Hopefully during the latest holiday season you’ve heard the Beatles “last” song ‘Now and Then’ released in November 2023 on the newly remastered ‘Red and Blue’ greatest hits collections.
What’s most interesting about these releases is how they showcase machine learning and artificial intelligence (AI) techniques that provide laymen a glimpse into the broad AI revolution not only in music but also across many industries.
In ‘Now and Then,’ newly developed computer programs were used to separate John Lennon’s voice from a poor-quality cassette recording from 1978. But the story really begins with award-winning director Peter Jackson and his work on the 2021 Beatle’s documentary Get Back. In this film that chronicled the Beatles’ final recording sessions, Jackson showcased AI-powered software that he called “Mal” (Machine Audio Learning) to clean and enhance 50+-year-old audio recordings captured at the Apple Records studio. This included removing background noise, hiss, and hum — significantly improving clarity and intelligibility of crowded rooms with many people talking at the same time.
How did Jackson do it? His team developed AI algorithms to identify the voices of John, Paul, George, Ringo, and others in the film. In turn, the programs isolated their respective voices and created separate tracks to clarify each versus listening to everything in the recorded room — revealing previously unheard conversations. Moreover, machine learning algorithms analyzed color patterns from old still photos and Beatles’ film footage to accurately colorize black-and-white sequences, creating a seamless and visually stunning experience.
Jackson also used upscaling technologies to increase the footage’s resolution, enhancing details and providing much sharper visuals. Moreover, other programs — such as facial and gait recognition software — sifted through and analyzed approximately 100 hours of footage to identify key moments and conversations, streamline the editing process, and helped Jackson locate specific events.
Not only does Jackson’s cool documentary and the new Beatles song showcase some of what AI can do, it also provides glimpses into other potential uses. For example, similar AI and machine learning technologies are revolutionizing the hearing aid industry. AI algorithms can analyze individual hearing loss patterns and preferences to create custom personalized sound profiles, ensuring optimal sound clarity and comfort for each user. Machine learning programs effectively identify and suppress background noise, allowing users to laser-focus and clarify conversations even in challenging environments. Like in Get Back and ‘Now and Then,’ AI can reduce surrounding clatter, significantly improving speech intelligibility for hearing aid users.
Over time as computing power increases, hearing aid algorithms could automatically detect a user’s environment and adjust sound profiles accordingly, optimizing for a crowded restaurant, noisy street, or quiet room. With wearables and remote technology, AI-powered platforms also will enable hearing specialists to monitor and adjust hearing aid settings to optimize and improve performance for users.
Jackson’s magic with the Beatles should inspire us to investigate how to incorporate AI into our own personal and business lives. Isolating specific data and finding patterns is a very common practice in AI, particularly in machine learning. It involves filtering a subset of data from a larger dataset that is relevant to a specific question or problem. This subset is then analyzed to identify patterns, trends, or correlations — and this can be done in any field from finance to healthcare.
For instance, imagine a dataset of patient health records. An AI program could be used to search and isolate information related to a specific condition, like heart disease. It then applies algorithms to this subset to find patterns in symptoms, diet, genetics, lifestyle factors, and treatment efficacy. In turn, such findings can provide doctors and hospitals with insights that can help in decision-making, predicting possible future trends, and personalizing treatments for patients.
This AI approach is fundamental and applicable to virtually all industries where understanding specific segments of data is crucial. Take marketing. Instead of making hunches about what customers want, AI can leverage data and machine learning to monitor purchasing patterns and optimize marketing efforts for more informed business decisions. This results in increased customer satisfaction and loyalty, better sales and productivity, and enhanced overall efficiency.
For finance professionals, AI can be used for everything from gathering market intelligence, summarizing text/video/spreadsheet information, producing commentary and presentations, and providing real-time regulatory checks and fraud detection. For investing, AI allows both retail and institutional investors to filter thousands of investment possibilities and create portfolios that meet their risk/reward criteria, create quantitative strategies, and even execute trades much faster than humanly possible.
So here’s my suggestion for a 2024 New Year’s resolution: learn more about AI and how it can be utilized for your specific firm or job. AI has the potential not only to boost your productivity but also to uncover information that may have been difficult without computer-assisted help.