. Thus, the process of consciousness can be explored in humans and animals or even in the context of artificial intelligence." Why this is important: As the authors state: "To what extent an artificial intelligence which is capable of independently solving a new and complex problem for which it has no predefined solution algorithm can likewise be considered conscious has to be tested." However, this theory has fascinating ramifications for AI and should be of interest to all data scientists. [Click here to read on!]( [Face Detection in Untrained Deep Neural Networks Detected]( In brief: Researchers have found that higher visual cognitive functions can arise spontaneously in untrained neural networks. A research team from The Korea Advanced Institute of Science and Technology has shown that visual selectivity of facial images can arise even in completely untrained deep neural networks. This new finding has provided revelatory insights into mechanisms underlying the development of cognitive functions in biological and artificial neural networks, also making a significant impact on our understanding of the origin of early brain functions. The study demonstrates that neuronal activities selective to facial images are observed in randomly initialised deep neural networks in the complete absence of learning and that they show the characteristics of those observed in biological brains. Using a model neural network that captures properties of the ventral stream of the visual cortex, the research team found that face-selectivity can emerge spontaneously from random feedforward wirings in untrained deep neural networks. Why this is important: The ability to identify and recognize faces is a crucial function for social behavior, and this ability is thought to originate from neuronal tuning at the single or multi-neuronal level. Neurons that selectively respond to faces are observed in young animals of various species, and this raises intense debate whether face-selective neurons can arise innately in the brain or if they require visual experience. [Click here to discover more!]( [A Call for Algorithmic Reparation]( In brief: In these SuperDataScience newsletters, we have frequently discussed issues of bias and ways in which we can seek to overcome our unconscious prejudices, which have come to form part of the technology created. This article by Wired highlights the fact that sociologists and computer science researchers are now saying that the builders and deployers of AI models should consider race more explicitly, by leaning on concepts such as critical race theory and intersectionality and apply these learnings by introducing “algorithmic reparation.” According to the academics who first coined the term: “reparative algorithms aim to name, unmask, and undo allocative and representational harms as they materialise in sociotechnical form.” Supporters of algorithmic reparation suggest that we should consider how to ethically collect data about people and propose considering not just whether the performance of an AI model is deemed fair or good but whether it shifts power. Why this is important: The suggestions confront the core issues at the root of inequality. They echo earlier recommendations by former Google AI researcher Timnit Gebru, who encouraged ML practitioners in a 2019 paper to consider how archivists and library sciences dealt with issues involving ethics, inclusivity, and power. [Click here to see the full picture!]( [AI Art App Offers a Glimpse at the Future of Synthetic Media]( In brief: If you’ve been hanging out on Twitter lately, then you’ve probably noticed a trend for rather strange-looking AI-generated images. These pictures have been generated using a new app called Dream, which lets anyone create “AI-powered paintings” by simply typing a brief description of what they want to see. The resulting artwork has its own particular aesthetic. No matter what you type, the app will generate something that is visually compelling and that matches your prompt in often surprisingly relevant ways. This sort of AI-generated artwork is not new, but it is becoming higher quality and more accessible. Past examples of these sorts of text-to-image models have included research-orientated programs like DALL-E and VQGAN+CLIP, as well as more specialized commercial projects like Artbreeder. Generally, programs like these are trained on vision datasets — huge libraries of images that are tagged based on objects and scenery. Why this is important: With tools such as these, the AI art scene has exploded in recent years, with practitioners creating everything from lifelike Roman emperors to infinite waifus. The Dream app takes things a step further with its speed, quality, and accessibility. [Click here to find out more!]( [Super Data Science podcast]( this week's [Super Data Science Podcast](, Dr. Brett Tully joins to share his interesting work at Nearmap with AI output systems and his previous academic work in fascinating applied biomedical issues, and his research into nuclear fusion. --------------------------------------------------------------- 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