https://ofstringsandkings.pointlinesurface.com Of Strings & Kings Sun, 05 May 2019 04:38:52 +0000 https://ofstringsandkings.pointlinesurface.com en https://ofstringsandkings.pointlinesurface.com/Cover Sun, 05 May 2019 04:38:52 +0000 Of Strings & Kings https://ofstringsandkings.pointlinesurface.com/Cover <img width="2000" height="2000" width_o="2000" height_o="2000" data-src="https://freight.cargo.site/t/original/i/a88a62299787691c1f55608c6e92a01f83c5cc7a041e9bc5c1c941384caa7b98/Of-Strings---Kings-1.jpg" data-mid="41499057" border="0" src="https://freight.cargo.site/w/1000/i/a88a62299787691c1f55608c6e92a01f83c5cc7a041e9bc5c1c941384caa7b98/Of-Strings---Kings-1.jpg" /> https://ofstringsandkings.pointlinesurface.com/Introduction Thu, 28 Mar 2019 01:12:20 +0000 Of Strings & Kings https://ofstringsandkings.pointlinesurface.com/Introduction Introduction "Every disease is a musical problem, every cure is a musical solution." ― Novalis&nbsp; &nbsp; &nbsp; &nbsp;&nbsp; This project is an odd confluence of three of my interests: political economy, machine learning, and luthiery of odd acoustic instruments. None of these affairs I can claim to be particularly knowledgeable about so, per usual, this project began by wielding my typical artist-as-excuse ethos to embark on an adventure wrought by intrepid ignorance. Of Strings and Kings is a set of musical compositions written for strings and machine learning algorithms intended to accompany this manuscript. The compositions interspersed throughout the essay will serve as aesthetic counterparts relating to each chapter’s subjects. The first chapter will draft a framework of some of the bleak vectors of platforms, big data, cloud computation, and centralized AI when drawn to their conclusions. I will tell a cautionary tale of this totalizing hegemony that is being established by harnessing cluster computation and transcontinental network infrastructure and will draw historical parallels to the despotic doctrine that entrenches sovereign rule. In the second chapter I will address the ideas of play and practice as a way of claiming agency in a domain of increasingly unique sets of uncertainty. I will use a multitude of musical metaphors to correlate to various deep learning techniques (training, prediction, learning, feedback, and bias) and will use these metaphors to articulate ways of retaining harmonious improvisational dynamics with the abounding swarms of behavior modifying algorithms that live in our pockets, networks, and cities. The third chapter’s aim is to review alternative tuning systems as a tool for recalibrating not just our music, but our social relations. Much like tuning an acoustic instrument, algorithms are tuned and optimized. I will posit different ideas for how we get our hands on these parameters to bring about more harmonious and equitable outcomes. I will discuss a number of topics including&nbsp;data-as-labor, distributed ownership, co-operative parametric design, right-to-explanation, decentralized machine learning, and optimizing for utilities that redistribute decision making and minimize extraction. Before diving into the first chapter allow me to provide a brief primer on the composition and instrumentation. These compositions were written for custom fabricated lyres, hammered dulcimer, lute, gittern, and popular machine learning techniques including LSTM RNN1 (long short-term memory recurrent neural networks), adversarial neural audio synthesis2, and a multitude of others cited in appendix B3. I will only partially expound upon on the relevance of the various tuning systems and algorithms’ aesthetic and technical significance throughout the essay but will provide detailed auxiliary reflections in appendix A4.&nbsp;In short, the music is stylized in neo-feudal minstrel songform with accompaniment from gothic automata. This synthesis of ars antiqua and deep learning portends a speculative future in which the class asymmetries of yore are resurrected through menacing algorithmic injunctions. 1&nbsp; &nbsp; “LSTM Networks” https://www.tensorflow.org/tutorials/sequences/recurrent 2 &nbsp; “Adversarial Neural Audio Synthesis | OpenReview” https://openreview.net/forum?id=H1xQVn09FX 3 &nbsp; “Appendix B: Algorithms &amp; Libraries", https://ofstringsandkings.pointlinesurface.com/Appendix-B-Algorithms-Libraries 4 &nbsp; “Appendix A: Auxiliary Musical Context”, &nbsp;https://ofstringsandkings.pointlinesurface.com/Appendix-A-Musical-Context https://ofstringsandkings.pointlinesurface.com/Chapter-I-Masks-of-Power Thu, 28 Mar 2019 01:28:29 +0000 Of Strings & Kings https://ofstringsandkings.pointlinesurface.com/Chapter-I-Masks-of-Power 1.0Masks of Power https://ofstringsandkings.pointlinesurface.com/Mystification-and-Naturalization-as-Power Thu, 28 Mar 2019 02:43:19 +0000 Of Strings & Kings https://ofstringsandkings.pointlinesurface.com/Mystification-and-Naturalization-as-Power 1.1Mystification, Naturalization, &amp; Literacy This section will outline three masks that contemporary power has adorned in the age of deep learning. These masks are ancient methods of refashioning authority, but are being retrofitted for planetary computational dominion. Used for distancing and obfuscating material jurisdiction away from the general population, the sovereign justify and cloak their authority with these historically vetted tactics. MYSTIFICATION Where realpolitik’s brute force approach to wielding power dwindles is where its twin, mysticism, intervenes. Across most cultures, authority can be found seated next to the keepers of transcendental doctrine: monarchs’ devotion to the papacy, maharajahs’ consultation with the brahmins, shahs’ relation to the caliph, emperors’ adjacency to the hierophants, and so on. The mystical consulate are positioned as the adjudicators of transcendental doctrine. Their creeds typically required their hermeneutic expertise due to complex, abstract, and remote precepts. The modern humanist project that began during the European Enlightenment quite literally dethroned old world autocratic rule as reason triumphed and laissez-faire capitalism formed a new political subject: the individual liberal citizen. This social subject was granted inalienable rights that superseded the divine decrees of pontiffs and monarchs. However, in the age of computational capital and ecological meltdown we are witnessing a crisis of the Westphalian nation-state, the primary maintainer of the principles of the humanist project. In an accelerating geopolitical landscape where intelligent algorithmic systems continue to agitate national identities, desubjectify modern democratic subjecthood, and retopologize new jurisdictions according to remote technocratic administration it is becoming evident that monolithic power is re-emerging. ︎ Refer to Appendix A1 for ancillary musical, historical, and technical details Given the emergent complexities in our global networks of coordination, algorithmic capitalism has anointed its entrepreneurial engineers to remodel the world to reflect the staggering rate of growth. To contend with the sheer amount of information processing human oversight is bypassed due to the magnitude of tasks. These convoluted processes of automation are developed and maintained by the embrocated. This software canonicate occupy the edges of computation where mathematics becomes metaphysics; where transcendental logic seeks to untether intelligence from the shackles of humanism. Some of these machine intelligence researchers utilize allegories of gods for framing models of intelligence that exceed the capacity of humanity. While mythology can serve as a mirror for reflecting insights into the human condition, these analogies with AI generally play into teleological and theological schemas of determinism and inevitability that are ultimately unhelpful misnomers. Deifying complex systems serves as another tactic for elevating and naturalizing investment capital’s capacity to aggregate and analyze colossal amounts data at planetary scale. Myth-making is to be expected as humans attempt to explicate these ineffable, foreign processes that play such an enormous role in our lives. Since its indigenous origins humanity has spun stories using in-built associations to construct meaning and negotiate with uncertainty. However, these types of animistic myths of enchanted relationality and ecological cohabitation with other intelligences are not the ones cropping up in Silicon Valley circles. These factions have categorically discarded pre-colonial models of temporality, including dreamtime and cyclical time, supplanting them with a chronological telos of autocatalytic productivity that will beget emergent computational supremacy. This belief in a deterministic a priori supreme being made of pure reason (logos) and is quite literally derived from ecclesiastical Christian dogma and rabbinical Hebrew scripture. In fact, the Jesuit priest Pierre Teilhard de Chardin postulated the Omega Point5, a theory that the universe is evolving towards a maximum state of complexity and consciousness. Despite its religious provenance, his ur-Singularity cosmology has been widely adopted by many secular executives and engineers helming the burgeoning technocracy. <img src="https://freight.cargo.site/t/original/i/eb43b477da82643050de6e8cf12e6872cb171a2f5ea93e6eb6553d3049486002/IMG_0947.JPG" alt="The Ethics of Big Data" style="width:100%"> The Ethics of Big Data, O'Reilly Media, 2012 <img src="https://freight.cargo.site/t/original/i/dc30fc708fe296331f65e0d1e10ad0627cd6529309e72d046654c7a4d1d83cff/priest_blessing_server_room.jpg" alt="Server Blessing" style="width:100%"> Data Center Blessing;&nbsp;# /etc/init.d/daemon stop Image Credit: India Times The Sophist notion of technē which forms our most essential figuration of technology is derived from Promethean myth: whereby fire was stolen from the gods and bestowed upon humanity spawning progress and civilization. This correlation is apotheosized in Yudkowsky’s Bayesian horrors6, Alexander's transhumanism7, Kurzweil's technological singularity8, Land's cosmological singularity9, Bostrom's Superintelligence10 and Levandowski's Church of AI. The irony in these rationalists’ accounts of intelligence is that they are all reifications of inherited cosmologies that aren’t predicated on formal logic, but on myth. When subjected to methodological rigor these narrative-based scare tactics are subsumed by heuristic biases and collapse into xenophobic rhetoric. Mystifying computation in this way tends towards a cosmic narcissism: gazing into the abyss as the abyss affirms its own preconceptions of itself. Regardless of one’s stance on the othering of hypercomplex computation I would argue that humanity should cultivate a lexicon for talking about aliens, others, or xeno-intelligence without dynamic divergence. The theories of breakaway recursive self-enhancing technology that undergird these conceptions of intelligence are often discussed in a bounded immaterial domain without the examination of the anatomical geographic scale. These exceedingly brilliant cerebral meta-linguists and transcendental number theorists often neglect to acknowledge the material earthen corpus upon which their symbolic logic is expressed.11 Researchers Kate Crawford and Vladen Joler’s Anatomy of an AI System12 &nbsp;is a graphical dissection of the infrastructural assemblage required to embody artificial intelligence; spanning the fabric of capital to include mineral resource extraction, human labor, supply chain logistics, data collection and distribution, analytics, prediction, and optimization, their project synopsizes the pipeline for constructing a deep learning system. This corporeal plexus is typically trivialized with the public relations nomenclature of the Cloud. Referring to the vast material architectural projects of data centers with the ephemeral parlance of the Cloud is not just a bit disingenuous, it is downright deceptive. As platforms deploy preemptive algorithms into the cultural apparatus it requires a delicate and tactical marketing narrative as its trojan horse. As architecture design critic Keller Easterling aptly suggests: “You can see the discrepancy between what organizations are saying and what they are doing. You can even see temperament in construction or potentials for violence. That disposition is propensity within a context, property or tendency that is unfolding over time.”13 By steering the public narrative away from its material operations, these supranational syndicates are able to truncate opposition by minimizing attention to what they are building. LITERACY Historically the doctrinal gatekeeping that delineated class and caste was maintained by ecclesiastical scribes. There are striking parallels to the emerging niche of computational literati contributing to the steepening disparity of wealth and technical literacy. Both clergy and programmers provide order, in the form of textual statutes, that designate the foundational axioms upon which a society rests. Similar to the vertical denominations that emerge with religious sects, emanant power has stratified social order and restricted access to its inner sanctum with media illiteracy and technical deficiency. The development of user experience serves as an exegetical layer between the code and its graphical representation. These simplified behavioral flows divert users away from the software’s extractive disposition and carry the hermeneutic subtext of “leave it to the experts”. UX is a set of inherited decisions that form interfaces to assist in navigating a computational domain. While user experience design affords those without literacy to use digital media with relative ease and is an essential facet of all computing, the Jobsian design ethos of “it just works” is synonymous with consumerism and is not designed to garner media literacy. By funneling users into compromised sets of autonomy these firms are able to maintain opacity by occluding access to the operational facets of their services. For example, the frictionless front-end of Amazon’s Alexa and its ilk are reverse portals into behavioral surplus supply chains. Developing formal design criteria, let alone literacy, for deep learning is a sophisticated process that even those who are are working in the field have failed to do. Arguably since artificial intelligence interpolates the latent fields between engineering, science, philosophy, mathematics, design, and spectacle, it doesn’t satisfy the methodologies that quantify any real metrics specific to each respective field.14 This provides us interesting output but becomes subject to motivated logic15 and is used to “prove” certain assumptions but without the rigor and criteria that each of the aforementioned fields uses. This effectively allows interesting, but not provable claims to slip through under the guise of technical or conceptual rigor.16 AI researchers are more likely educated in fields that take formal problems as inputs: engineering, computer science, mathematics, or theoretical physics. Yet the problems being tackled are mostly ones in which a design approach, maintaining a continuous, open-ended relationship with nebulosity, may be more appropriate.17 These applications assume a highly technocratic solutionist position on social life; entire fields and industries are “disrupted” by platform engineering that ignores the specialist knowledge held by experts in their respective practices. This approach circumvents domain specific expertise and supplants it with big data. With its appeals to bottom line, businesses can afford to overlook the resultant margin of error bypassing proficient professionals with automated solutions. What the adopters of these “disruptive solutions” may not realize is that these platforms are using tautology to substantiate their claims. In essence the proponents create the criteria that they need to fulfill, fit the data to qualify their results, and prove that their “solution” will cost less than hiring experts. ︎ Refer to Appendix A2 for ancillary musical, historical, and technical details NATURALIZATION Venkatesh Rao alleges that “deep learning has an authoritarian right wing bias. It feeds on vast data sets created by natural behavior, has a tendency to inherit and reproduce endemic biases, and codify them in favor of conservative authoritarians who see the incumbent balance of power as natural and just.” Rao goes on to state that the management class organizing the business and social formations for how deep learning is codified claim that trying to “regulate” the functioning of deep learning algorithms directly, through human political processes, or by demanding ‘justifiable AI’ that can explain itself, is a fool's errand. By adopting this framing of deep learning it becomes a tautological justification for itself. Outside of the market-based rhetoric of profit motivation, how is this being justified as data science? These algorithms use techniques that leverage what are called adversarial networks. These techniques use computation to accelerate simulated evolutionary processes by determining “data fitness” through a mathematical sorting process. These are sometimes referred to more broadly as genetic algorithms.18 These information processing models compute data by simulating Darwinian natural selection where only the best data “survives” making it through a statistical gauntlet of adversarial regression. By using these types of sorting algorithms many Silicon Valley executives leverage and vindicate the results as “natural” data science; the truth is that they are anything but natural. Their reasoning generally claims that if you have enough data the algorithm will be able to arrive at a statistical equilibrium having run through enough permutations of evolution. This assumption is predicated on the cum hoc logical fallacy19 which, in statistical lexicon, can be summed up as “correlation doesn’t imply causation.” As more of big data’s conclusions are subjected to scientific rigor outside of its own self-affirming means-tested regressions it has been shown that more data can often lead to erroneous results.20 21 Cultural critic Mike Pepi articulates the hubristic naturalization of platforms as biological organisms in his astute analysis of Silicon Valley’s Sublime Administration: Between Platform and Organism22. The careful use of biological and evolutionary language around these techniques is intentional because the proponents position their critics as being against “science”. This has been a continual assertion of those advocates seeking to deepen the entrenchment made by these systems. The evolutionary justification for these conclusions are not just fallacious but are epistemologically falsifiable. Francis Galton, a pioneer in eugenics and biometrics, was also a progenitor in the field of statistics. The statistical techniques that Galton invented23 (correlation, regression) and phenomena he established (regression to the mean, bivariate normal distribution) form the basis of the biometric approach and now operate at the core of deep learning data analytics. Historically, we’ve seen this type of evolutionary rhetoric attempt to draw sinister erroneous conclusions about race and criminology. In the age of deep learning we are experiencing the resurgence of outmoded 19th century ‘race (pseudo)science’ [e.g. criminal anthropology23, biological determinism24, social Darwinism25, phrenology26, physiognomy27, and eugenics28). Like much of the induction biases latent in deep and reinforcement learning systems these fallacious notions are predicated on racist methodological weaknesses (poor sampling technique, bias in gathering data, poor statistics)29. Even though these claims have been categorically disproven we are observing deep learning with these biases rapidly mount the American martial systems30 of ICE31, NYPD32, U.S. Army33, Orlando Police34, New Orleans PD35, and Washington Sheriff’s Department36. 5&nbsp; &nbsp; "Teilhard de Chardin and Transhumanism." https://jetpress.org/v20/steinhart.htm 6&nbsp; &nbsp; "Rationality: A-Z - LessWrong 2.0." https://www.lesswrong.com/rationality 7&nbsp; &nbsp; "Transhumanism | Slate Star Codex." https://slatestarcodex.com/tag/transhumanism/ 8&nbsp; &nbsp; "Ray Kurzweil | Singularity" https://www.kurzweilai.net/futurism-ray-kurzweil 9&nbsp; &nbsp; "Fanged Noumena, Nick Land" http://azinelibrary.org/trash/fangednoumena.pdf 10 &nbsp; "Nick Bostrom | Superintelligence" https://nickbostrom.com/ 11&nbsp; &nbsp;"Total Consumer Power Consumption Forecast - ResearchGate." https://www.researchgate.net/publication/320225452_Total_Consumer_Power_Consumption_Forecast 12 &nbsp; "Anatomy of an AI System." https://anatomyof.ai/ 13&nbsp; &nbsp;"Keller Easterling — Extrastatecraft: The Power of Infrastructure Space." http://kellereasterling.com/books/extrastatecraft-the-power-of-infrastructure-space 14 &nbsp; "How should we evaluate progress in AI?" &nbsp;https://meaningness.com/metablog/artificial-intelligence-progress 15&nbsp;&nbsp; "Artificial intelligence pioneer says we need to start over - Axios." https://www.axios.com/artificial-intelligence-pioneer-says-we-need-to-start-over 16&nbsp; &nbsp; "Why AI Is Not A Science - Stanford University" https://web.stanford.edu/group/SHR/4-2/text/matteuzzi.html 17&nbsp; &nbsp; "Troubling Trends in Machine Learning Scholarship” http://approximatelycorrect.com/2018/07/10/troubling-trends-in-machine-learning-scholarship/ 18 &nbsp;&nbsp; "Genetic Algorithm [Deep Learning Patterns]" http://www.deeplearningpatterns.com/doku.php?id=genetic_algorithm 19&nbsp; &nbsp; "Logical Fallacies » Cum Hoc Fallacy." https://www.logicalfallacies.info/presumption/cum-hoc/ 20&nbsp; &nbsp; "Causal Inference and Statistical Fallacies" http://www.math.chalmers.se/~wermuth/pdfs/96-05/CoxWer01_Causal_inference_and_statistical.pdf 21 &nbsp; "Issues with data and analyses: Errors, underlying themes, and ..." https://www.pnas.org/content/115/11/2563 22&nbsp; &nbsp; "Jenna Sutela - Orgs: From Slime Mold to Silicon Valley - Printed Matter." https://www.printedmatter.org/catalog/49553 23 &nbsp; "Francis Galton: Pioneer of Heredity and Biometry | Johns Hopkins University Press" &nbsp;https://jhupbooks.press.jhu.edu/title/francis-galton 23&nbsp; &nbsp; "Neural Network Learns to Identify Criminals by Their Faces - MIT." https://www.technologyreview.com/s/602955/neural-network-learns-to-identify-criminals-by-their-faces/ 24 &nbsp; "OSF | Deep neural networks are more ...." https://osf.io/zn79k/ 25&nbsp; &nbsp; "Researchers Want to Link Your Genes and Income—Should They?" https://www.wired.com/story/researchers-want-to-link-your-genes-and-incomeshould-they/ 26&nbsp; &nbsp; "FACEPTION | Facial Personality Analytics." https://www.faception.com/ 27&nbsp;&nbsp; "Automated Inference on Criminality using Face Images - Brown CS" http://cs.brown.edu/courses/cs143/2017_Spring/lectures_Spring2017/27_Spring2017_SocialGoodandDatasetBias.pdf 28&nbsp; &nbsp; "Sociogenomics is opening a new door to eugenics - MIT Technology" https://www.technologyreview.com/s/612275/sociogenomics-is-opening-a-new-door-to-eugenics 29 &nbsp; "Machine Bias — ProPublica." https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing 30 &nbsp; "AI is sending people to jail—and getting it wrong - MIT Technology." https://www.technologyreview.com/s/612775/algorithms-criminal-justice-ai/ 31 &nbsp; "ICE Extreme Vetting Initiative: A Resource Page | Brennan Center for Law." https://www.brennancenter.org/analysis/ice-extreme-vetting-initiative-resource-page 32&nbsp; &nbsp;"Palantir Contract Dispute Exposes NYPD's Lack of Transparency" https://www.brennancenter.org/blog/palantir-contract-dispute-exposes-nypd%E2%80%99s-lack-transparency 33&nbsp; &nbsp; "Palantir wins competition to build Army intelligence system", https://www.washingtonpost.com/world/national-security/palantir-wins-competition-to-build-army-intelligence-system/2019/03/26/ 34&nbsp;&nbsp; "ZeroEyes AI Threat Detection - ZeroEyes." https://zeroeyes.com/ 35&nbsp;&nbsp; "An improved kernelized discriminative canonical correlation analysis - IEEE" https://ieeexplore.ieee.org/document/6359400 36&nbsp; &nbsp; "Orlando Pulls the Plug on Its Amazon Facial Recognition Program" https://www.nytimes.com/2018/06/25/business/orlando-amazon-facial-recognition.html https://ofstringsandkings.pointlinesurface.com/Vectoralism-Content-Isn-t-King Thu, 28 Mar 2019 01:12:21 +0000 Of Strings & Kings https://ofstringsandkings.pointlinesurface.com/Vectoralism-Content-Isn-t-King 1.2 Vectoralism: Content Isn’t King&nbsp; Reminiscent of rentier capitalism or, even worse, absolute technocratic monarchism, we are increasingly finding that our interactions are mediated by platform economies, or what McKenzie Wark refers to as the vectoralist class37. I’ve adopted this term because it redefines the terms of the game. It seems that society is no longer operating according to capitalist rules, but something worse, and even more extractive. Vectoralism tracks the space of possibility. It annihilates the familiar concepts of ownership and property which capitalism is predicated upon. To retrieve or access a good or service for a reasonable price it is now temporarily leased as a service; vectors of access are monopolized by those that operate the infrastructure, protocols, and applications by which we have historically interacted and transacted in an unsurveiled peer-to-peer marketplace. By driving a wedge in between our social interactions to mediate our economies and social lives these private entities are able to learn intimate knowledge about us and siphon off a rentier tax from our labor, property, and interactions without any of the risk or responsibility to the commons. Contemporary power works as an environmental form of pre-emption; it is perpetually produced, monitored, refactored, and presupposed. Venture capitalist Marc Andreessen has infamously said that “software is eating the world”38 and unfortunately it seems to be hungry to replace democracy’s checks and balances with something that runs more efficiently. Vectoralists envision governance models where citizenship is relegated to the status of a user. Using computer networks this vision aims to circumvent nation-states granting of rights and supplant them with privileges and permissions managed by remote administration. In addition to physical walls, their prototypes are augmented with paywalls to access sites, services, and goods by embedding software and sensors into everything. The platform operates as the proxy layer through which all transactions are coordinated. Typically obscured under the auspices of ‘sharing’ and ‘convenience’ the actual subtext signifies the end of ownership. ︎&nbsp;Refer to Appendix A3 for ancillary musical, historical, and technical details <img width="960" height="958" width_o="960" height_o="958" data-src="https://freight.cargo.site/t/original/i/5bfdb2b3ad72ee95a3e0327cb9010f6216baf3105939f717f4448d16706faee0/IMG_0797.JPG" data-mid="38677986" border="0" src="https://freight.cargo.site/w/960/i/5bfdb2b3ad72ee95a3e0327cb9010f6216baf3105939f717f4448d16706faee0/IMG_0797.JPG" /> [Zuckerberg Family: How to Train Your Dragon, 2018. Image Credit: Facebook] Allow me to position this in the neo-feudal aesthetic framework of the compositions written for Of Strings and Kings. The task of the class of lords &amp; scribes (managers, programmers) is to cultivate, legitimize, and reify the knowledge systems of the kingdom (company, platform). These scribes take direction from the monarchic technocrats (CEOs) presiding over the kingdom and its wealth (data, software, infrastructure, capital) from which the serfs (users, netizens) temporarily lease from the fiefdom. The court jesters’ and minstrels’ (the creative class) role is to entertain the kingdom’s inhabitants and hypostatize absolute power through aesthetics (streaming content, marketing, branding, PR). 37&nbsp; "The Vectoralist Class - e-flux journal 56th Venice Biennale." http://supercommunity.e-flux.com/texts/the-vectoralist-class/ 38&nbsp; "Why Software Is Eating the World – Andreessen Horowitz" https://a16z.com/2011/08/20/why-software-is-eating-the-world/ https://ofstringsandkings.pointlinesurface.com/Intellectual-Property-and-AI-The-Ownership-of-Possibility Thu, 28 Mar 2019 03:27:43 +0000 Of Strings & Kings https://ofstringsandkings.pointlinesurface.com/Intellectual-Property-and-AI-The-Ownership-of-Possibility 1.3 Intellectual Property and AI: The Ownership of Possibility Those monitoring the deep learning related patent filings39 occuring in the U.S. over the past few years have observed an inordinate amount of lawsuits regarding intellectual property for machine learning algorithms and heuristics.40 This is obscenely controversial because these optimization and classification techniques are essentially mathematics wrapped up in applications. When properly understood, the implications of this should exceed concern and rapidly escalate into outright panic. This is, simply put, a legislative power play to stake out a monopoly on the future from the emerging vectorialist class. Allow me to elaborate insofar as I am capable: by owning these types of connectionist systems (neural nets, etc.) these companies are able to make complex projections about consumer behavior, cultural vectors, market proclivities, transnational interests, and natural &amp; biological systems that exceed anything we've ever encountered. These companies are claiming to own statistical techniques which predate these companies’ existence. By obfuscating the underpinning mathematics with technical applications for patent law purposes these companies are alleging ownership on things that they didn't create. Owning these types of learning algorithms not only assures the imminent obsolescence of worker-ownership but it guarantees no dividends or returns for those that have created the value. This power lies in monopolizing intellectual property — patents and brands — and the means of reproducing their value — the vectors of communication. This transcends the ownership of the means of production into the means of possibility. This is what Wark means when they say Vectoralism.41 <img width="500" height="467" width_o="500" height_o="467" data-src="https://freight.cargo.site/t/original/i/bdf16e4d7574e634a53d94fb5be55e38cebf0f9942da7e68ea6d58ab7ded4b81/politcal-compass-ai.png" data-mid="38678027" border="0" src="https://freight.cargo.site/w/500/i/bdf16e4d7574e634a53d94fb5be55e38cebf0f9942da7e68ea6d58ab7ded4b81/politcal-compass-ai.png" /> [Image courtesy of ribbonfarm.com] In the quadrant above we see the mapping of artificial intelligence’s ideological proclivities tending toward centralized power. I think this assessment is largely correct given the resource necessity and investment capital required to compute that much data. This rendition of the political compass axis meme was likely created with a sense of hyperbolic wit, but it ensnares an elemental veracity: ideologies are latent within these types of frameworks. I am not quite as polemical as Rao’s memefied compass and believe that it cannot be solely reduced to the authoritarian (“platform knows best”) vs. libertarian ("I own my data"). This is where we need to break open the differing layers of infrastructure ownership and discuss the merits of decentralized asynchronous back propagation, platform co-operatives, and unpack the data-as-labor discourse. I will address these subjects in chapter three of this essay and will elaborate upon them with some musical developments. The subjects of this chapter’s compositions and theses are to draw attention to the risk of totalizing ownership that may happen over the course of a single generation. The capacity to surveil, collect, and compute the magnitude of data currently being amassed has the potential to end half of existing labor markets in a matter of a few decades.42 Data aggregation has modified the technical composition of capital and has produced changes in value composition. Automation combined with the rate of capital accumulation will inexorably produce unforeseen levels of unemployment assuring the constant presence of a “surplus population” or “reserve army of labor.” Without some of the interventions outlined in the third chapter, this engine of indifference runs the risk of stratifying a class disparity that the West hasn’t seen in millenia. The title Of Strings and Kings isn't an empty aesthetic parable but is invoked to reveal a stark convergence of imminent realities. 39&nbsp; &nbsp; "’Quantum leap' in AI-related patent filings: UN (Update) - Phys.org"&nbsp; https://phys.org/news/2019-01-agency-asian-companies-ai-patents.html 40 &nbsp;&nbsp; "Patenting Algorithms: IP Case Law and Claiming Strategies - IPfolio" http://blog.ipfolio.com/patenting-algorithms-ip-case-law-and-claiming-strategies 41 &nbsp;&nbsp; "McKenzie Wark on the Rentier Vectoral Class - P2P Foundation Wiki."&nbsp; http://wiki.p2pfoundation.net/McKenzie_Wark_on_the_Rentier_Vectoral_Class 42&nbsp; &nbsp; "Automation and local labor markets - MIT Economics." https://economics.mit.edu/files/15254 https://ofstringsandkings.pointlinesurface.com/Chapter-II-Duets-Practice-Play Thu, 28 Mar 2019 03:35:52 +0000 Of Strings & Kings https://ofstringsandkings.pointlinesurface.com/Chapter-II-Duets-Practice-Play 2.0 Duets, Practice, &amp; Play https://ofstringsandkings.pointlinesurface.com/Xeno-Duets-Data-Minstrels-Machinic-Satyrs Thu, 28 Mar 2019 03:43:00 +0000 Of Strings & Kings https://ofstringsandkings.pointlinesurface.com/Xeno-Duets-Data-Minstrels-Machinic-Satyrs 2.1 Xeno-Duets &amp; Learning Systems The conception of this project came with the idea that I would perform musical duets with an AI. At the time I was exploring ancient tunings systems on some harps (lyres, specifically) that I designed with the fabrication assistance of Koumartzis Anastasios, a luthier in Greece. When grappling with some of the pernicious cultural formations discussed in the previous chapter most traditional strategies languished with insufficiency. I wanted to use musical improvisation as a way of playing with these impermeable and grim concepts. I wanted to embody the aesthetics of a pre-modern era to reflect the daunting hegemonic systems mentioned above while adopting tuning and improvisation as constructive methods that embodied agency. By adopting tuning systems that predated the Enlightenment I felt I could evoke the ancient world through song. But instead of writing counterpoint for two humans, I would be accompanied by algorithms. The third chapter will address tuning as a theme for recalibrating our cultural relations and political economy, but first this chapter will focus on learning, improvisation, play, and practice as modes for navigating indeterminacy. ︎ Refer to Appendix A4 for ancillary musical, historical, and technical details The early phases of this project were characterized by wide-eyed anthropomorphic fantasies in which I envisioned myself frolicking through the forest as a minstrel followed by a AI satyr playing gleeful harmonious lyre duets of Phoenician hymns. Of course I’m being melodramatic with my admittedly naive Brautigan-ish43 idea of cybernetic relations; however, I did fool myself into how this would actually transpire. I, too, fell into the trap of mythologizing alien learning systems with my presuppositions and apriorism. Since then my naïvete has waned measurably and my initial preconceptions matured as I continued to write the pieces of music and became more involved in the technical facets of machine learning. Performing a successful duet is contingent upon having at least some mutual understanding. Machine listening is still in its naissance and can be a bit frustrating when trying to coax out a human legible “musical” result. The learning process has not been exclusive to the machine and has required an enormous amount of trial and error. Unlike traditional software that uses structured logic to construct rule sets, these learning systems form, develop, and adapt their ‘intelligence’ contingent upon the music it is exposed to. To mitigate unintelligible musical behavior I spent dizzying amounts of time dedicated to the tedious yet paramount task of organizing musical data. The adage of ‘garbage in, garbage out’ is apt here. Unlike traditional duets where two musicians share a tonal lexicon, a common listening history, and oftentimes a musical form that they traverse in time, this duet took place over a longer intervallic timescale. The period where the dialogue occurred exceeded the classical human temporality of performing a form in simultaneity and took place in an inhuman, nonlinear temporality unfolding in the pre-critical multidimensional indices of model training. The duet was actually happening during the data selection, cleaning,44 and training rather than while I was playing music with it. I began to refer to this process as a xeno-duet (or a meta-duet) whereby I would train the model, perform music with it, and then return to retrain the model for a more “musical” output. The dynamic back and forth of recalibrating the system to bring it into an optimal state for musicality was where the most nuanced repartee took place. It was only after I internalized these second-order dynamics that the music took on a life of its own. The presuppositions upon which human musicality rests were undermined when attempting to establish even the most basic common parlance upon which to communicate. The non-computational behavioral methodologies that I considered for model integrity drew heavy influence from constructivist theories of cognitive development45 (Piaget, Bachelard, et al.), the Palo Alto Group’s work on communication theory46 (Bateson, Von Foerster, et al.), and Chilean radical constructivism47 (Varela, Maturana, et al.). The developmental cycle for arriving at decipherable outcomes was achieved by iterating through cognitive stages of selection, distinction, comparison, and connection of elements into information and phenomena forming coherent output (by my aesthetic standards). 43 &nbsp; "Richard Brautigan &gt; Machines of Loving Grace." http://www.brautigan.net/machines.html 44&nbsp; &nbsp; "Data Cleaning: Overview and Emerging Challenges." https://www.cc.gatech.edu/~xchu33/chu-papers/data-cleaning-sigmod-tutorial.pdf 45&nbsp; &nbsp; "Constructivist Foundations." https://constructivist.info 46&nbsp; &nbsp; "Bateson &amp; the Palo Alto Group | The Children &amp; Adolescents: Clinical Formulation &amp; Treatment." https://www.sciencedirect.com/topics/neuroscience/communication-theory 47&nbsp; "Radical Constructivism" http://www.radicalconstructivism.com/ https://ofstringsandkings.pointlinesurface.com/You-Can-t-Teach-an-Old-Form-New-Tricks-Overfitting-Underfitting-Bias Thu, 28 Mar 2019 03:46:16 +0000 Of Strings & Kings https://ofstringsandkings.pointlinesurface.com/You-Can-t-Teach-an-Old-Form-New-Tricks-Overfitting-Underfitting-Bias 2.2 You Can't Teach an Old Form New Tricks: Overfitting, Underfitting, Bias, and Variance Some of the particulars that I found fascinating while attempting to coax responses from the machine satyr that I deem to be musically acceptable have been the use of ‘overfitting’ and ‘underfitting’48 techniques. Overfitting happens when the neural network is very good at learning its training set, but cannot generalize beyond the training set (known as the generalization problem49 ). This results in a sort of mimicry in the system. The image below is a biased overfit model responding to the question: “if all of your friends were jumping off a bridge, would you?” <img width="555" height="436" width_o="555" height_o="436" data-src="https://freight.cargo.site/t/original/i/36fb766fdf37be18b16b0bc840306306315ac96f90cce7150827dda13625516d/IMG_0835.JPG" data-mid="38678764" border="0" src="https://freight.cargo.site/w/555/i/36fb766fdf37be18b16b0bc840306306315ac96f90cce7150827dda13625516d/IMG_0835.JPG" /> ︎ Refer to Appendix A5 for ancillary musical, historical, and technical details Overfitting is like trying to teach music by using a very specific training set. For example, listening or training only with madrigal chorale music. Someone who learns music only having listened to madrigal chorale music will learn a very specialized form of music and may not be able to stay in tune, timbre or idiom in another style. This model isn’t generalizable. Admittedly, most of the pieces in this project were overfit and deeply biased towards non-generalizable outcomes. On the other hand I experienced ‘underfitting’ which happens when the network is not able to generate accurate predictions on the training set—not to mention the validation set. This ends up with some alien accompaniment that seemingly focuses on odd, myopic facets of data that has no correlation to general anthropometric semantics. An example of underfitting could be the following: consider an AI satyr that tried to learn music solely from listening to Baroque lutenists, but ignored most of the phrasing, nuances, and general song structure and obsessed over micro-dynamics and timbre like the plucking of a string, trills, etc. This machine satyr will have an extremely limited understanding even of the notion of the Baroque musical forms, not to mention an insufficient ability to understand broader musical precepts. 48&nbsp; &nbsp; "Model Fit: Underfitting vs. Overfitting - Amazon Machine Learning." https://docs.aws.amazon.com/machine-learning/latest/dg/model-fit-underfitting-vs-overfitting.html 49&nbsp; &nbsp; "Generalization | Machine Learning | Google Developers." https://developers.google.com/machine-learning/crash-course/generalization/video-lecture https://ofstringsandkings.pointlinesurface.com/Practice-Makes-Perfect-Practice-Makes-Permanent Thu, 28 Mar 2019 03:51:56 +0000 Of Strings & Kings https://ofstringsandkings.pointlinesurface.com/Practice-Makes-Perfect-Practice-Makes-Permanent 2.3 Practice Makes Perfect Practice Makes Permanent Many of our listening habits are the result of ear training. To accept musical relationships and idioms as consonant they first must be familiar. This requires prolonged exposure over time and, in a formal domain, it takes practice and training. To achieve this at a cultural scale this requires a remarkable amount of exposure and endurance. Many 20th century composers spent their entire careers developing musical systems that rejected the tyrannies of inherited tonal systems established in the preceding centuries often at the peril of their careers. With the rapid coalescence of deep learning, platform preference analytics, and behaviorism formalized into user experience design it seems that altering culture doesn't take nearly as long as it used to. Training, whether embodied in musical practices, data modeling, or cultural engineering, still requires time and repetition to modify behavior. However, shifting the Overton window in mediated environments requires a nominal amount of uptake time for normativity to emerge. Any dedicated musician knows that they need to be careful not to practice the wrong thing. The colloquialism “practice makes perfect” is actually a stark misnomer with potentially negative consequences. It should more accurately be modified to “practice makes permanent.” Everything is practice. Habits are formed when repeated over time. Whether in the rehearsal studio or data center it doesn’t matter whether the habits are ‘good’ (true), ‘bad’ (false), or inert (neutral) they are computed into the model of the musician or the algorithm. This is what happens when we train. We cannot afford to form bad habits when we train these algorithms. As these systems are designed and developed we need to be exceedingly cautious that we aren't using the models to retrain ourselves and our social relations in ways that undermine our agency. These prediction methods supplant rationality with efficiency. Since subverting the rational mind is axiomatic to the current paradigm of attention economics it is imperative that we define healthy metrics for cultural tuning outside of consumer preference optimization. Austrian economist Ludwig von Mises’ notion that the unknown information in the world can only be excavated by the market50 has been embedded itself into deep learning data analytics. The neoliberal notion of the Market, as the singular super-information processor, positions the individual’s desires as the engine of creativity. This fracturing account of association systematically disables our ability to create common goals or problems. Supplanting humans’ sense of interconnectedness with transactional game-theoretical relations fundamentally undermines our ability to coordinate for the types of species-wide issues that we are imminently facing [e.g. ecological meltdown, precarity in transnational markets growth projections, refugee crises, etc.]. By placing the onus on the individual rather than the cultural or commercial institutions we absolve firms of proper accountability for pernicious outcomes in the public domain. If you’ll allow another outlandish sonic metaphor, this is similar to billions of simultaneous wankery guitar solos not playing in concert. When deep learning is subjugated by market dynamics it is unrealistic to expect that utility maximized consumer behavior will steer our culture to feasible explications for the impending existential and extinction crises. It is abundantly evident that under the current paradigm our culture is rehearsing for optimal market efficiency and utility maximization. This is the musical equivalent of practicing for solo recitals rather than rehearsing as an ensemble. Perhaps by developing systems that make use of reinforcement learning that orbit around common species goals we could change the tune and begin to play in concert. 50&nbsp; “Nonparametric von Mises Estimators for Entropies, Divergences and Mutual Informations | Carnegie Mellon” https://www.cs.cmu.edu/~bapoczos/articles/kandasamy15NIPS.pdf