AhIn October, singers, songwriters and music creators uploaded 100,000 new songs to streaming services like Spotify every day. That’s too much music. The reality that even in a thousand lifetimes someone could probably hear all of it, alternative or not. Whether you’re into Japanese noise, Russian hardcore, Senegalese Afro-house, Swedish doom metal, or Bay Area hip-hop, the sheer scale of listening options available is paralyzing. This is the monumental problem that data his scientist Glenn McDonald is trying to solve. In the excerpt below, Computing Taste: Algorithms and Maker Music Recommendationsauthor and anthropologist at Tufts University, Nick Seaver, explores McDonald’s unique landscape-based methodology, surfacing every truck you didn’t know you couldn’t live without.
Reprinted with permission from Computing Taste: Algorithms and Maker Music Recommendations By Nick Seaver, published by The University of Chicago Press. © 2022 University of Chicago. All rights reserved.
world of music
“We are now at the dawn of an era of infinite music,” said the data alchemist from beneath the Space Needle. Glenn McDonald chose the title himself, preferring the esoteric relevance of “alchemy” to the now commonplace “data science.” As he explained from the stage, his job was to “use math and typing and computers to help people understand and discover music.”
McDonald has put his alchemy into practice for music streaming service Spotify. So he worked on turning the big data foundation (listener interaction logs, digital audio snippets of his files, and whatever else he could get his hands on) into precious gold. . It has the potential to attract and retain paying customers. The mystical power of McDonald’s alchemy lay in the way ordinary data, when properly processed, seemed to transform from thin traces of interaction into thick cultural significance.
It was 2014 and McDonald’s was at a pop conference. The Pop Conference is an annual gathering of music critics and academics held in a pile of crumpled Frank Gehry-designed buildings in the heart of Seattle. I’m on the other side of the country and followed you online. The theme of that year’s conference was “Music and Mobility,” and McDonald began by recounting his personal musical journey while playing samples. “When I was a kid, I discovered music by sitting still and waiting,” he began. As a child, he used to listen to folk music at home that his parents played on the stereo. But as he grew up, his listening broadened. Car radios offered heavy metal and new wave. The Internet has revealed a world of new and obscure genres to explore. A passive observer of the music he happened to pass by, he eventually came to measure his progress in his own life by an ever-expanding musical horizon. McDonald has managed to turn this passion into his profession, making what others have called “the world of music” more accessible than ever with his Streaming On Demand service. Helped me explore.
Elsewhere, McDonald (2013) describes the world of music as a landscape. Over the years, Australian hip-hop, Hungarian pop, micro-house, Viking, like metal, he has reconstructed the world of music in methodically and idiosyncratically altered miniatures. ”
Those who travel through the world of music will encounter familiarity and surprises, sounds they never imagined, songs they longed for. McDonald’s marveled at this new feature that allowed them to hear music from all over the world, including Scotland, Australia and Malawi. “The perfect music for you may come from the other side of the world,” he said, but this didn’t matter. “Music has a teleporter.” It provided musical mobility, allowing listeners to instantly travel through the world of music.
However, he suggested that the scale of this world can be overwhelming and difficult to navigate. “To make this new world really appreciable, we need to find ways to map this space and build machines that guide us along interesting paths,” McDonald said. Recommendation systems offered by companies like Spotify were machines. MacDonald’s recent work focused on maps, or, as he explained in another lecture, “the writhing, proliferating, insatiably expanding global musical information space. A sort of thin layer of vaguely intelligible order over the beast.”
Very poetic as his wording may have been, McDonald was expressing an understanding of musical diversity that is widely shared among the creators of music recommendations. Music exists in a kind of space. In a sense, the space is like a normal landscape, and you may encounter new things while walking. But in another way this space is very strange. Behind the valleys and hills lies a writhing, surging beast, ever-growing, linking the dots in space, connecting them to infinity. The music space naturally looks like the mountains seen from the top of the Space Needle. But it can also seem like a hodgepodge of underlying artificial topologies. Organic and intuitive. It’s technical and chaotic.
Spatial metaphors provide a primary language for thinking about differences among music recommendation creators, much like machine learning and, more generally, Western cultural differences. Within these contexts, it’s easy to imagine certain similarities coming together. over hereand many more clustered that sideGestures often evoke musical spaces in conversations with engineers. Gestures envelop the speaker in an imaginary environment made up of short pinches and hand waves in the air. One genre is on the left and another on the right. The whiteboards and windows scattered around the office have musical spaces rendered in two dimensions, with arrays of points clustered across the plane.
Music Space finds similar music nearby. When you are in such a space, you should be surrounded by your favorite music. To find more of it, just look around and move. In the music space, genres are like regions, playlists are like trails, and tastes are like drifting archipelago territories. There may be.
But despite how familiar it is, a space like this is strange. Similarities are everywhere, and points that seemed far apart can suddenly be adjacent. If you ask, all of these spatial representations are not 2D or his 3D, but rather a much more complex reduction of potentially thousands of dimensions of space. This is McDonald’s Information Space Beast, a mathematical abstraction that extends human spatial intuition beyond its breaking point.
Such spaces, commonly called “similar spaces”, are the iconic terrains on which most machine learning works. To classify data points or recommend items, machine learning systems typically arrange them in space, collect them into clusters, measure the distance between them, and draw boundaries between them. increase. As cultural theorist Adrian Mackenzie (2017, 63) has argued, machine learning “represents all differences as distance and direction of travel.” Thus, musical space is in one sense an informal metaphor (a landscape of musical variations), but in another sense a highly technical formal object (the mathematical basis of algorithmic recommendations).
Spatial understanding of data movement through technical infrastructure and everyday speech. They are both figurative forms of expression and concrete computational practices. In other words, “space” here refers to both formalism — a restricted technical concept that facilitates precision through abstraction — and what anthropologist Stefan Helmreich (2016, 468) calls informalism — formalism. Both — what I would call a less disciplined metaphor that moves with technology. In practice, it is often difficult or impossible to separate technical specificity from figurative accompaniment. When creators of music recommendations talk about space, they immediately speak figuratively and technically.
For many critics, this “geometric rationality” (Blanke 2018) of machine learning is an abomination to “culture” itself. It quantifies quality, rationalizes passion, extracts cultural objects from everyday social contexts and relocates them to sterile and isolated places. computational grid. For example, mainstream anthropology has long been opposed to such a formalism that seems to lack the depth, sensitivity, or appropriateness to the lived experiences we seek through ethnography. has been defined. As political theorists Louise Amoore and Volha Piotuk (2015, 361) suggest, such an analysis “reduces heterogeneous organisms and data into a homogeneous space of computation”.
To use the terminology of geographer Henri Lefebvre (1992), analogous space is a clear example of ‘abstract space’. This is a kind of representation space in which everything is measurable, quantifiable and controlled by the central authority of capital services. Media theorist Robert Prey (2015, 16) applies Lefebvre’s framework to streaming music, and people like McDonald’s (“data analysts, programmers, engineers”) focus primarily on computational and measurement abstractions. It suggests that he is interested in a conceptual and conceived space. In Lefebvrian’s view, the envisioned space is parasitic on the social, living space that his Prey associates with listeners who resist and reinterpret the work of the technician. The expansion of abstract space under capitalism portends, within this framework, “the devastating conquest of the living by the imagined” (Wilson 2013).
But to those who work with it, musical space, even the most mathematical, does not feel like a sterile grid. I will not stay. In the course of their training, they learn to experience musical space as normal and livable, despite its underlying strangeness. as heterogeneous as engineering objects in higher dimensions. To exploit an often problematic distinction with cultural geography, they treat ‘space’ like ‘place’, as if an abstract uniform grid were a kind of livable local environment. to handle.
A similarity space is the result of many decisions. They are by no means “natural,” and people like McDonald recognize that their choices can deeply reorganize them. It helps to make patterns in cultural data feel real. The confusion between map and territory, malleable representation and objective topography, is for those interested in producing objective knowledge and in explaining their subjective influence on processes. productive for These spatial understandings change the meaning of musical concepts such as genre and social phenomena such as taste, and express them as a form of clustering.
All products recommended by Engadget are selected by an editorial team independent of the parent company. Some stories contain affiliate links. When you purchase something through one of these links, we may earn an affiliate commission. All prices are correct at the time of publication.