#16: Biochemistry of love | Hands-on VR | Bad Peppa, Bad Google
art | science | digital culture | design | creativity | words | tech
art | science
The biochemical nature of love
A current exhibition in New York by Heather Dewey-Hagborg explores the biochemistry of love through the stages of desire, attachment, and grief. At the Temperature of My Body consists of three rich works of “speculative narrative” projecting real science into a not-implausible future.
[Spirit Molecule (2018-19) by Heather Dewey-Hagborg]
From the catalogue:
Spirit Molecule is an ongoing research project with artist and botanist Phillip Andrew Lewis, attempting to engineer psychoactive plants to contain human DNA. Upon entering the gallery, visitors walk through a living greenhouse and a functional DIY genetic engineering lab staffed by a working scientist. If the project is successful, a plant engineered with the DNA of a lost loved one could be consumed as a genetic memorial.
I haven’t seen it in person but wish I could! The portfolio of the show is well worth a look (PDF here), and if you go to the gallery website, you can read a lot of press about the show. It’s better to read some of the accounts of people who’ve seen it in person than my inferences here.
design | technology
Touching the virtual world
A new paper published in Scientific Reports, one of the Nature stable of journals, describes a really interesting new project (shown above).
In my Wearable Technology class, students often come up with the idea of making haptic feedback gloves—gloves that respond to some trigger in a way that the wearer can feel. The idea has been toyed with for decades and there are numerous examples of them at the experimental and commercial stages. Unfortunately, many of the students neglect to research that history and don’t quite appreciate the technical challenges to innovate in this area. They are significant and there are ongoing research developments.
One of the issues of haptic glove feedback is that the actuators in the gloves can’t reproduce any sense of texture. The researchers in this case have made soft actuators that can wrap around skin, and activate them pneumatically. Typically, pneumatic actuators need bulky and often loud air compressors but in this work the pneumatics are controlled via electrostatic effects. (Think static electricity put to better use than making your hair stand on end or zapping you as you touch metal after walking across carpet.)
The actuator is just 15mm in diameter, creates presses against the fingertips of about 0.1mm (enough to feel convincing), and can react in a few hundred milliseconds. The researchers tested the system through a virtual chess interface in which a person could feel themselves picking up and moving a chess piece.
As it develops, this kind of tech is really going to move virtual reality beyond its current emphasis on visual representation, which is never going to be completely effective when decoupled from the other senses.
digital culture
Second-order algorithms gone awry
Look closer. Yep, at that image of Peppa being tortured at the dentist, in the second row.
If you’re a parent, you have almost certainly been warned about the dangers of letting kids wander YouTube (even the kids version) because of some of the videos they’ll find. There are a huge number of videos of children’s cartoon characters being assaulted, tortured, and involved in definitely-not-kid-friendly acts.
I personally haven’t come across it in my attempts to find these videos by clicking from recommended video to recommended video to recommended video but plenty of people have. There’s a good discussion of the problem in James Bridle’s essay “Something is wrong on the internet” from November 2017. He revisited the topic and it doesn’t seem to have gotten much better, writing that the people involved at YouTube, though well-meaning, don’t seem to recognise the significance of the problem.
I happened to encounter a flow-on effect of this the other day. When I did a Google search for Peppa Pig, the summary image mosaic shown above appeared. The image of Peppa being tortured at the dentist has been used as an illustrative image in many articles about the YouTube problem and is why it appears here.
Google’s search algorithm can’t tell that this image comes from discussion of a problem that it has caused through its other algorithms (Google owns YouTube). I’m calling this a second-order algorithmic failure as it’s one algorithm reinforcing a failure of a previous algorithm. With this kind of process in place—algorithms feeding other algorithms—not only do we have to deal with the unanticipated consequences of algorithms, those consequences might spread far beyond what is reasonably controllable, and certainly not within the scope of any particular algorithm “owner”. It’s an emergent collective effect that seems particularly difficult to solve.
What do we do about it? One starting point is to recognise that there are flow-on effects from a bad algorithm and to realise that there is even more responsibility to get things right at the first step. Will tech companies accept that responsibility? Probably not if they are driven almost solely by a profit motive, as most of them have demonstrated they are, despite public claims they make. I don’t know the answer to this.
Might the solution be in a competition of algorithms? It feels like there could be more to learn from the idea of a 1984 computer game for uber-geeks called “Core War”.
In it, players write pieces of code (algorithms) that battle to dominate a programming environment. It’s insanely geeky but the kind of thing that I suspect many of the algorithm writers of today would find interesting and fun if they know about it. I just don’t know how many of the programming cohort being hired by the big tech companies, with their well-established bias for hiring young coders, will know about it given that it existed well before they were born.
Let the battle of algorithms begin! (Actually, it’s well underway and the bad guys are winning…)
If you’re reading this on the website or have had this forwarded to you by email, would you think about subscribing (free!) so I have a better sense of how many people are finding the newsletter useful? As always, let me know if you have comments or reactions directly at physicsdavid@gmail.com.