- In this blog post, write about 100-200 words (1-2 paragraphs). You are asked to re-explain Compton’s “10,000 Bowls of Oatmeal Problem” in your own words. Can you think of a scenario in which this might be a problem, and another in which it isn’t? What are some artistic or technical strategies for overcoming this problem?
Kate Compton’s “10,000 Bowls of Oatmeal Problem” explains the issue of perceptual uniqueness when generating at a large scale. Her metaphor with oatmeal easily displays this idea – noting how even though each bowl is technically different, at such a large scale the differences become much harder to read. Soon, it all begins to look somewhat the same. When generating at a large scale, the products may have many differences but that does not necessarily mean they will look unique. This brings up the idea of perceptual differentiation – which is not the same as perceptual uniqueness. Perceptual differentiation does not go further than making each generated product different from one another. This may be as small as changing the color of a person’s eyes, a detail that may be overlooked when put alongside hundreds of other people. If what you are going for is just some large crowd, this is perfectly fine. However, if this crowd needs to have distinctive figures, perceptual uniqueness takes a step further to make each product perceivably unique to the human eye. But, as Compton explains, making these large generative works have perceptual uniqueness is no easy task. I think this could certainly be a problem if the recipients of a generative work expect the product to be unique each time. It may become boring if they feel like it is getting too repetitive. Maybe a more technical way of overcoming this problem could be to increase the randomness.
Response to Katie Compton’s 10,000 Bowls of Oatmeal Problem:
Oatmeal can often be regarded as a plain and uninteresting meal. For certain people, including me, it is one of my favorite breakfast options. The greatest property of oatmeal is its diversity and ability to mimic your favorite desserts. There is always one distinct element that makes or breaks the oatmeal. For me, it is a creamy flavor. Katie Compton uses an analogy to oatmeal to describe the issue with neglecting a key element to hook the viewer. In the visual sense, each oat could be generated differently every time, but without the ability to perceive the uniqueness of each bowl, the issue rises about whether viewers will truly be able to differentiate each generation. Oatmeal can have many different flavor profiles and combinations, but ultimately, it can become boring if there is no key flavor or element. Katie Compton pushes the question of how artists can produce work that can be perceptually unique.
The “10,000 Bowls of Oatmeal Problem” describes a problem that some algorithms in certain contexts can run into when generating a large amount of content. Though generative programs are technically capable of producing any amount of content, most are unlikely to generate perceptually “interesting”, “novel”, or “unique” results when many artifacts have to be made.
Perceptual uniqueness is not necessary in contexts where “perceptual differentiation” is satisfactory enough – for example, if one was generating the appearance of multiple small details such as grains of sand or waves on water, slight variations to their appearance would be 1) less work for the creator of the generator 2) beneficial to creating a pleasing sense of visual uniformity. However, for larger, more noticeable things, such as say, fish in an aquarium, higher levels of differentiation are needed to prevent an “uncanny valley” effect. One could increase the amount of options that can be generated, or make the generative options as different as possible from each other.
Compton’s “10,000 Bowls of Oatmeal Problem” explains the phenomenon that it can be hard to generate a system that has perceptual differentiation. Although each output may be different, when making art, it is up to human aesthetics to determine the uniqueness of each. An example of when this would be a problem would be if you did 10,000 loads of laundry, each time using different types and amounts of soap because the clothes would turn out looking the same each time. However, if you took 10,000 bananas and dipped them into different types of color dye, they would all come out looking different. A strategy to overcome this problem might be to think about the differentiating aesthetics of your possible outputs before beginning the project so that you will know what to aim for.
Kate Compton’s “One Thousand Bowls of Oatmeal” problem refers to the ways in which one must account for human perception when making a generator that might make things that are technically unique, but not so to the human eye. The author raises the concepts of perceptual differentiation and perceptual uniqueness as two ways by which she explains this “problem”–the former being an object’s differentiation from its iterated predecessor, the latter being a sense of “personality” about an object; does it stand on its own against the whole body of objects? Striving for perceptual uniqueness would lead me to try and create different “archetypes” for whatever I would be generating–perhaps making it so that an object is made either of cool or warm colors, has rounded or sharp edges, or is against a light or dark background.
Kate Compton’s “10,000 Bowls of Oatmeal Problem” has sparked a new view on orientation and possibility based on procedural generation. Compton states, “I can easily generate 10,000 bowls of plain oatmeal, with each oat being in a different position and position and different orientation, and mathematically speaking they will all be completely unique, but the user will likely just see a lot of oatmeal.” From this, you can really see the way procedural generation can be viewed to the viewer and from the generator.
One example that Compton’s blog got me interested in were tattoo shop deals on days such as Friday, the 13th and Halloween where tattoos are given for free but at the expense of not knowing what their tattoo looks like. I think about how a random generator can be part of the process of your daily life and implemented as a “thing” that happened in your life, or a meaningful destined tattoo. But in conclusion, Compton’s blog opened my eyes to the definition of purpose and meaning in our choices, emotions, and environment.
The 10,000 bowls of oatmeal problem is particularly interesting because it is a problem that resides almost completely in the mind of the viewer, and not a truly tangible problem that the computer recognizes. The problem is essentially of relative minuteness in variation when generating something using code. You can create a generator that spits out every variation of a specific ruleset. These variations can be incredibly minute such as a single leaf on a tree being missing, or a single oat in a bowl of oatmeal having a different orientation than it’s other variants. This variation is completely true in the eyes of the computer, but has a lack of significance in the eyes of the viewer. The problem then becomes how to make the permutations both complete and significant enough for them to be interesting. It may not even be a problem for some types of projects, such as simulation, where every slight variance can provide a different result when some type of time changing variable is added, but is particularly unhelpful when trying to create games as sometimes the variations are too slight to have it worth implementing. The problem then becomes how to make visual significance in every variation, such as paring down your parameters to certain key variables.
In the “10,000 Bowls of Oatmeal Problem,” we face a generator that has many mathematically unique outputs, but the outputs are so ever so slightly different (like 10,000 bowls of oatmeal where each oat is uniquely placed and rotated) that they are no longer perceived as unique to the viewer… they are all mundane, monotonous.
Kate Compton says that just achieving perceptual differentiation is important: the viewer should see the differences in your outputs. However, the more difficult achievement is perceptual uniqueness: the viewer should find each of your outputs unique and interesting.
In the oatmeal problem, even if it is possibly clear that each bowl of oatmeal is different, 10,000 bowls of oatmeal won’t all individually feel fascinating and important. We should strive to make our generators produce meaningful work, not a large quantity of work with muddied and unclear differences.
The “10,000 bowls of oatmeal” refers to the issue of generating 10,000 completely unique bowls of oatmeal with each oat in a unique position and orientation but having the user perceive it as all the same. Mathematically, the generative artworks (bowls of oatmeal) are all different, but to the user, all those artworks look the same, losing their uniqueness.
The lack of perceptual uniqueness may become a problem if the user or audience expects a more personal approach to a product or artwork. One example the article mentions is the creatures in “No Man’s Sky,” which generates new creatures every time but nothing “too surprising.” An example where the lack of perceptual uniqueness doesn’t become a problem is when the user wants a similar outcome to what they were expecting. In my previous Looking Outwards assignment, for example, the artist used a mathematical concept called the Voronoi Diagram. This algorithm produced artworks that were physically unique every time but still similar to what users were expecting.
A strategy for overcoming this may be to use an algorithm that replicates patterns within our environment. This would ensure that the artwork/product will generate different permutations every time.
So, regarding the 10,000 Bowls of Oatmeal Problem.
As I understand it, the problem that Kate Compton outlines here is one I actually find a lot when listening to contemporary jazz music. There are some wonderfully brilliant technical players, all of whom can move their fingers in ways that blow my fucking mind. That said, so many of them focus on chipping away at each note and creating incredible variations with dozens of notes, rhythms and patterns that soon it becomes incomprehensible to the listener. Impressive and groundbreaking, but unclear. So in the Oatmeal Issue, we have something similar: hundreds of groundbreaking permutations on a generative system that are all certifiably unique, but at a certain point become functionally the same and unclear. The jazz example lies more on the philosophy of the musician, but it’s what came to mind first for me. Another thing I think of is the new breed of open world adventure games that are pushing the boundaries of how many in-game miles they can make their worlds, without filling it with everything. Technically expansive, yet functionally indistinguishable. An artistic way to overcome this is to invest in populating or iterating on those permutations or that empty space by hand after its finished. Technically speaking, Compton mentioned adding constraints to create enough variability that ‘perceptual uniqueness’ and ‘perceptual differentiation’ are achievable.