Machine learning creates professional level photographs

Discussion in 'Philosophy' started by movingfinger, Jul 15, 2017.

  1. Those of us with a few years under our belts will remember a commercial back in the cassette tape days that asked "Is it real or is it memorex?". It showed a singer who could hit notes that would break a glass, then the memorex recording of the singer did the same thing. 'Real' implied it was done by a human. Now-a-days we ask "Is it real or is it AI?" That summarizes the Turing Test.
     
  2. How can a machine sense a moment of love, fear, awe, juxtaposition, grace, happiness, spirit, ? It would have to have a sense of itself. Otherwise it's just a camera shooting licenses of cars going through toll gates.
     
  3. Machines may be smart but they don't crave. Don't worry folks in the future only two kinds of people will still be employed: artists and interfaces.
     
  4. How does A HUMAN sense a moment of . . . ?

    Though humans have a sense of themselves, I don't think that's key in recognizing moments of, for example, happiness. We often go off of signs. We see a smile, we recognize happiness. A machine can be programmed to do that.

    Also, the machine wouldn't have to "recognize" happiness. It just has to learn how to create a picture that conveys it. For that, it doesn't need to FEEL happiness. It just needs to know what happiness looks like. A machine can learn, for example, that dark shadows, tears on a face, and falling rain don't signal happiness, but that bright sunshine, a smile on the face, and a field of daisies do.

    This is why photographers don't have to feel the feeling they're creating with a photo when they're shooting it. I recently read about a musician who said he could only write depressing songs when he was not depressed. When he was depressed, he was too depressed to write music. A photographer will work with visual signs and symbols to create a certain feel. A machine can be programmed to put together that kind of sign-symbol visual composition.

    I think this works two ways. It not only shows how we can "humanize" machines. It shows how programmed and mechanized the human brain and emotional reactions already are. The hard part for some to swallow is that discoveries about how the brain, consciousness, and emotions work demythologizes and deromanticizes those aspects of human life.
     
    sjmurray likes this.
  5. I understand your point Fred. But aren't you describing mimicry? I'm struck by how many pictures we can look at and then one particular one just stands out. Most viewers can feel it. But when they describe it, their words don't sound much different than the words they use to describe similar but less emotional and impactful pictures. That one picture is extra special yet indescribable. If we can't describe the differences among ourselves, how can you program a computer to see it or create it? If art truly comes from the soul, how could a soulless object create art? Could a computer catch the looks and feelings among the people in your Plowshare series? I doubt it.
     
  6. I never say always:)
     
  7. I don't think, we can design a machine to create feelings or emotions as yet, but as Fred said, and Alan rightly pointed out, we can make a machine mimic or simulate the effect of such emotions in pictures. How will then a machine produce that special picture out of many similar ones. The answer is in numbers, and selection. Many photos can be composed covering a specific mood such as happiness (which a machine can do efficiently), and that increases the possibility of creating a special picture. Photography after all is a product of skill, serendipity and creativity. While a human director can provide the creativity, the machine can bring in the skill, and the serendipity can be in numbers and permutations. Together, it has the potential to create something special that we can connect to.
     
  8. Since I think art can be made by machines, obviously i think it doesn't take a soul to make art. As an aside, I think the idea of soul is a made up fantasy, so having one is not necessary for anything.
    Yes! If I set up several nanny cams, it would catch plenty of what I catch. I will do it with more economy because I'll intuit when to turn my camera on to capture a good moment. The nanny cam will capture many wasted as well as good moments. Then we just program some signs for the computer attached to the nanny cam to look for in making its selection. When photographing at Plowshare I'm not often on the lookout for feelings so much as I am looking for signs of feeling. When I see one person reach out a hand toward another, I trip the shutter because that's a sign. There are plenty of times lots of feeling is present but I don't get a good picture because I haven't captured a sign of it. All photographers experience that. So, no, I don't believe we capture feelings by camera. I believe we capture signs and symbols of feelings and I believe a machine could be programmed to do that.

    While there are certain dangers in the power that machines can have, one possible benefit is in getting humans to have a little more humility. As we've evolved, we've become more humbled by how we view ourselves relative to a much more powerful environment and by how we see ourselves in relation to animals. Machines have already been a big step in how we see ourselves and that will likely increase as we build more and more sophisticated ones.

    Humans like thinking they have all the control, though they've come up with the idea of God as a foil to all that human power. But certainly our shifting sense of power over the environment and animals has humbled us a bit, which I think is a good thing. I welcome at least that aspect of machines, though it works both ways because we can be proud that we've created such incredibly powerful tools while being humbled by what they can do better than we can, such as making quick calculations.

    That a machine can create art doesn't lessen human art-making any more than a machine's ability to calculate has lessened our respect for people who can calculate what they need to calculate in order to make machines that can calculate. My art-making isn't threatened by other people who also make art. As a matter of fact, knowing other people make art enhances my own endeavors in that field. I'm not threatened by machines doing it either.
     
    Supriyo likes this.
  9. Mozart was 5 when he started making art. I'm not sure it was his lived experience doing much of the heavy lifting.
     
  10. Huh? What I said was that, because I think souls don't exist, I don't think machines have one. And I don't think souls are necessary to make art. I didn't say the CONCEPT or FANTASY of souls would be irrelevant to art. I said HAVING ONE is. I don't think Mad Hatters exist, but obviously the concept of Mad Hatters is important to understanding some art and that same concept could help enrich future art. All that has nothing to do with an art-maker having a soul, which was Alan's suggestion.

    I don't think art makers have souls, though some effectively use and understand the concept and fantasy of souls.

    Whether souls are fantasy has a lot to do with whether artists have to HAVE them to make art, regardless of whether they can think about, have faith in, or use the concept of them in making art.
     
    Last edited: Jul 24, 2017
  11. You misunderstood me. I hope I've cleared it up. I said HAVING A SOUL isn't necessary to the making of art.
     
  12. The relevance of the soul being a fantasy is to HAVING A SOUL, not to the concept of soul as it relates to art throughout history.
     
  13. I think Picasso addresses this . . .
    There's something about lived experience that often works against art. We see so many artists whose earlier work is more significant than their later work, and I think the innocence and experimentation present in those with less experience plays a role. I think there's something new, fresh, and experimental about machine-generated art that isn't dismissible but is, rather, exciting.
     
    sjmurray likes this.
  14. I don't necessarily think Mozart at age 5 "knew" how to use musical symbols and signs effectively. I think he had an innate or given sense of it. In an important sense, Mozart was probably more like a music-making machine than he was like most other humans. I say that as a compliment and don't think it detracts from his own humanity. It's a description of it.
     
  15. Actually, Phil, I'd phrase it a little differently. I'd say it's not the human-made art that has soul. It's the homunculus lurking inside the human-made art that has the soul! ;-)
     
    Landrum Kelly likes this.
  16. Phil, I think you're making it too complicated.

    Machines follow the rules. Artists don't (have to).
     
  17. It's clear from the article that this AI is processing photos taken by people, based on some norms it's "learned" from a dataset of photos taken by people, deemed by people to have a statistically "correct" internal computer representation. Granted that this is cooler than just applying auto-levels, and I'm sure eventually we'll have cameras with those kind of smarts for the benefit of interfaces (aka people) without their own taste, but that's a long ways off from a machine that knows what a photograph is, much less wants to take one. Do not forget in these AI scenarios that machines don't want to take photos, or cure cancer, or enslave mankind. Machines don't know anything and therefore are interested in nothing. Not even to be powered on. In that sense the most powerful supercomputer is not even as smart as an amoeba, which at least knows what it likes. Maybe you can attach a smart camera to a roomba to take pictures "autonomously", but they'll only be good pictures by accident, and then only because a sentient human being said they're good. Is this technology smarter than artfully dropping your camera on its shutter? Yeah, but not essentially.
     
    Last edited: Jul 24, 2017
  18. Some of the reasons that the results will be so fascinating and open up new possibilities for art.
    Photographers are often thrilled by accidents that happen when they're taking pics. Couldn't machine accidents be thrilling, too?
    The camera you artfully drop hasn't been extensively programmed to act in certain ways when dropped. Presumably, an AI camera would be carefully and thoroughly programmed. We can call that "smarter" or not, but it's definitely something!
     
  19. I didn't want to start a side discussion about God by mentioning soul. So let me change my sentence and say that art comes from the heart and from human feeling and cameras as well as computers have no feelings, no heart. Even if you could input a kind of artistic intelligence into a computer, it would be a program by one person or group of people who programmed the computer to their way of seeing. It would be their art, used over and over again in the same way. What if they're great programmers but lousy photographers? Lousy artists"?

    Which raises another point. Do we want to live in a world where there would be art but no artists? How sterile. In spite of Photoshop, or maybe because of it, we have millions of people who can try to express themselves even if their art is liked only by their mothers. I think we have a need to express our feelings and our sense of aesthetics. "Nice photograph, son."
     

Share This Page