CMU School of Drama


Tuesday, October 01, 2019

“Alexa, dance”: Translating language into movement

The Tartan: While we are getting accustomed to having a black cylindrical device responding to our questions and playing music for us, our sci-fi dream remains unfulfilled: when can we have our own home robots that move and finish tasks according to our verbal commands?

4 comments:

Alexa Janoschka said...

My name is in the title so I had to post. I'm pretty sure there are cameras and a TV screen in the CS building that pinpoints your movement into a stick figure as you walk by. Ok... but why do they have to use my name for this technology? I can deal with the little amazon echos but any more and whenever I meet someone knew they start asking me the weather. And no, I will not play despacito. AI and robots to do humans bidding are honestly just strange. Can we not do things for ourselves anymore???????? I get that what they are doing is awsome and cool but sometimes I question why all of this advancement is necessary. Having a piece of technology that can understand your words is one thing, but now they are having robots understand human movement? Technology is confusing to me and the debate about where society should draw the line is a constant moral and ethical problem. I get that they are trying to develop this tech to interpret film (movement) into a script but once the technology is created there is always someone who will find a way to use it in a not so wholesome way.

Olav Carter said...

I think that articles and ideas like these are what lead the human race through a technological revolution, though ultimately will also overly corrupt our society. With any piece of technology, the use can be manipulated to do harm or wrong, and especially with a more technologically-savvy generation of children growing up today, hacking and technological manipulation becomes an incredibly dangerous drawback. For instance, if a robot can recognize when a person is going to the bank, or have access to a person’s bank account, then one who knows how to access that robot’s software also has that ability. With the expansion of technology that this article and many others describe and celebrate, I think society has to readdress and regulate what is necessary, and what will send us into a post-apocalyptic world like Wall-E, where robots reign superior. And if you run into a T-1000 on the street one day, then we take the necessary hint that we need to slow down and manage what technology is truly necessary versus harmful.

Emily Marshburn said...

As cool as artificial intelligence and robots seem on the surface, I honestly think that this “technical revolution” that first world countries are participating in. In my opinion it is not a renaissance but a revolution; and not a human revolution either. We are creating machines and software that is getting increasingly better at learning and, subsequently, acting based on what they have learned. Even if a revolution led by sentient robots is not as close as it can sometimes feel, humans are making it easier and easier to attack through the digital world. Additionally, we are allowing all sorts of pieces of technology to silently record us audiovisually when we sign user agreements that no one ever reads (which is low key by design). As we put more and more of ourselves into the aether world, we open up our information (whether willingly given or not) to any sort of access. Any computer science kid with enough coding knowledge could theoretically hack into pretty well anything they want.

Chase T said...

This is all very cool. Thinking about robots as performers is interesting in itself (there is a play about that very subject, but I can not remember the name). On a different level, this group seems to be making very significant strides in machine learning. It shows that we are not so far from having personal service robots that can adapt and learn tasks with ease. That said, unless I am misunderstanding, the technology is focused on mimicking human movement and associating it with words or commands. That is itself limited, unless there could be human volunteers that assist with teaching robots tasks that their owners cannot, similar to the idea where people volunteer via a phone app to assist sight-impaired people with grocery shopping. Besides that, I can imagine applications where the chain of movements a human would execute to accomplish a goal or nagivate an obstacle would be too complex to send as coded instructions. Instead, the robot could mimic an actual human in a training environment, and then execute a task in the real world that would be dangerous for a human.