Royal Society Science Matters panel discussion, “Machine Learning & Artificial Intelligence”
Some interesting things that came up in the video I watched of the Royal Society Science Matters panel discussion, “Machine Learning & Artificial Intelligence”, 10 Jan 2017:
Joanna Bryson, Reader in AI Ethics at the University of Bath makes some interesting points about ‘projections’ humans themselves make on the idea of something human-like. For instance, words that come up when discussing general AI are ‘intelligence’, or ‘consciousness’, but Joanna claims these terms are often ill defined in such discussions and that humans are not that capable in these supposedly human-like abilities anyway. For instance, if intelligence is the ability to “act in the right way in different situations”, as in adapt to surprising situations in a dynamic world, then plants or automatic thermostats are intelligent. Plants can grow towards the sunlight depending on where it happens to be coming from, and thermostats can adapts to any change in temperature to adjust heating accordingly. One interesting thing she mentioned is concerning consciousness. If consciousness is to be ‘self-aware’ then maybe machines suffer from too much self awareness! She says a computer is aware of and can access any part of its memory at any time – we humans cannot, and maybe it is our inability to access all our knowledge or the ways in which we function that is more what we mean by consciousness (i.e, kind of the opposite of self awareness, a lacking of it).
Coming from a background in psychology, what Bryson seems to be proposing is that we have too little intuition about what makes us human, what are our human-like attributes (perhaps even particularly an intuition of our ‘failings’, that we are in fact not very intelligent and not very self-aware, and that maybe this is important as to what constitutes human-ness) to make something human-like artificially.
What I might do during and after Tate concerning performance practice
Tate is a good opportunity to show some work in “Oral Tradition” and experiment with making work in public.
Bring: 3 usb pens with 4 different films | Laptop | Phone charger and usb transfer | Ext Hard Drives | selfie stick | Lapel mics from uni | A note for the film
Go around the space experimenting with making videos with selfie stick and headphones/lapel mic. Ask a couple of studio mates to film me telling one of the video stories, to experiment with repetition. Use the toilets as dressing rooms. There can be a note by the screenings of my films asking audience to ‘help me make work’ and to ‘find me’ walking around the Tate. I’m lucky to be attending the pre-press thingy tomorrow because I can experiment a bit before the show opens and see how best to show the video work as well.
I’m still finding it really difficult to discuss this strand of my practice with people, or to feel motivated to do it myself for some sort of crippling lack of confidence, but recently I felt encouraged by some pieces: I had another listen to The Ovine Principle and realised that the four recent works I filmed in studio were all made in something like half an hour in the studio. I’m trying to take a look at what is impeding the development of this practice.
- I’ve been troubled by the feeling that I have nowhere to film. Troubled by the idea of duplicating the background too much, i.e, making too many videos in the studio with the same walls behind me. But maybe this isn’t too important.
- Maybe I need to make the whole procedure a lot more portable and less fussy. So my room or studio don’t produce perfectly beautiful backgrounds. Maybe that doesn’t matter in my work? Try film even without perfect backgrounds.
- Try make more sound-only performances. There is a distinct quality to those, and I am often more skilfull in voice alteration because I can focus on voice acting alone. Having a good sound recorder about my room, which is so quiet it makes a great recording studio.
- Maybe, on the other hand, I need to try planning a performance. This has worked before, with things like The Widow or Technicalities. Sometimes it might be worthwhile to systematically choose an existing character to dress up as again, or work on an idea of a look I might like to try. Get the right look (looking a bit other than myself can help me play, act) and go somewhere specific, and allocate time for just that, without fuss about quality of results (this is really my main obstacle).
- I seem to find it very difficult to do this in front of people I know, especially ones close to me, and it is usually these people that are most convenient to ask for help holding the camera! I can’t shake off the awareness of my usual relationship to people I already know and find it so hard to snap into another character.
- This is why I want to try work with visitors to Tate, strangers to me. In future Oral Tradition proposals I can keep trying to collaborate with strangers.
- Work alone, that is the easiest option at the moment.
- I am afraid of exhausting myself making lots of videos I will just cringe at after. I must meditate on this and manage my expectations. Think not what can I do for the performance, what can the performance do for me! If it really serves an outlet for me, for example I choose to develop a character I actually currently feel jealous of, then even experimental performances will be fulfilling enough for me to do them. The focus must return to the excitement of doing it rather than aiming for watchworthy results. It is actually past successes which have made this hard for me, I am more concerned with more success that the experience of performing, and that’s losing the point (from my point of view as artist)
- I find the process exhausting because I am trying to reinvent a whole new story from scratch every time. It has been said to me by two people that I needn’t put such a strain on myself, that it would be worthwhile to repeat stories and to develop specific characters, strengthening what’s already there, but I am not sure. I distrust this advice for some reason, and cannot decide whether I am just being cowardly or whether there is something in telling a whole story from scratch that is rather the point of this practice. Either point might be right but I haven’t given the repetition idea a fair try, even though it interests me theoretically. Maybe rewatch some material and pay attention to something in there that tickles me currently. Maybe there is a character I’d like to be again. I think altering appearance slightly, and the way I feel in front of a mirror might help get me in the mood. Maybe music also can, actually, I’ve often used music before/during a performance to help me feel like somebody else.
- Recording in outdoor spaces is a bit difficult because I am overcome by embarrassment at being seen filming myself and drawing attention to myself. It seems really silly but it makes me very self conscious. There have been times that I have recorded outdoors. “Cover yourself”, “All Over the World”, different sound works recorded on my phone, or anything I’ve filmed with another person.
- Cover Yourself was filmed in a secluded place, a cemetery. I could start trying to purposefully elect secluded outdoor settings to film in. When I go about errands, to pay attention to spots that may work for performance. Use selfie stick in secluded places where I don’t feel self conscious being seen with one, ha! Hopefully one day I will stop being embarrassed about such minor things but currently I respect it because it’s such an obstacle. Get a better phone to film with.
- Sound works on phone where I either pretend I am actually speaking on the phone, or merely recording otherwise, draw less attention
- All the other examples involved somebody else with me, holding a camera, and that helped make what I was doing look like a thing, so then it was easier. This has worked with my dad (with difficulty), or friends that I shared some creative affinity with me but who weren’t super close to me so I wasn’t afraid of embarrassment…
- Using a lapel mic attached to smartphone could be an interesting way to make work outdoors. I might like diversifying where I make work and figuring out just how diverse my studio can be, especially because I am somebody who likes to be out and about so much. If I could marry performance work with being mobile, I’d be happy.
I do feel I have to systematically think about this. Something isn’t right about how I feel about this strand of my practice. On the one hand it seems I have something of a knack for acting and that it can be powerful when implemented with my (oral) writing, and both I and others seem to see something in some of the videos. On the other hand I can feel incredibly demotivated about this, and feel like it’s all a waste of time, foolish, and pointless, with too much investment for little return (the return is usually just great feelings of unease and a ‘what happened?’ kind of disorientation). Going forward after my Tate experiment, I want to allocate a little bit of time to actual experimentation and not only sporadic executions of things I expect to be perfect. No wonder I am deterred from playing with this.
Going forward with Oral Tradition
Here is then a summary of new attitudes and approaches I hope will help me welcome the Oral Tradition experiments into regular practice.
- To allocate time regularly and specifically for performing
- Try film even without beautiful backgrounds
- Try make sound-only performances
- Having a quality sound recorder about my room
- Try a more planned performance, with a worked out look and location
- Work alone
- Experiment with strangers filming
- Meditate on characters that can serve as an outlet for me
- Do not be afraid to re-watch past material, either already published or unedited
- Try being the same character again but playfully
- Use music, and/or slightly altered appearance to get in the mood
- Plan suitably secluded outdoor settings to film in
- When outdoors, take selfie stick & lapel mic
- Involve somebody to film in planned performances, if such a person can be found
“Through the Conduit of Language”
This was interesting. Enacting possession by a fictional character. This is a third take, and I might try to repeat this story to see if it develops, maybe do this one at Tate.
*Note: my reformed ideas about approaches to oral performance practice worked in this case. I tried out filming in my room with the new setup (selfie stick), and it worked. I made a performance.
Susan Sontag: Against Interpretation (1964)
On the one hand: “Interpretation is a radical strategy for conserving an old text, which is thought too precious to repudiate, by revamping it.” On the other, it just destroys the artwork by implying a need to replace it with the interpretation, Sontag claims.
She makes the connection that the interpretation of texts in Western thought naturally implies the idea of a text having a ‘content’ and ‘form’, and that interpretation is finding the true content underlying a more literal one. Sontag compares Marx and Freud, who look at the events of revolution and war in the case of the former, and individual psychoses in the case of the latter, as symptomatic of a deeper reading of the issue, an underlying drive for these effects – an assumption of content and form.
“In some cultural contexts, interpretation is a liberating act. It is a means of revising, of transvaluing, of escaping the dead past. In other cultural contexts, it is reactionary, impertinent, cowardly, stifling.”
“By reducing the work of art to its content and then interpreting that, one tames the work of art. Interpretation makes art manageable, conformable.”
This may be true, but what if interpretation is treated as an art? Interpretation as the creation of a new thing, text, that correlates with the one it talks about – rather than a replacement, i.e. “this is what the work really means…”
“Sometimes a writer will be so uneasy before the naked power of his art that he will install within the work itself—albeit with a little shyness, a touch of the good taste of irony—the clear and explicit interpretation of it. Thomas Mann is an example of such an overcooperative author. In the case of more stubborn authors, the critic is only too happy to perform the job.“
(Feels like it’s directed at me…)
“But the temptation to interpret Marienbad should be resisted. What matters in Marienbad is the pure, untranslatable, sensuous immediacy of some of its images, and its rigorous if narrow solutions to certain problems of cinematic form.”
Can these sensuous experiences not be simply more immediate readings, but readings no less? Is reading the same as interpreting, as I think I have assumed so far? Maybe Sontag and I don’t have the same idea about the word ‘interpretation’. To her it seems to be a means of translating one work into a ‘manageable’ account of its meanings and their value. Interpret to me merely meant to take a subjective position or processing of a work, making it ‘mine’. To rewrite it in my own terms, merely by experiencing it.
“What is needed, first, is more attention to form in art. If excessive stress on content provokes the arrogance of interpretation, more extended and more thorough descriptions of form would silence.“
But of course she mentions… “(It [programmatic avant-gardism] also perpetuates the very distinction between form and content which is, ultimately, an illusion.)”
“Transparence is the highest, most liberating value in art and in criticism today. Transparence means experiencing the luminousness of the thing in itself, of things being what they are […] Interpretation takes the sensory experience of the work of art for granted, and proceeds from there.”
“The function of criticism should be to show how it is what it is, even that it is what it is, rather than to show what it means.”
“In place of a hermeneutics we need an erotics of art”
Lisa Zunshine (Cognitive Literary Theory, Theory of Mind)
Lisa Zunshine is an English professor at University of Kentucky.
Theory of mind is an area of cognitive science that studies the social skill of ‘mind-reading’, which is not to say telepathy, but the various ways in which we guess or predict someone else’s internal mental states. According to Zunshine, whereas “theory of mind evolved to track mental states involved in real-life social interactions, […] on some level our mind-reading adaptations do not distinguish between the mental states of real people and of fictional characters […] The pleasure afforded by following minds on the page is thus to a significant degree a social pleasure […]” (from her website)
And Joshua Landy:
“each work, in other words, contains within itself a manual for reading, a set of implicit instructions on how it may best be used.”
Oral Tradition (planned) ideas
- A muse devoted to the task of supporting her artist partner (a bit like Jennifer Lawrence in Mother!) talks about him and barely about herself. She talks about his tendencies and what she does for him. He’s so busy. He is always working, even when it looks like he’s resting, he’s getting ideas that will flow into his work just by looking at you. Macho art cliches. She is unwaveringly committed. But it is apparent that he is nowhere in the video, he seems absent by his silence and the amount of her talking. She talks more about what she does for him. It turns out she does quite a lot. It turns out she has a part in some of the making and decision making. It turns out that because he is so busy, she sometimes takes over for him. It turns out he is a figment of her imagination. It turns out he’s her muse.
TECHNE Proposal and PhD RCA
I received offers from Melbourne, Goldsmiths and the Royal College of Art to undertake a PhD in art practice. This is highly successful, especially considering that I am and have been going through perhaps the worst depressive episodes of my life in the time I was applying.
I have deferred my offers from Goldsmiths and Melbourne for 2019 to be able to apply for a scholarship then. But the RCA are offering me to apply for the AHRC funded TECHNE scholarship now. I am dependent upon full funding to take up the position now.
There are a few bits I am admittedly stuck on on the application and they have to do with concrete timelines and promises – I can’t as of yet say what work I will make. Therefore, it is hard for me to fill out the ‘Training and Resources’ sections regarding what I am going to need.
My tutor also recently looked at my proposal to the RCA and said that it reads ‘woolly’ and lacks clear research questions, methodology and outcomes. It worries me because I thought it did have all those things, and the questions seem to me backed up by plenty of previous research. Actually the research question seems a bit absurd on its own, but it has its basis on accepted theories in other disciplines that makes it worthwhile asking, if at least to probe the theories it stems from.
Maybe I can handle both, the criticism about a lack of substance to this research and the barrier to imagining future projects and how I might concretely spend my time during the PhD, by reading my scholarship application and extracting the questions, research, methodologies and outcomes in the plainest terms:
- In disciplines where an agent, person or character is studied or developed (for example genetics, philosophy, psychology, sociology, performance studies, biography, acting, artificial intelligence studies, fiction), what are the assumed prerequisites that make up a person in each field and how do they compare? Do they radically differ on their emphasis on what constitutes personhood, or are there features common to all? E.g, how does and actor produce the sense of a living person? How does an AI theorist speculate on the conditions for general AI? How does a biographer write a faithful account of a life and does that reanimate the subject of the biography in any way? What does a writer have to write to conjure a compelling character?
- Now I am also concerned not only with looking at the various theories and assumptions about personhood that must naturally come out of the fact that these disciplines are obliged to describe the agents or characters that they study or produce. What are the kinds of agencies we live with – can this notion be expanded to include humans, animals, fictional characters, robots, deceased persons who only exist as stories? And then what assumptions are made in various circumstances that evaluate the degree of their status as persons, as autonomous and authentic selves? Can these assumptions be challenged by the theories mentioned above, theories about what it takes to make a person?
- What features of personhood do fictional characters share with human beings, and especially with the human beings that interact with them as authors or readers, creators or spectators. Can these similarities be probed to design fictional characters that seem like AI? What would happen if you ran a literary corpus such as Proust’s In Search of Lost Time or an autobiography through a computer neural network and then asked it questions as though to ask the author himself?
- Can a person be thought to be written by its genes and its environment, and in that case, can a person write a person?
Theories to back up questions
- I ask ‘what makes up a person’ rather than ‘what a person is’ as a consequence of my preference for describing ontological positions as performative rather than essential. I think of persons as produced by certain conditions and as functioning to produce the world in turn. This follows from Foucault, Butler, Blanchot, Derrida. It is an approach to ontology that is inclusive of the relativity between things and how what makes something what it is is always subject to change, whereas essentialism is problematic because it depends on the existence of some inner essence, some constant factor that exists in the core of a thing, and yet it seems impossible to find such a thing – in ourselves or in objects.
- There is the idea in theories of constructed selves and objects that relativity can produce the effect of essentialism, and produce identity, for example, the various biological and sociological details of one’s life coalesce to form an impression of an “I”, which is firm and palpable but inessential and always subject to change, sometimes radical change (for example Douglas Hofstadter’s notion of epiphenomena)
- The theory that an organism develops from its genetic information, a text that is parsed and interpreted by ribosomes to produce effects in the body
- Machine learning has endowed computers with equal to or by some reports better image and sound recognition abilities when compared to humans. This was achieved by immersing the virtual neural network (a computer program) in human context (large amounts of image/voice data), suggesting that this context is something that must be fed through the system we would wish to engage with as a fellow agent
- The theory that it only takes very simple rules to create complexity