Data Science Insider: February 25th, 2022

In This Week’s SuperDataScience Newsletter: Meta Outlines AI Plans. AI and Wildlife Conservation. QAnon Founders May Be Identified Thanks to ML. Survivorship Bias in Data Science. AI Tool Detects Global Fashion Trends but Raises Privacy Concerns. 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. --------------------------------------------------------------- [Meta Outlines Plans for Universal Speech Translation]( brief: Meta has announced an ambitious new AI research project to create translation software that works for “everyone in the world.” The company says that although commonly spoken languages are well catered to by current translation tools, many aren’t. Often, these under-served languages do not have easily accessible corpora of written text that are needed to train AI systems or sometimes have no standardised writing system at all. Meta says it wants to overcome these challenges by deploying new ML techniques in two specific areas. The first focus, dubbed No Language Left Behind, will concentrate on building AI models that can learn to translate language using fewer training examples. The second, Universal Speech Translator, will aim to build systems that directly translate speech in real-time from one language to another without the need for a written component to serve as an intermediary (a common technique for many translation apps). Why this is important: The project was announced as part of an event focusing on the broad range of benefits Meta believes AI can offer the company’s metaverse plans. This includes testing an AI system, called builder bot, that lets people build parts of virtual worlds by describing them. These aspirations are impressive. [Click here to sign up!]( [AI and Wildlife Conservation]( brief: At a time when conventional methods of wildlife conservation are failing, conservationists are turning to AI for more innovative tech solutions to protect species from the edge of extinction. According to a report by Wildlabs.net, AI has turned out to be one of the top three emerging technologies in advancing conservation over the next 10 years along with environmental DNA (eDNA) and genomics, and networked sensors. With camera traps and satellite images, AI can easily detect an animal call or identify a rare species, doing the job of hundreds of people at a time and reducing the labour required to collect data. The application of technology in wildlife and biodiversity preservation helps researchers and rangers, as well as protecting threatened species by using algorithms to sort the huge amount of data, enabling a better understanding of the behaviour of animals such as their foraging routes, reproduction patterns, and hunting habits. Why this is important: At present, there are five projects working with AI that are contributing to our understanding of biodiversity and species. Zambia’s Kafue National Park: Zambia’s department of national parks and wildlife, the Game Rangers International (GRI), and other partners have started the Connected Conservation Initiative that uses AI to enhance conventional anti-poaching efforts. As the technology evolves, this number is bound to increase. [Click here to read on!]( [QAnon Founders May Be Identified Thanks to ML]( In brief: Using ML, two separate teams of forensic linguists have claimed that they have discovered the origins of the QAnon mass political movement. Often referred to as a cult, QAnon has spread various conspiracy theories. Now, two teams of Swiss and French computer scientists have identified Paul Furber, a South African software developer, as the first to post conspiracy theories using the Qanon handle before Arizona congressional candidate Ron Watkins took over the handle and started posting theories himself. The Swiss team used ML to study QAnon's posts by breaking them down into patterns of three-character sequences. Then these sequences were checked against the posts to see how often the patterns repeated. The French team fed the posts through AI to look for identifiable patterns in writing. Both teams limited the scope of the investigation to social media posts made from the handle. Why this is important: The two different methodologies both reached the same conclusion, although Furber and Watkins both deny the claims. It is hoped that the unmasking may result in a lessening of QAnon’s hold over people and politics but it is feared that genuine whistleblowers may be more reluctant to post online. [Click here to discover more!]( [Survivorship Bias in Data Science]( In brief: Survivorship bias is a type of sample selection bias that occurs when a data set only considers “surviving” or existing observations and fails to consider observations that have already ceased to exist. Generally speaking, survivorship bias tends to create conclusions that are overly optimistic, and those that may not be representative of real-life environments. The bias occurs because the “surviving” observations often tend to have survived due to their stronger-than-average resilience to difficult conditions, and leaves out other observations that have ceased to exist as a result of such conditions. In this article in Forbes by the Head of Data and AI at CentralNic Group PLC., there is an exploration of how survivorship bias applies to data scientists and an examination of the importance of having a robust understanding of the quality of data we are working with rather than merely concentrating on high volumes. Why this is important: As data scientists, we need to appreciate the different types of bias we may bring to our data or that already exists within the numbers. This article gives useful tips on how to avoid bringing survivorship bias into our work. [Click here to see the full picture!]( [AI Tool Detects Global Fashion Trends]( In brief: Researchers at Cornell University developed an AI tool that scans millions of publicly available photos to effectively identify fashion trends around the world, as well as traditions and events with signature styles. The GeoStyle tool analyses public Instagram and Flickr photos to map trends using computer vision and neural networks. The models help researchers understand existing trends in specific cities, and across the world as a whole. GeoStyle's trend forecasts are up to 20% more accurate than previous methods. The team also created a visualizer that allows users to view the popularity of a certain attribute—such as a pattern, hat, or color—by city, over time. Cornell researcher Kavita Bala explained: "One of our follow-ups from this work is improving the technology so that if you add a little expert information, you can improve the recognition and get an even finer-grained understanding." Why this is important: The researchers at Cornell claim that how people dress in an area can tell you a lot about what happens there, or is happening at a particular time, and knowing the fashion sense of an area can be a useful tool for visitors, new residents, and even anthropologists. However, the project’s connection with Amazon and Meta raises questions around the application of user data, often without explicit permission. [Click here to find out more!]( [Super Data Science podcast]( this week's [Super Data Science Podcast](, the gifted author and software engineer, Wah Loon Keng, joins us to dive deep into the world of RL and discusses its history, limitations, modern industrial applications, and so much more. --------------------------------------------------------------- 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