The song machine May 20, 2016Posted by Cameron Shelley in : STV302 , add a comment
“The Song Machine” is the title of a new book by New Yorker writer John Seabrook. It is about a group of mostly Scandinavian musicians who generate many of the world’s big hit songs. In the following short interview with Steve Paikin of TVO, Seabrook discusses the gist of the book.
A couple of central themes are relevant to this blog. First, there is the matter of authenticity. As Seabrook notes, the ideal of singers writing and performing their own music was established by the Beatles. For the most part, musicians in the industry have performed songs written for them by others. The continuing power of the Beatles ideal is manifested by the concealment of the “song machine”, so that pop singers continue to appear to author their own works.
Second, much of the development of pop music has been shaped by developments in technology. The development of music over the internet, today often streamed, has broken down album as a unit of music consumption. (The album, of course, was itself a product of a previous generation of technology.) The focus of songwriting has thus fallen on the catchiness of individual songs. This focus has been served by the development of research equipment, such as the fMRI machine used to study how people’s brains react to certain aspects of musical performance.
The composition of pop singles has changed. Seabrook notes that computer composition equipment allows writers to establish how a sing should be performed even as they hash out the ideas and the lyrics. Indeed, the sonic aspect of pop music has come to dominate the aims of composition. This has opened the door wider for non-native-English speakers, such as the Scandinavians who are so central to the “song machine”, who pay less attention to the meanings and implications of their lyrics and more to the sounds and their relation to the instrumental side of the product. Of course, the ability of people to collaborate over the Internet has facilitated the transition of pop songwriting from a personal endeavor to one more like script writing, in which a dozen or more people may participate.
Perhaps the best-known technology involved in the music-making process is Autotune. This software allows producers to correct and composite different tracks from an artist’s rendition of a song in order to come up with the optimal result.
The discussion manages to avoid any dystopian undertones. That is, the “song machine” is not obviously the ultimate degeneration of music, a once human art form brought low by advances in technology. Still, it does demonstrate the continuing power of the notion of authenticity in human affairs, as well as the interaction of technological development and artistic expression.
DNN: 17 May 2016 May 17, 2016Posted by Cameron Shelley in : STV302 , comments closed
Drones can perform a variety of useful services. In some of those roles, drones act largely as intermediaries. In others, they are themselves the objects of interest. In yet other cases, the role of drones is mixed.
A good example of a drone as an intermediary is found in the recording made by a drone of Bertha, an enormous tunnel-boring machine being used to construct a highway under Seattle. Here, the drone is an invisible substitute for the human eye, as it flies through the tunnel from its beginning to the drilling face itself.
Sometimes, the role of the drone is not merely (or primarily) to see but to draw a response. Consider a program developed by the Desert Research Institute of Nevada to try cloud seeding by drone. A fixed-wing drone was used to haul aloft gear to dump silver-iodide on clouds, the idea being to prompt precipitation. The work is in a testing phase—no rain has been made yet, nor is it clear what its ultimate use will be. Still, given the level of water stress in areas of the southwest US, hopes may be flying high with the drone.
Researchers from Duke University planned to monitor elephant populations in Gabon by drone. However, the elephants were not impressed:
And it wasn’t just that the elephants noticed them; in many cases, the elephants were clearly agitated. Some of them took off running. In at least one case, an elephant used her trunk to hurl mud in the drone’s direction. “She had her baby with her,” said Missy Cummings, the director of Duke’s Robotics Lab.
It seems that the noise of the drones reminds elephants of swarms of bees, something they are not fond of. The researchers are hoping to modify the drones so that they do not make the offensive sound.
Elephants are not the only ones who may respond defensively to the presence of drones. One participant at a Russian renaissance fair shot down a drone by hurling a spear, the most advanced anti-drone weapon apparently available to him.
I note from the row of port-a-potties in the background that the fair-goers are not completely inflexible about the use of authentic technology.
In some cases, interest in drones lies in seeing them and not in what they see at all. This is suggested by a recent Mountain Dew commercial featuring drone hunting. Dirt bikers on their steeds chase after and bag a drone, rather like the mounted, fox-hunting toffs of yesteryear.
If it has not already been done IRL, it soon will be.
In any event, such a pursuit shows how interest in drones can vary from a means of observation to the object of observation themselves.
Gene drive discussion May 16, 2016Posted by Cameron Shelley in : STV203 , comments closed
As noted in an earlier post, researchers are looking for ways to implement the concept of gene drives using the CRISPR gene-editing technology. In rough terms, this technology would allow them to change the genome of many species quickly and in a more-or-less arbitrary fashion.
The most prominent application mooted for the technology is the elimination of malaria. This disease is transmitted by certain species of mosquito. A gene drive could be used to alter the genome of the mosquito so that it can no longer carry the cause of the disease. Given the harm that malaria does, such a prospect is quite inviting.
Journalist Jennifer Kahn gives an accessible overview of gene drives and the medical or ecological effects their deployment might have.
The take-home quote is the question, “Are we gods now?” I take it that this question refers to the usual concept that to modify nature is to “play God” in the sense of blasphemously changing the natural order.
That question is an interesting one but I prefer another interpretation. Aristotle argued that gods are distinguished by the ability to foresee all the consequences of their actions, as opposed to humans who are able to foresee only some. When Kahn talks about the use of one gene drive to reverse another, for example, that proposition assumes that the results of the interaction of two gene drives (or more) in nature can be anticipated with a high degree of certainty. Such an idea may look good under laboratory conditions where external factors can be controlled. However, I would be quite concerned about generalizing that notion to interactions in the wild where control is lacking. So, the idea of simply hitting “undo” after an ecological boo-boo seems somewhat dubious.
Her conclusion is that perhaps gene drives should be deployed permissively in nature. After all, there is a large and immediate benefit to be reaped, at least in the case of malaria. Of course, any harms stemming from fundamental changes in ecology might be equally as rapid, suggesting a more precautionary strategy. It is a good illustration of a fundamental dilemma of progress.
And the conversation continues.
Requiem for BlackBerry May 13, 2016Posted by Cameron Shelley in : STV302 , comments closed
Journalists Jacquie McNish and Sean Silcoff published a book called “Losing the signal: The spectacular rise and fall of BlackBerry.” I suppose, as an employee of the University of Waterloo, I should have read it by now, but other projects have come first.
In any event, the book recently won the National Business Book Award. In honour of the occasion, Ms. McNish appeared on TVO’s The Agenda with Steve Paikin. If, like me, you have not had the chance to read it, then the interview will give you some idea of what it’s about. As usual, Mr. Paikin conducts a good interview.
In spite of having lost its footing in the global phone market, BlackBerry lives on in significant ways.
Am I a computer? May 11, 2016Posted by Cameron Shelley in : STV302 , comments closed
A recent class in Artificial Intelligence at Georgia Tech actually featured an AI chat agent. “Jill Watson” was a Teaching Assistant for the course, interacting with students through the class’s Piazza chat site. Jill was actually an instance of Watson, IBM’s AI system, thus her moniker.
One day in January, Eric Wilson dashed off a message to the teaching assistants for an online course at the Georgia Institute of Technology.
“I really feel like I missed the mark in giving the correct amount of feedback,” he wrote, pleading to revise an assignment.
Thirteen minutes later, the TA responded. “Unfortunately, there is not a way to edit submitted feedback,” wrote Jill Watson, one of nine assistants for the 300-plus students.
Evidently, none of the students identified Jill as an AI.
The success of Jill is of some significance. For one thing, Jill’s ability to handle routine requests would enable human TAs to concentrate on more challenging tasks:
[Prof.] Goel estimates that within a year, Ms. Watson will be able to answer 40% of all the students’ questions, freeing the humans to tackle more complex technical or philosophical inquiries such as, “How do you define intelligence?”
Given that students post about 10,000 messages per semester to the site, that would be a significant result.
At the same time, use of an AI could prompt concerns about job losses. Will the course receive 40% fewer TAs when Jill is up and running?
Then there is the interesting point that Jill has passed a kind of Turing Test. That is, she interacted through text messages with hundreds of people without them realizing that she is an AI. Of course, the students were not aware of the situation, which is different than in Turing’s formulation. Next time, however, the students will know:
Mr. Goel plans to tell students next year that one of their TAs is a computer, but won’t say which one.
With the advent of AI chatbots, the task of distinguishing people from software will become more difficult. Soon, it may be hard to know whether or not blog posts like this one are authored by a human. By that point, will it matter?
Are Facebook photos biometric data? May 10, 2016Posted by Cameron Shelley in : STV202, STV302 , comments closed
New Scientist has an interesting article on a matter of Facebook and privacy before the courts. If you are a Facebook user, then you may well know that Facebook has some state-of-the-art facial recognition software. It uses this software in order to automatically tag people in photos that are uploaded to the service.
Last year, a group of dissatisfied users filed a suit against Facebook for its use of this technology. In Illinois, the state’s Biometric Information Privacy Act (BIPA) prohibits companies from collecting biometric data from users without their (express) consent. The complainants contend that Facebook’s facial recognition process violates this law.
Biometric data usually encompasses things like eye scans and fingerprints. However, the suit claims that Facebook is using photos in the same way, creating “faceprints”, in effect. Well, a person’s facial appearance does constitute individuating information from their physical form, so the contention enjoys prima facie plausibility.
If so, then the Internet giant might need to obtain express consent of users in order to apply this technology. If Mark Zuckerberg is correct that we live in an age of sharing, then that should not be too difficult.
In any event, Facebook is defending its practices. A move to dismiss the case on jurisdictional grounds was dismissed by a judge, so it appears that both sides will get to argue their case in court.
This situation illustrates some interesting issues. The first is the notion of biometric data. What constitutes biologically based, individuating information depends, in part, on what sorts of individuation are technologically feasible at a given point. The idea of tagging a vast album of portraits was just a fantasy only a few years ago.
It also illustrates an issue with privacy. On the view that privacy amounts to control over personal information, then it makes a big difference as to how that control is authorized. Express consent, as apparently required by BIPA, means that biometric data is under the control of its subject by default and may be accessed only when the subject gives permission. Implicit consent, which Facebook has assumed, means that biometric data is accessible by default and limited only when the subject says otherwise.
It is not yet clear what BIPA might require of Facebook. Would it have to seek users’ approval each time it tags them in a photo? That might be quite a bother for avid users. Could they give blanket consent under defined circumstances? We await further developments in the court room.
Is gene therapy finally here? May 6, 2016Posted by Cameron Shelley in : STV203 , comments closed
Gene therapy is a medical approach to treatment that, roughly speaking, involves manipulating a patient’s genes in order to improve their health. This might involve disabling a malfunctioning gene, replacing it, or even inserting a novel gene. Gene therapy has long been thought to be a natural candidate to treat genetically conditioned ailments such as cystic fibrosis.
In spite of its obvious promise, gene therapy has spent many years in the wilderness. The death of Jesse Gelsinger in a gene therapy trial in 1999 sent research into a long tail spin. However, certain forms of gene therapy have now made their way through clinical trails and are emerging into the medical marketplace. The results are mixed.
The first gene therapy in medical service is alipogene tiparvovec, known as Glybera, which is a treatment for lipoprotein lipase deficiency, a malfunction of people’s fat-processing metabolism. Glybera consists of trillions of viruses containing the correct lipoprotein lipase gene, which the viruses prompt a patient’s body to reproduce.
In spite of its arrival in clinical use, Glybera is already considered a flop. The basic problem is that lipoprotein lipase deficiency is very rare. Partly for this reason, doses of this treatment cost $1 million each. Needless to say, there have been few individuals or insurance companies lining up for it.
On the bright side, in getting approval and proving salutary, Glybera led the way for other gene therapies.
So, prospects are finally looking up. For example, a genetic therapy for severe combined immune deficiency, also known as Bubble Boy disease, is set to emerge from trials. The therapy is touted as a cure since it works by replacing the defective genes that give rise to the ailment. It has a good track record:
The treatment is different than any that’s come before because it appears to be an outright cure carried out through a genetic repair. The therapy was tested on 18 children, the first of them 15 years ago. All are still alive.
It is not yet clear what price will be asked, although officials with GlaxoSmithKline say it will not approach $1 million.
Investors and researchers in gene therapy are happy with the way things are going. Still, new developments have not stirred the hype accompanied by earlier efforts. Gene therapies may work well for ailments caused by single-gene errors. Yet, even most genetically conditioned medical problems are too complex to be addressable through the work that brought about these latest treatments.
Self-driving cars coming soon? May 5, 2016Posted by Cameron Shelley in : STV202 , comments closed
Google has recently announced a partnership with Chrysler to install its self-driving car system in the Chrysler Pacifica. The immediate aim is to further test its system with the aid of 100 modified units of the popular minivan.
Adrienne Lafrance of The Atlantic considers the project a genius move. Google has taken pains to present its self-driving car project in a non-threatening way, as evidenced in the form of its cute little bubble buggy. Who would be afraid of a car—even a robotic one—when it looks like a marshmallow?
Roughly the same logic applies to the Pacifica. As a minivan, it fits into an established, even staid, category of vehicle:
But even more than that, picking a minivan for Google’s first direct collaboration with an automaker is really about what the Chrysler Pacifica—and minivans more broadly—represents. Boring ole safety and reliability. This is the kind of car that gets a tumble of kids to and from field hockey practice without incident; it is not a vehicle you expect to see speeding down the freeway, weaving in and out of traffic.
So, the public will be quicker to feel OK with, and to trust, a self-driving car that arrives in this form.
Beyond the issue of how people will respond to self-driving cars, there is the matter of what they will do once a computer takes the wheel. Barrie Kirk, of the Canadian Automated Vehicles Centre of Excellence, speculates that people will have more sex in cars as a result. Certainly, sex in cars is nothing new, as visitors to a Lovers’ Leap or a drive-in movie theatre can attest. However, in those cases, the cars were stationary.
Tesla drivers have already recorded themselves doing odd things in their autopilot-equipped Model Ss, such as brushing their teeth or reading books. As confidence in such services increases, minds may turn to other pursuits. And, minivans seem particularly well suited to them.
So, perhaps technology is about to change our perception of minivans from boring brat haulers to mobile love lairs. One hopes that they will have tinted windows.
DNN: 4 May 2016 May 4, 2016Posted by Cameron Shelley in : STV202 , comments closed
It is interesting to observe the forces that shape the adoption of new technologies. Drones, for example, have proved challenging for regulators. However, they are being adopted readily in other social spheres.
In the arts, drones have been recruited into dance performances and they have been deployed for making graffiti. Now, Gizmag reports, Japanese advertising firm MicroAd has used drones as part of a choreographed performance with Mt. Fuji as a backdrop. The purpose of the performance is to promote this “Sky Magic” service for aerial advertising at special events. One may be coming soon to a festival near you!
Fastcompany reports on efforts to make drone racing into a spectator sport. The Drone Racing League has approached marketing firm Culprit Creative to turn drone races into broadcast-worthy events. The result is the design of a race course that allows for spectacular stunts and, of course, crashes. The aim is to be a cinematic as possible, evoking high-movie races scenes from Tron and Star Wars. If these designers can capture the look and feel of these popular experiences, then we may be seeing much more of drone racing on our screens.
Although watching drones may become a pastime, having drones watch us is already pretty well established. Some ski resorts already offer a service where drones follow skiers downhill to record their exploits. Now, Gizmag reports that there is an underwater drone that can follow divers around to record their dives. The iBubble can function autonomously for about an hour. It can also be steered remotely. I assume that many dive videos will be put on social media for others to admire.
So, the urge for both promotion and self-promotion has prompted some love for drones.
The social construction of bike share systems April 26, 2016Posted by Cameron Shelley in : STV 201, STV202 , comments closed
An interesting article has appeared in The Journal of Urban Technology: “Disassembling Bike-Sharing Systems” by Fabio Duarte, a professor of urban management at Pontifícia Universidade Católica do Paraná. In brief, the article analyzes the adoption of bike share systems (BSSs) in accord with Bijker’s account of the social construction of technology.
The article examines the social aspects of the recent success of BSSs in major world cities. Of course, bike share systems help to solve certain urban mobility problems. Yet, their acceptance has also been conditioned by non-technical factors. Duarte focusses on three:
- Electronic keys: Users of BSSs typically buy a membership in a system and are issued a unique ID to use when borrowing and returning bikes. Besides solving management issues, these keys also provide a surveillance function, tracking where and when their users are moving. BSSs are a sign that people seem to be OK with the commodification of this personal data.
- Credit cards: Payment is typically made through credit cards. Although this practice does solve a management issue, it also excludes people who do not have access to credit cards, e.g., the urban poor. It is worth noting that there are exceptions, such as Philadelphia’s BCycle, which accepts cash.
- Advertising: BSSs are often subsidized by private companies, especially banks, which then get to place their logos prominently on the bikes. Thus, the bikes become traveling billboards for the sponsors. In some cases, this move helps sponsors to skirt restrictive public advertising bylaws. Also, it means that BSSs are more likely to be found in urban areas where there are customers wealthy enough to afford banking services.
As Duarte stresses, none of these points mean that BSSs are nasty or ought not to exist. It simply means that their design is the result of social as well as technical criteria.