Algorithmic Superstructuring: Aesthetic Regime Of .
Transformations issue 34 (2020)www.transformationsjournal.orgAlgorithmic Superstructuring: aesthetic regimeof algorithmic governanceISSN 1444-3775AUTHOR BIOAlex Anikina (b. Kolomna,Russia) is a media artist,researcher, film-maker andcurator, currently completing herPhD at Goldsmiths, University ofLondon, and teaching at LondonSouth Bank University. Herdissertation investigatesprocedural films as researchmedia art practice. Her work hasbeen shown internationally,recently at Sanatorium gallery,Istanbul, Krasnoyarsk MuseumBiennale and VI MoscowInternational Biennale for YoungArt. She co-edited Cosmic Shift:Russian Contemporary Art Writing(ZED Books, London, 2017). In2018 she co-curated IMPAKTmedia art festival (Utrecht,Netherlands) with the themeAlgorithmic Superstructures.Alex AnikinaABSTRACTIn this paper I suggest the idea of algorithmic superstructuring as a way toexplore aesthetic regimes of algorithmic governance, drawing on work ofJacques Rancière, Luciana Parisi and Wendy Chun. Algorithmicsuperstructuring presents as pervasive expansion of algorithmic processingand logic, installed under the techno-capitalist drive for quantifying,consolidating and regulating human experience. Algorithmic superstructuringis built into networks of distribution and circulation of affect and flourishes inthe cognitive frameworks of interfaces and protocols. Building on previouscuratorial work and drawing on media art practices, this paper aims toinvestigate how inhumanity of algorithmic modes and models of reasoning isreflected in the distribution of the sensible, and how the aesthetic regimes ofalgorithmic governance could be articulated.KEYWORDSalgorithmic superstructuring, media art, algorithmic governance, distributionof the sensible
Anikina 36[1] See Danaher et al. (2017) forcomprehensive outlines of thestakes and challenges ofalgorithmic governance.[2] In the last two decades,theories of media have adapted tothe emergence and proliferationof algorithmic processing. Someof the pivotal discussions in thissense are represented by softwareand new media analysis by LevManovich (2001), and the criticalstudies of software by MatthewFuller (2003) and AlexanderGalloway (Essays 2006). Due tothe challenges that have beenbrought on by advanced dataprocessing techniques such asmachine learning, data mining andpredictive analysis, the algorithmicitself, seen as a concatenation ofdesign, economic, cultural andpolitical concerns, has become thefocus. The term “algorithmicculture” was offered by Galloway(Essays 2006); further mapping ofthe term can be found in Gillespie(2014) and Striphas (2015).Media art has traditionally responded to questions of algorithmic governanceby opening the ‘black box’ of technology, by revealing the power structuresinside the machine, be it explicitly or implicitly, through hacking, commentary,speculative narratives, glitching and other methods. In the current moment itseems that such blackboxing takes on more menacing forms, as algorithms aremade more and more opaque not only by proprietary claims, but also bynarratives of technological mastery and progress. This is reflected in theinvestigations of algorithmic governance, or even ‘algocracy’ (Aneesh 2006,2009; Danaher 2016), characterised by the unprecedented consolidation ofaccess to big data and to proprietary algorithmic solutions in the hands ofvarious state and corporate bodies. [1] Zeynep Tufekci (2015) identifies themain dangers of algorithmic governance as “lack of transparency, informationasymmetry and hidden influence” (p.207). The shift to focus on the“algorithmic” from “software” has been also reflected in media studies. [2]“Drawing on contemporary media art practices and on studies of visual andalgorithmic cultures, I would like to develop the idea of algorithmicsuperstructuring as a reading of aesthetic regimes of algorithmic governance.Through the work of Jacques Rancière and Luciana Parisi I will discuss howthe inhumanity of algorithms can be regarded in relation to political aesthetics.First, I will outline how algorithmic superstructuring can be seen asdistribution of the sensible and suggest the readings of inhumanity ofalgorithms within this setting. Secondly, I will draw on media art works in orderto situate the processes of algorithmic superstructuring as localised in mediaart, labour, in blackboxing and imaginaries of technology. Finally, I willconsider how interfaces and ‘persistence of vision’ affect algorithmicsuperstructuring as a larger condition and as an aesthetic regime of algorithmicgovernance.The curatorial concept, Algorithmic Superstructures, was developed as a theme forIMPAKT media art festival by Yasemin Keskintepe, Luba Elliott and I. It tookplace in Utrecht, Netherlands, in October 2018. Algorithmic Superstructures in itsinitial iteration was aimed at investigating widely, through artistic, theoreticaland design approaches, the epistemic and affective shifts brought on byalgorithmic processing. The idea of algorithmic superstructures appeared as anattempt to describe the ways in which the traditional systems of politics, media,labour and art are being overlaid and displaced by new algorithmicframeworks, interfaces and protocols, installed under the techno-capitalistdrive for quantifying, consolidating and regulating human experience.Referring specifically to these processes of displacement, we imaginedalgorithmic superstructures as pervasive, expansive, open vectors ofalgorithmic processing and logic that flourish under the auspices of theattention economy, where codes, images, software and protocols serve asprimary mediators in the networks of the commercialisation, capture andcirculation of affect. The spaces of knowledge and affect production that arecreated within algorithmic culture were the primary focus of our festivalconcept. In the course of the festival and its many public discussions andconversations with artists, it seemed that there was a particular ambiguitywhere artistic fascination with the inhuman nature of algorithmic reasoningintermingled with the inhumanity of rationality clearly seen in the proprietarystructures of algorithmic governance. These two kinds of inhumanity – the
Anikina 37“alien” and the “proprietary” one – seemed to be at times confused, equated,or intentionally interchanged.In this paper I aim to articulate the idea of algorithmic superstructuring (nowused as a verb) in order to see more clearly how these two kinds of inhumanityare actualised as political aesthetics of algorithmic governance. Thesuperstructuring in algorithmic superstructuring takes on a different meaningfrom its original Marxist debate of base and superstructure. In the transition,as Raymond Williams puts it, “from Marx to Marxism,” the economic base hasbeen often interpreted as determining the political and legal (later alsoideological and cultural) superstructure (“Base and Superstructure”). The stricteconomically reductionist approach was already criticised by Engels in a letterto Joseph Bloch in 1890 (Marx and Engels 498), and many theorists argued fora more nuanced reading of the interrelation of base and superstructure.Williams in his insightful analysis suggests that instead of a rigid, staticunderstanding of base and superstructure, there needs to be a “more activeidea of a field of mutually if also unevenly determining forces” (Problems inMaterialism and Culture 36-37). Alex Callinicos suggests that the forces andrelations of production merely set limits to the “superstructure” rather thandetermine it (97). Similarly, the use of superstructuring in this paper is notaimed at representing algorithms as a rigid hegemonic superstructure, butrather at investigating them as processes, and at discovering their potentialaffective and cognitive agency within governing structures, as well as theaesthetic configurations that result from it.Theorists of digital culture have often had recourse to Rancière’s formulationof the distribution of the sensible to describe the regulatory function ofcomputational processes in the acts of concealing and revealing (Steyerl “ProxyPolitics”; Dieter 222). In this sense algorithms, software, protocols andinterfaces can be seen as politico-aesthetic regulators, in Rancière’s sense ofsuch phenomena serving the function of “delimitation of spaces and times, ofthe visible and the invisible, of speech and noise, that simultaneouslydetermines the place and the stakes of politics as a form of experience”(Rancière 8). Where the distribution of the sensible “reveals who can have ashare in what is common to the community based on what they do and on thetime and space in which this activity is performed” (8), algorithmicsuperstructuring also establishes modes of being, sensing and acting.In Rancière’s account, the scope of the “aesthetic” is not confined to questionssuch as the status of the art object, but rather pertains to the general field oflife and its sensible forms and practices. Seen in this light, aesthetics inalgorithmic capitalism refers to the cognitive production and movement ofaffect. As Michael Dieter notes, “The alteration of sense and perception inCTP [critical technical practice] speaks to the classic meaning of aisthesis, butnow explicitly defined by sociotechnical events” (220), where aisthesis is meantas perception through the senses. In this sense Rancière’s argument for thedistribution of the sensible directly aligns with the vectoral character ofalgorithmic superstructuring. Algorithms produce meaningful and affectiveaspects of life, and more often than not this occurs along the vectors ofalgorithmic governance. As algorithmic infrastructures underlie the conditionsfor working, learning, consuming and creating, the algorithmic superstructuring
Anikina 38re-distributes the sensible aspects through its interfaces, analytical modes andchoices. If one applies Rancière’s call for aesthetics to serve “the invention ofthe sensible forms of the life to come” (24), the ethical task of the media artistseems to be located precisely in uncovering and re-inventing forms ofexperience produced by technical media.The key capacity of algorithmic superstructuring to circumvent modern modesof organisation and install its own logic can be seen as an ability to affect thevery conditions of knowledge production. Where the value chain andinformation distribution are controlled, the communication space is alsoreorganised accordingly. Tarleton Gillespie argues that the algorithm hasbecome “a key logic governing the flows of information” (167). Algorithmicsuperstructuring, crucially, has to be seen not as a solely materialistreorganisation of economical structures, but also of meaning itself. Whileinterfaces participate in the distribution of the sensible in the most direct way,by offering and limiting choices of action of the user, it is the invisiblealgorithmic processes and power formations that affect the conditions ofmeaning-making. It is in this sense that Ganaele Langlois argues for the shifttowards understanding meaning not only as a human process, but as “one thatis increasingly dependent on media technologies” (5). In her investigation ofsocial media algorithms, she finds that software contributes to meaningmaking as “a semiotechnology in charge of producing both meaning and theconditions for the experience of meaningfulness” (19).So how do we address the distribution of the sensible, if the distributionprocess involves an active algorithmic renegotiation of pre-cognitive aspectsof ‘sensible’? By pre-cognitive aspects I mean those that occur either outsidethe scale of human cognition and senses (such as high-frequency trading), orpassing below the threshold of user’s media literacy (such as interfaceelements). I would like to suggest superstructuring as the process of technocultural construction of meaning which, importantly, acknowledges the agencyof the technical components that modulate and direct cognitive processes.That does not mean assigning human-like agency to algorithms, but, rather,considering their part in the decision-making processes. Parisi suggests, in theessay “Reprogramming Decisionism,” that with the incursion of algorithmicautomation into decision-making processes, it becomes possible to speak of akind of “technological decisionism, which values making a clear decisionquickly more than it does making the correct one” (para 2). She positsalgorithmic processing, following N. Katherine Hayles, as a “nonconsciousform of cognition, solving complex problems without using formal languagesor deductive inference” (“Reprogramming Decisionism”, para 13), as well asworking at scales and speeds inaccessible to human perception. Where Haylessuggests the “the exteriorization of cognitive abilities” (11) to technicalsystems, this suggestion seems to echo one of the definitions that Williamssuggests for superstructures – that of the “forms of consciousness” arisingfrom the conflict introduced by real relations of production (Marxism andLiterature 76).In other words, the transformation of the computational paradigm from therationalist top-down causality to correlation in advanced algorithms also meansan epistemological shift towards what Parisi calls “soft thought” (Contagious
Anikina 39[3] For example, the Institute forEthical AI & Machine Learning inUK (opened in 2018), the Ethicsand Governance of ArtificialIntelligence Initiative launched byMIT Media Lab and the HarvardBerkman-Klein Center in 2018,and the European AI Alliance(2019). This is not to excludeearlier work accomplished in thisdirection, or a wide range of opensystems that are community-run(such as Women in AI orPlatform CooperativismConsortium), but to underline arecent surge in the wideracknowledgement of necessity forethical regulation of algorithmicprocessing.Architecture) and, later, “inhuman thinking” (“Reprogramming Decisionism”).To take one of the popular examples, in neural networks, the direct causal linksbetween the algorithm and the datasets are foregone in favour of moreefficient meta-analysis that creates its own algorithmic relations as it analysesdatasets, and bases the ensuing analysis on those algorithmic relations. ForParisi, this new paradigm represents a radical departure from the rationalisttheories of computation precisely because it is inseparable from indeterminacy,noise and unknown data: machine learning “is indifferent to the entropic noiseof increasing data volumes insofar as this noise is precisely part of the learningprocess” (“Reprogramming Decisionism”). “Inhumanity” in this case can beattributed to the technical impossibility of tracing the entirety of microdecisions that went into building a specific model. However, it should also beconsidered from the point of virtual opacity of such models as they enter theareas of decision-making, often staying within the proprietary copyright oftheir owners. While there has been a noticeable increase in recent initiativesfor ethical guidelines and legislation surrounding the use of datasets, [3] in thecurrent moment such models are freely built into the processes of analysingexperience and channelling affect.The participation of “proprietary” inhumanity in the algorithmic distributionof the sensible can be clearly seen in how it is organised around the axis of thecommodification of experience. The process of data commodification runsparallel to the commodification of affect and the design of user experience.Following the vector of algorithmic superstructuring from the initial economicmotivation to the design and implementation of algorithms, it becomespossible to see how the design of user experience perpetuates the automationof various cultural operations through software abstraction. This can be seenin platforms such as Netflix, as well as other services using recommendationalgorithms. In this sense, the dangerous aspect of algorithmic superstructuringlies not only in its pervasiveness, but in the loop of commodification ofexperience and affect that it enables. Where data collection participates ininfinitely updating feedback loops, it guarantees continuous commodification:as data analysis turns human choices, experience and attention into rationalisedmodels, these models, in their turn, create more and more refined and precisedefinitions of what kind of experience is marketable. As Brian Massumi notes,“the ability of affect to produce an economic effect more swiftly and surelythan economics itself means that affect is a real condition, an intrinsic variableof the late capitalist system, as infrastructural as a factory” (45). The affectivecapacities of algorithmic procedures are therefore embedded in the softwareregime of abstractions.But while it is generally possible to trace the “proprietary” inhumanity ofalgorithms to the economic motivation of the platform, does it account for theentirety of experience and affective space that it creates for the user? In “TheIncomputable and Instrumental Possibility,” Antonia Majaca and Parisi drawon Judith Butler’s writing to suggest a possibility of feminist re-claiming ofmachinic instrumentality. They suggest that machine logic, primarily seen as apart of a “paranoid techno-industrial apparatus” relying on collecting andflattening data as predictive models, could be also reclaimed on its own terms,as an alien logic that embraces its own instrumentality and repurposes it for itsown ends, potentially disrupting the white-male concept of humanness as a
Anikina 40whole. While this suggestion is speculative in nature, it offers a way toacknowledge that the “proprietary” inhumanity of algorithmicsuperstructuring does not exclude the possibilities for other kinds ofinhumanity, built in different ways and experienced in different ways. Theinhuman scales of technical infrastructures introduce affective and cognitiverenegotiation of human experience. Algorithmic superstructuring, understoodas distribution of the sensible, can be then seen as a scalable dynamic thatnegotiates the infrastructural design of algorithms as an aesthetic regime thatconditions the production of meaning. Considering the local aestheticconfiguration, where machine logic is normally hidden behind the persistentsprawl of well-interfaced business solutions, the process of ”un-blackboxing”should include a similarly persistent inquiry into the processes of design andlabour.[4] See description anddocumentation on the work’swebsite, ImageNet.xyz.In order to look at algorithmic superstructuring through the inquiry of artisticpractice, I will draw on several projects from the festival, selected forexhibition by Yasemin Keskintepe. These projects, in particular, engaged withthe question of the correlational rather than causal links introduced by themachine learning technologies, investigating how these technologiesparticipate in meaning-making in embedded cultural contexts. The work ofConstant Dullaart, The European Classes, Euronet (2017), developed incollaboration with Adam Harvey, uses convolutional neural networks thatrecognise objects within images in order to create an image dataset. The artistsretrained the networks on “European artefacts” in order to investigate howEuropean cultural output can be presented in a dataset, and how the networkcan classify something as European. [4] The 152 classes for image recognitionranged from common – “guitar,” “beret” – to more specific: “Hagelslag,”“Chancellor Angela Merkel.” The neural networks, essentially tasked with thequestion of European identity, are solving the technical question of semanticsegmentation – what parts of the images are recognised as a particulardescriptor, and what images can be reconstructed from these correlationallinks. Highlighting how algorithmic automation can enter the areas that havebeen considered a cultural domain, this work also reconstructs the capacity ofmachine learning to produce meaning, referring to the cases in which machinelearning techniques, when
“algorithmic” from “software” has been also reflected in media studies. [2] “Drawing on contemporary media art practices and on studies of visual and algorithmic cultures, I would like to develop the idea of algorithmic superstructuring as a reading of aest
An aesthetic judgement is a decision made by an individual regarding the aesthetic value, the quality, of an artwork. The outcome of an aesthetic judgement is expressed through observable expressions of aesthetic judgements, such as a person speaking positively about a work or purposely reexperiencing an artwork. Aesthetic judgements
Founded in 1970, the International Society of Aesthetic Plastic Surgery (ISAPS) is the world’s leading professional body for board-certified aesthetic plastic surgeons. Regarded as the leading global authority on aesthetic and cosmetic surgery, ISAPS’ threefold mission is to offer aesthetic education worldwide, to provide accurate
Please credit The Aesthetic Society when citing statistical data. 2 Contact: The Aesthetic Society 562.799.2356 media@surgery.org www.surgery.org fax: 562.799.1098 The Aesthetic Society was founded in 1967 when aesthetic surgery was only beginning to be recognized as the important subspecialty it is today.
Nº Contribuinte NIF Regime Repartição Fiscal de Domicílio LISTA DOS CONTRIBUINTES ENQUADRADOS OFICIOSAMENTE NO REGIME GERAL DO IVA Versão actualizada em 04/01/2021 41 AFRICHAHEL FARMACIA SU LDA. 5417539678 Regime Geral RF CACUACO 42 AFRICONSULT-CONSULT.DE ENGENHARIA, LDA 5403076875 Regime Geral 2ª REPARTIÇÃO FISCAL
Algorithmic Trading Table: Proportions of trading volume contributed by di erent category of algorithmic and non-algorithmic traders in the NSE spot and equity derivatives segment (for the period Jan-Dec 2015) Custodian Proprietary NCNP Total Spot Market Algo 21.34% 13.18% 7.76% 42.28% Non-
aforementioned model. Thus, the Triangulation Algorithmic Model is provided and defined to aid in understanding the process of measurement instrument construction and research design. The Triangulation Algorithmic Model for the Tri-Squared Test The Algorithmic Model of Triangulation is of the form (Figure 2). Where,
In these notes we focus on algorithmic game theory, a relatively new field that lies in the intersection of game theory and computer science. The main objective of algorithmic game theory is to design effective algorithms in strategic environments. In the first part of these notes, we focus on algorithmic theory's earliest research goals—
MPhys Astrophysics with a Year Abroad (2021-22) This specification provides a concise summary of the main features of the programme and the learning outcomes that a typical student might reasonably be expected to achieve and demonstrate if s/he takes full advantage of the learning opportunities that are provided. Awarding Institution University of Southampton Teaching Institution University of .
