What is Performance Reality, Really?
CounterPulse Festival's Kinetech Arts Merges Dancers, Audience and Artificial Intelligence: Interview with Daiane Lopes da Silva and Weidong Yang
The upcoming CounterPulse Festival in March asks audiences to "join in the experiment of festival, both human-scale and mind bending." One of the Festival's performances, Super Lab by Kinetech Arts, which epitomizes the spirit of experimentation and mind bending, is directed by co-founders Daiane Lopes da Silva and Weidong Yang. Daiane is a dancer and choreographer. Weidong Yan is a dancer and a quantum physicist who is CEO of Kinevix, a virtual reality company. BayDance.com had the pleasure of interviewing Daiane and Weidong.
BayDance.com: Weidong, in your job you work with virtual reality. How does that work? What does that have to do with dance and Super Lab?
Weidong: At Kineviz we map data into visual elements and dynamics that we can appreciate intuitively. I visualize data, turning data into something we can sense. We borrow a lot of research from Kinetech Arts, artistic practice, and turn it into data visualizations.
BD: By data you mean numbers?
BD: And turn that into movement?
Weidong: Into movements, colors, vibrations.
BD: I'm trying to imagine that...
BD: How did you get into dance?
Weidong: After I had been taking dance lessons, after awhile my teacher told me that when I started I had two left feet. But I kept practicing and taking lessons, and I got better.
Daiane : He's fascinated by things he doesn't know.
BD: That's a good attitude to have, but aren't dance steps mathematical progressions?
Daiane: (smiling) He's a physicist.
Weidong: I like to think that dance represents a state of movement, or state of the body, that cannot be described by any other media, including words or physics. It defies physics. I'm a physicist. I once read a book about dance written by a physicist working with American Ballet Theater. He made a lot of observations about how physics works in dancing. But as a dancer, I know from experience that what he described was wrong. So it's something that can't be described by physics because it involves so many dimensions, that need to come together in your body, that it's an eclectic, emergent behavior, that it cannot be described by some other principles.
Daiane: It can be described, but every person will describe it differently.
BD: How do you describe it?
Daiane: Dance? It depends on what I'm looking at, right? Every dance is different. But, you know, when I watch dance, depending on the knowledge of the performer, the theater, the setup, the space, the time-what did I do before that changed my perception of what I'm seeing. That's the power of dance, that I can look at it with my own interpretation and my own bias or my own dreams, and kind of just let my imagination figure out what the heck those dancers are doing, what thy're trying to express, and what I feel about what they're doing there on the stage. It changes-if it's in the morning, if it's in the evening. Depending on what's going on in your own life, you'll see different things in the same dance. So, I guess you're right. It cannot be described.
BD: I know ballet fans who go to different performances of the same ballet so see how they're danced by different dancers, so the audience gets a different impression.
Daiane: Even if they're doing the same steps, they're creating different magic. It's a real, if I may say, spiritual experience. It can be a spiritual experience, just to sit and watch a dance performance. Sometimes you might feel things that you can't really understand why you're feeling that way. If you're laughing, if you're crying. Getting in touch with these feelings is a healing process. It's very healing. You don't have to understand, you just have to get in touch with the feelings that arise when you are watching dance.
BD: What are you hoping to convey with Super Lab for the CounterPulse Festival? It's about the future of artificial intelligence. People communicating with cell phones or whatever. What the heck are you guys doin'?
Daiane and Weidong: (laughing)
Weidong: For me it's about what questions we should be asking.
Weidong: In 2016 we started collecting tweets for the election, and from those tweets we did some experiments. One experiment ended up as a brainwashing machine that we performed at Yerba Buena last year, where we discovered the labels people use to label each other. It kind of reminds me of how the Communist party brainwashes people in China. There's an unsettling resemblance. So that's what prompted us to create this brainwashing machine, to let people brainwash themselves with the labels they used in the campaign. We also discovered that the machinery of thought, of the fake Twitter handles, was taking over communications on the Twitter space about the election. We were working in collaboration with data scientists from the UC Davis data science team. They found that the propertion of fake tweets from bots went up dramatically as the election approached. So there is a heavy influence of artificial forms of information in life.
BD: Russian bots?
Weidong: Russian bots and others. There was an article in Newsweek that they discovered an ultra-right group orchestrated a Twitter bot attack that took down Al Franken. So, with that you can think about a future where this machine-learning AI mixed with social networking that affects how we perceive things-what we believe is true, what we believe is right. At the same time, computing power is doubling every year. Right now, this tablet represents about 1% of our brain's ability to calculate, to think and possess information. If it doubles every year, in fifteen years we can have a device like this as powerful as the human brain. What does it mean for us? It means how we experience things, how we experience each other, how we communicate. So those are some really unsettling questions then that we are just trying to explore.
Daiane: I hope we will still be dancing. (laughing)
BD: We're a long way from dancing robots. What do you think?
Weidong: Ten years ago I thought it would take 50 years for computers to beat humans at playing go. Go requires intuition to play it well. Computers are really good at deriving things, but when it comes to creativity, to intuition, that's where humans are really good. But in three to four years time, a computer beat the best human go players. So that was a wake-up call for me-Okay, don't take things for granted. If computers can already do things that predominantly rely on intuition, what's the next step? In the last couple of days we started researching about AI composers and AI compositions. We came across a few scores composed by artificial intelligence. Beautiful!
Weidong: So we had people perform the AI-composed music.
Daiane: We played the music and asked people in our labs, was this made by AI or was it made by humans. They couldn't guess.
BD: People couldn't tell the difference?
Daiane: No. It's scary.
Weidong: When it comes to dance, we think of a robot as a machine walking. But if you look at the robot that came out of MIT, trying to stand up, trying to learn how to walk, it really looks like animals. So you can see, it's not hard to imagine that in the future you can have some robot doing some highly sophisticated aerial movements. Jumping up and down, turning, better than the best trained dancer. So I think it comes down to human expression, but in terms of the basic ability, I don't think we can compete with the machines.
BD: That sounds frightening.
Weidong: Yeah. (laughing) I'm frightened.
Daiane: I don't like to think robots will replace us. Of course, as a dancer I don't want to think that will be real. It's unthinkable to me. To think that they will be able to do a better job than us, because...would it be touching to watch robots doing acrobatics? Who cares? I want to be touched. Can it touch me?
BD: There's a parallel, I think, in figure skating at the Olympics. I saw a news report that some American skaters are not doing so well because they are trained traditional ly to perform beautifully. But the new emphasis is on technicalities, how many jumps and turns skaters can do. There's also been a decline in attendance at figure skating as people are not as interested in watching it.
Daiane: Interesting! Very interesting. So when robots perform at the Olympics, the stadium will be just empty. Who will want to see robots?
BD: Other robots.
Daiane: (laughing) That's hilarious.
BD: Daiane, do you use your education in psychology in collaborating with Weidong's artificial intelligence?
Daiane: I have a BA in psychology, so I'm not an expert, but I feel like I learn much more about psychology in our labs and working with people than when I was at school. I have no doubts about that; the group dynamics, how do you get people to do what you want them to do. It's always different for each person. So, in a way, yes. It's fascinating to see when you ask people to imagine things, to create stories on a certain subject, the things that come up. With dance, everything's so buried, so deep, that you can't hide it, it just comes up. When you write something, you can always try to hide something about how you're feeling, but in dance you cannot. You go out there, you do an improvisation, and things just come up.
BD: Some of your performances involve audience participation. How does that work?
Daiane: It depends.
Weidong: It depends on the setting. Is it a formal theater, and what kind of performance it is.
Daiane: Yeah. For example, at CounterPulse Festival we will have audience participation. Because we are now doing a formal performance we'll be showing what we've been experimenting with in the lab. So we have created multiple scores on the theme of artificial intelligence and we show the scores but we also ask the audience to participate.
Weidong: We designed a performance like a Markov Chain, with the audience actually running around the stage.
Daiane: That was a really fun one. We created tasks, very simple tasks that every human being can do. And it just creates these very interesting patterns in space and people having fun, and it's interesting.
BD: That was the one where people worked in pairs, with one reading out instructions on their cell phone and the other acting out the instructions? And a screen behind them with occasional commands, like Run? It's hilarious.
Daiane: (laughing) Yes. Run! Walk! Levitate! Hide! Drop dead! Hug! Talk to someone. Introduce yourself.
Weidong: I love to reverse the roles, with the audience running around and the performers standing around watching. (laughing) Really cool!
Daiane: That was a good one. People had a good time. And it was a weird way of using technology to help people connect with each other.
BD: Is that part of what you're hoping to do with Super Lab?
Weidong: The lab is where we explore ideas and ask questions. So the Super Lab is super-sized, where we invite people to participate in our exploratory process. So there will be a set exploration where we dancers will have an informal performance, and in some of it we will invite anyone in the audience who wants to participate. If they don't want to, they can just observe.
BD: A few years ago Diablo Ballet had people in the audience Tweet suggestions on creating choreography.
Weidong: It begs the question: Where is performance going?
Daiane: Yes, exactly. What is the role of the audience?
Weidong: In tribal times, there was no division between performer and observer. It was a collective experience. Formal theater came about in Greek and Roman times. Then ballet and opera created a formal space. The audience is far from the perfomers and they have to be quiet. Originally, opera was not like that, right?
Daiane: People walking up and down, talking, eating. Still there was a separation of space between the audience and performers. But it was not like the audience could not make decisions, like I want to go grab a bottle of beer or I want to go out for a little bit and come back.
Weidong: I think movies fundamentally challenged this tradition of the separation of audience and performer, of putting the performance on a pedestal. In a movie when you zoom in, the blink of an eye can give you so much, at the right moment, can evoke such a strong emotional response. On a stage the dancer has to work so hard in a big theater to register the majesty, the beauty of movement. An even deeper problem is more psychological. In a movie people get used to seeing things as prerecorded. Even though people get killed, you know it's a movie.
Daiane: You still cry.
Weidong: You still cry. You sit down in the theater, the psychological experience brings you right back into the movie theater. You start seeing the performers as human beings, and the performer, of course, is not hard to relate to, either, because of distance. So you start seeing the performance as a movie being played back. That's why in ballet when somebody is injured and cannot dance, suddenly the whole theater becomes alive. Because that's the moment where, "Oh! It's a human." So this is a question we are trying to explore: what is the role of performance with the audience in the age of big screen movies.
Daiane: It depends on the kind of work a company is doing, but a lot of people have been experimenting for a long time already, especially in Europe. Many years ago when I was living in Europe, in Brussels, often we became part of the performance, as audience members.
Weidong: In Chinese opera the traditional way of appreciation was very different from the modern way, like the Western opera setting. Traditionally, when performers were singing people were chatting , drinking tea, singing along. It was a very rich, visceral experience. Nowadays people sit down and listen. In the past they would chat and go in and out and walk around and there was noise. It was not sacred.
BD: Like church.
Daiane: Yeah. Like church.
Weidong: To me, that's more like a genuine human experience. And everybody can sing along. It's a much richer experience. Today if you sing along with opera, people are like, shut up! (laughing)
BD: What would you like audiences to experience or take away from Super Lab?
Weidong: If we can get them to start thinking, start noticing things, then we'll be successful.
Daiane: Maybe they'll find the right questions. We're still looking for them. We want to create an experience for people to think. We're curious to know what they're thinking about and what they have to share.
BD: How will you ask them?
Daiane: We create questions; people can answer or not.
Weidong: Also, because our lives are ongoing, people go and they come back. So conversations do not happen at just one spot. It's more of a continuous conversation. When they come back to the conversation, they bring what they experienced before and take what they experienced outside of the lab.
BD: I think I'd have to see it.
Daiane: Every Wednesday night the lab at CounterPulse is open to the community. Everyone is welcome to participate and share their own projects with us. We get together and we play, we talk.
BD: Sounds like fun.
Weidong: I think we can describe a couple of scores we developed based on this AI concept. One of them is from super surveillance. You know in China they have suveillance cameras all over the country. Now they put AI with it so they can identify people, and they're networked so they know where everyone is. So what's the implication of this surveillanced country? The obvious thing is the super power imbalance. Based on this concept, we created a movement score for dancers to explore-how is it to be watched and to watch.
Daiane: We have three states: watching, invading, and hiding. So we can make choices during our improvisation, what kind of action do we want to do. We never know. We have to be totally connected to see who is doing what.
Weidong: The dancers get to choose their state.
Daiane: This is not my favorite. My favorite is the Deep Dream, Google Deep Dream.
Weidong: Deep Dream is like where people start to see things from nothingness.
Daiane: Say I give the computer a picture of the Mona Lisa, and it will start seeing different patterns.
Weidong: This is an example of artificial intelligence that sees things that totally do not exist. It's an interesting transfer of like style.
Daiane: It creates really kind of psychedelic images. We're translating that idea into movement. For example, a group of people can go into a space and create a landscape with their bodies. Anything they want to. Standing, sitting, or moving. Maybe they're together, maybe they're separate. It's doesn't matter. They're just creating a picture. One person sits facing the performers and closes their eyes and opens their eyes in a split second. Whatever they see, they have to translate that image into movement. You don't really see it. You use your imagination to create your own patterns with movement.
Weidong: It's a fundamental AI problem. AI learns from observation and observations are usually never complete. AI is very prone to creating bias by observing, like Google AI is starting to analyze the difference between humans and chimpanzees, and the next thing you know they start classifying African Americans as chimpanzees. So it's very much based on everything they observe they start interpreting and labelling. We're reproducing that through dance exploration, having the dancers purposely only taking a slice of information and from there explore into a totally weird landscape.
Daiane: It's very dreamlike.
Weidong: So that's an example of how we take AI into action with humans in life or in the future into movement scores and movement exploration.
Daiane: We also have electronic music going on, a musician doing the same thing as he watches us, playing with sounds, creating soundscapes. So it's like conversation between dancers and musician.
BD: It's like watching movements in strobe light flashes. You only see bits of the movements.
Daiane: Exactly! You got it. (laughing)