Compass CIO: Chess AI, Neural Network Training, The Voice as a Health Creator, etc.

First autumn month tells us to start with some not-so-serious news in the field of AI and analytics. That’s why we will start with astonishing accusations, made by five-time world chess champion Magnus Carlsen. After the tournament in St.Louis, Missouri, the Norwegian grandmaster made a Twitter statement explaining his defeat to an American competitor and saying Hans Niemann was cheating. Although it is difficult to cheat in chess when rivals sit directly across the board with a number of officials over their shoulders, the modern advances in AI algorithms are here to help.

A chess engine is a software tool which analyzes chess positions and returns what it calculates to be the best move options. Although chess engines have helped players train better, finding their ways against the perfect moves, they also allow some players to cheat more easily. During the game, a player can go to the bathroom and check its smartphone easily. Other complicated cheating involves video recording in a pendant on the player’s neck, which transmits the signal of the next best move to a device in a player’s pocket.

Another sport where we hope not to see any cheating is a football. With the World Cup coming up in Qatar in November, FIFA developed a new player app. All players at the finals will be able to browse their performance data on this new app. It will allow athletes of all 32 teams access and analyze information by making use of input from FIFA's performance analysts, tracking data and physical performance metrics such as distance covered, sprints and positional heat maps.

From sports to arts. Manas Bhatia, an architect and a graphic designer from New Dehli, created a project "AI x Future Cities" where he used AI imaging tool to create images of buildings in the future. In a series of images Midjourney, an AI engine that Bhatia used, depicted futuristic metropolis based on text descriptions such as "futuristic towers", "utopian technology," "symbiotic," and "bioluminescent material". Midjourney needed about 20 minutes to make each of the surreal artworks.

Bhatia could also have used Phraser, a new AI algorithm that employs machine learning to assist users in creating neural network prompts. The key function of Phraser is how it breaks down the creation of a prompt into a few steps, e.g, selecting a style, content type, colour, quality, camera settings, etc. Instead of using keywords, Phraser enables users to search by the meaning of the prompt and look for multiple neural networks in the results. Which is more, users can instantly see how different keywords and styles affect the output due to a function in the prompt editor.

And the last news bit from a non-business world is related to … pigs. A team from the University of the West of England developed a software tool that recognizes pigs’ faces. Identifying pigs based on their unique facial features could enable them to receive personalized food and veterinary care and be easily tracked.

Back in the city streets September saw an inspiring project in San Jose, California. Planning and transportation authorities there started using a cloud-based decision support system for a citywide access and mobility plan. A software tool called Move San Jose uses Big Data and analytics to help city officials determine the next best move. Datasets for the systems are prepared by data analytics startup UrbanLogiq, and include internal data about a number of streets, a number of lanes, how they’re laid out and whether they have sidewalks. By integrating this information in a unified format into the system users can see what the major patterns in data are.

Other serious projects are being carried out in Colombia at Transportadora de Gas Internacional (TGI). This largest natural gas transporter uses digital twins by Emerson to optimize its pipelines. Emerson’s PipelineManager simulation software provides uses with predictive analytics models which run automatically and rely on current operating conditions to let users examine what conditions will be like in six hours, as well as show how actions they perform now will resonate across the network and affect future operations. With this tool TGI identifies an exact window of opportunity for production site shutdowns.

American Airlines, the world’s largest airline, is using data and analytics to minimize disruptions and streamline operations. Having established DataOps framework across the organization AA significantly improved its ability to ingest and consume new data sources in a matter of hours rather than weeks. American Airlines has partnered with Microsoft to use Azure as its preferred cloud platform. Its partnership results in reduced taxi time and real-time information for maintenance staff, ground crews, pilots, flight attendants, and gate agents.

Nikkei partners with S&P Global to develop a new service based on AI to distribute news and data about Asian companies. The companies plan to develop a database centered on organizations in the region and make Nikkei's original content more broadly available. A data platform, scoutAsia, aims to develop a next-generation service to provide information beyond regular financial statements. All with the assistance of the AI.

Panera Bread restaurants in New York are testing AI for taking drive-thru orders. A rise in drive-thru ordering during the COVID pandemic in the US led to long lines of cars, making chains focus on speed of service and order accuracy. That is why Panera Bread uses OpenCity’s voice-ordering technology, Tori. At Panera, customers will interact with Tori when they pull up to the drive-thru speaker thus reducing waiting time.

In the world of science, the AI also proves to be useful. Researchers at Flatiron Institute's Center for Computational Quantum Physics in New York City managed to compress a quantum problem which requires 100,000 equations into a bite-size task of as few as four equations. The AI assistance did not sacrifice the accuracy. In the future, such approach can help in the design of materials with sought-after properties such as superconductivity or utility for clean energy generation.

The US National Institutes of Health has secured $14 million funding for the training of AI that can analyze patients' voices to diagnose and study illness. Twelve research institutions led by the University of South Florida will work on collecting a database of people's voices that will be used to train applications that doctors can use to potentially detect diseases and neurological disorders by examining a person's speech. The Voice as a Biomarker of Health project will focus on such deсeases as voice, neurological and neurodegenerative, mood and psychiatric, respiratory and pediatric voice and speech disorders.

Meta AI researches presented artificial intelligence that can decode words and sentences from brain activity. Using only a few seconds of brain activity data, the AI guesses what a person has heard. That includes many patients in minimally conscious, locked-in or “vegetative states”. Researchers trained the algorithm to detect words and sentences on 56,000 hours of speech recordings from 53 languages. The team applied an AI with this language model to databases from four institutions that included brain activity from 169 volunteers. In these databases, participants listened to various stories and sentences by, for example, Ernest Hemingway or Lewis Carroll, while the people’s brains were scanned. Then with the help of a computational method that accounts for physical differences among actual brains, the team tried to decode what participants heard using just three seconds of brain activity data from each person.

The last but not the least innovation for our health comes from the Department of Electrical Engineering and Computer Science at MIT. A new ML model can detect Parkinson’s disease by reading a person’s breathing patterns. The algorithm mimics the way a human brain works, capable of assessing whether someone has Parkinson’s from their nocturnal breathing.

Vendor news

SAS made its Viya analytics suite available on Microsoft’s Azure Marketplace to help organizations access all features of the suite in on-demand model. Viya, now as a part of the Azure Marketplace, will include an in-app learning center to help support onboarding. The platform includes all SAS applications such as Visual Analytics, Visual Statistics, Visual Data Mining and Machine Learning, Model Manager, Studio Analyst, Information Governance, Econometrics and Optimization.

Altair acquires RapidMiner, one of the leaders in data analytics and machine learning. RapidMiner is said to reinforce Altair data analytics market position in several verticals, with a focus on manufacturing and financial services. RapidMiner's cloud platform will be seamlessly integrated with Altair’s existing tools, such as Knowledge Studio, SmartWorks, and SLC.

That would be all for September. But next month we hope to share more of these exciting news stories.