Re: Spamming the Data Space – CLIP, GPT and synthetic data
On Mon, Dec 19, 2022 at 3:55 AM Francis Hunger wrote: > While some may argue that generated text and images will save time and > money for businesses, a data ecological view immediately recognizes a major > problem: AI feeds into AI. To rephrase it: statistical computing feeds into > statistical computing. In using these models and publishing the results > online we are beginning to create a loop of prompts and results, with the > results being fed into the next iteration of the cultural snapshots. That’s > why I call the early cultural snapshots still uncontaminated, and I expect > the next iterations of cultural snapshots will be contaminated. > Francis, thanks for your work, it's always totally interesting. Your argumentation is impeccable and one can easily see how positive feedback loops will form around elements of AI-generated (or perhaps "recombined") images. I agree, this will become untenable, though I'd be interested in your ideas as to why. What kind of effects do you foresee, both on the level of the images themselves and their reception? It's worth considering that similar loops have been in place for decades, in the area of market research, product design and advertising. Now, all of neoclassical economics is based on the concept of "consumer preferences," and discovering what consumers prefer is the official justification for market research; but it's clear that advertising has attempted, and in many cases succeeded, in shaping those preferences over generations. The preferences that people express today are, at least in part, artifacts of past advertising campaigns. Product design in the present reflects the influence of earlier products and associated advertising. One of the primary fields of production in the overdeveloped societies is the field or product range of culture itself, such as movies and TV shows. In the case of TV, feedback loops have been employed systematically since the early 1950s, with the introduction of Nielsen's audiometer, a device that was directly attached to thousands of TVs. Today, TV shows and especially movies are not only used to define the cultural context of successive "generations'' (Gen X, etc). Marketers also use them as surrogates for the memories and affects of those generations. Of course these proxy memories cannot cover the full range of generational experience, but they have the immense advantage, for advertisers, of being fully knowable and therefore, calculable in their effects. The calculations may be more or less bullshit, but they are still employed and acted upon. Blade Runner vividly demonstrated this cultural condition in the early 1980s, through the figure of the replicants with their implanted memories. The intensely targeted production of postmodern culture ensued, and has been carried on since then with the increasingly granular market research of surveillance capitalism, where the calculation of statistically probable behavior becomes a good deal more precise. The effect across the neoliberal period has been, not increasing standardization or authoritarian control, but instead, the rationalized proliferation of customizable products, whose patterns of use and modification, however divergent or "deviant" they may be, are then fed back into the design process. Not only the "quality of the image" seems to degrade in this process. Instead, culture in general seems to degrade, even though it also becomes more inclusive and more diverse at the same time. AI is poised to do a lot of things - but one of them is to further accelerate the continual remaking of generational preferences for the needs of capitalist marketing. Do you think that's right, Francis? What other consequences do you see? And above all, what to do in the face of a seemingly inevitable trend? best, Brian # distributed via : no commercial use without permission #is a moderated mailing list for net criticism, # collaborative text filtering and cultural politics of the nets # more info: http://mx.kein.org/mailman/listinfo/nettime-l # archive: http://www.nettime.org contact: nett...@kein.org # @nettime_bot tweets mail w/ sender unless #ANON is in Subject:
Hypothesis 891. Beyond the Roadblocks
Now available for ordering and/or free download… Hypothesis 891. Beyond the Roadblocks Colectivo Situaciones & MTD Solano Translated by Dina Khorasanee & Liz Mason-Deese Important collective theorization on the meaning of the 2001 Argentinean uprising In 2001 a mass popular uprising overthrew the neoliberal government in Argentina: thousands upon thousands of people, both in organizations and on their own, took to the streets, defying the government’s curfew, shouting “they all must go” until the president was forced to flee by helicopter. The uprising, a response to years of economic and political crisis, cannot be understood without understanding the broader ecology of movements and what Colectivo Situaciones defined as “new social protagonists”: the unemployed blockading highways, neighborhood residents coming together in assemblies, vast segments of the country surviving through alternative currencies and barter networks. This work, translated into English for the first time, brings together the conversations and theorizations of two key participants in that environment: militant research collective Colectivo Situaciones and the Movement of Unemployed Workers of Solano. The encounter and writing in common constituted a formidable experience for all those who participated, bringing to life a novel form of relation between thinking and doing, subject and object of research and political action. Bio: Colectivo Situaciones is a collective of militant researchers based in Buenos Aires. They have participated in numerous grassroots co-research activities with unemployed workers, peasant movements, neighborhood assemblies, and alternative education experiments. PDF available freely online: https://www.minorcompositions.info/?p=1172 Ordering Information: Available direct from Minor Compositions site. Release to the book trade May 1st, 2023. Minor Compositions is a series of interventions & provocations drawing from autonomous politics, avant-garde aesthetics, and the revolutions of everyday life. Release to the commercial book trade May 1st, 2023 # distributed via : no commercial use without permission #is a moderated mailing list for net criticism, # collaborative text filtering and cultural politics of the nets # more info: http://mx.kein.org/mailman/listinfo/nettime-l # archive: http://www.nettime.org contact: nett...@kein.org # @nettime_bot tweets mail w/ sender unless #ANON is in Subject:
Spamming the Data Space – CLIP, GPT and synthetic data
Dear Nettimers, honoring the institutionalized format, I'm posting this speculative text in the hope for comments. best Francis @databasecultures@dair-community.social / www.irmielin.org *** Spamming the Data Space – CLIP, GPT and synthetic data *** ** Introduction ** For the last time in human history the cultural-data space has not been contaminated. In recent years a new technique to acquire knowledge has emerged. Scraping the Internet and extracting information and data has become a new modus for companies and for university researchers in the field of machine learning. One of the currently largest publicly available training data sets to combine images and labels (which shall describe the images content), is Laion-5B, with 5,85 billion image-text pairs (Ilharco, Gabriel et al. 2021).[1] The scope of scraping internet resources has become so all-encompassing, that researcher Eva Cetinic has proposed to call this form ‘cultural snapshot’: “By encoding numerous associations which exist between data items collected at a certain point in time, those models therefore represent synchronic assemblages of cultural snapshots, embedded in a specific technological framework. Metaphorically those models can be considered as some sort of encapsulation of the collective (un)conscious […]” (Cetinic 2022).[2] The important suggestion which Cetinic makes, is that these data collections are temporally anchored. The temporal dimension of these snapshots suggests that digital cultural snapshots taken at different times document different states of (online-)culture. So how will a 2021 snapshot differ from a 2031 cultural snapshot? ** Consequences ** Multi-modal models, like CLIP, trained on large-scale data sets, such as LAION-5B provide the statistical means to generate images from text prompts. In the CLIP Model, pre-trained models merge two embedding spaces, one for images and one for text-descriptions which with mathematical methods get layered together, so that the vectors in the one space, the image domain, align with vectors in the other space, the text domain, assuming there is a similarity between both, and one can translate into the other. In three short examples I’ll discuss some of the consequences of the underlying data for large-scale models from the perspective of cultural snapshots. 1.) Data Bias: A critical discussion of these large-scale multi-modal models for instance, has pointed out how they are culturally skewed and reproduce sexist and racist biases. Researchers Fabian Offert and Thao Phan, for instance, describe how the company Open AI decided not to mitigate the problem of whiteness by changing the model’s underlying data. Instead, Open AI added certain invisible keywords to users’ prompts to have more people of color included, without changing the model. Obviously, the calculations for creating these models or even curating the underlying data are so tremendous that for economic reasons even major problems cannot be corrected in the embedding space itself. Discussing the prevalent ‘whiteness’ in these models further, Offert and Phan suggest to turn to humanities in order to “identify the different technical modes of whiteness at play, and understand the reconceptualization and resurrection of whiteness as a machinic concept” (Offert and Phan 2022, 3).[3] 2.) Uneven spatial distribution: Users of large-scale multi-modal models have tested their limits when generating images. ‘Crungus’, and ‘Loab’ are two examples. ‘Loab’, the image of a women appeared when AI artist Supercomposite looked for the negative of a prompt: “DIGITA PNTICS skyline logo::-1”. Loab appears to be a consistent pixel accumulation, which repeatedly emerges in different configurations and cannot easily be traced back to a single origin.[4] The creator/discoverer of ‘Loab’ felt during intensive testing, that Loab might exist in its own pocket, because it was relatively reproducible, compared to other prompts, as if it was populating a certain statistical region within the larger latent space. Another, similar phenomenon of uneven spatial distribution in latent space is ‘Crungus’, basically a phantasy word which as a prompt nevertheless created results: a snarling, zombie-like figure with shoulder-long hair, which could be part of a horror movie.[5] Both examples demonstrate that the cultural snapshots also contain material which cannot be easily identified or traced back and they demonstrate, how the latent space is an uneven spatial distribution by design. Since the models are built by a process called zero shot learning in difference to for instance the supervised learning used in ImageNet, there are no longer intentional ontologies used in the knowledge creation of these models. The human involvement involves the uncoordinated captioning of images by users online, and the setting up the scraping algorithms and excluding certain domains from being
