Introduction by Professor Maurice Pagnucco, Deputy Dean (Education), UNSW Engineering
All right, ladies and gentlemen, welcome and thank you for joining us for this lunch and learn, our bite size research talks from leading researchers at UNSW. We've got a fantastic talk lined up for you today. My name is Maurice Pagnucco, I'm the Deputy Dean for Education at UNSW Engineering. I have the pleasure of introducing my wonderful colleague here in a moment.
Before we get started, I'd just like to acknowledge the Gadigal people of the Eora nation, the traditional custodians of the land upon which this building sits. I'd like to pay my respects to the elders past, present and future. In the last 70 years, since University was founded, we've continued to refine and grow our commitment to providing lifelong learning opportunities to University alumni and friends. We're delighted that you can all join us here today.
I'd like to also thank those in attendance from our Scientia Circle programme, our supporters who enable UNSW to advance knowledge through research and education by giving in they will. Thank you very much. A few housekeeping notes before I introduce our speaker today. Please be aware that we're recording today's session, and it will be available for podcast via the UNSW alumni website. Also, please have your phones on silent during the talk.
Now, as I said, I'd like to introduce my wonderful colleague, Professor Mari Velonaki. She's a Professor of Social Robotics at UNSW Faculty of Art and Design. She's the Founder and Director of the Creative Robotics Lab and the Founder and Director of the National Facility for Human Robot Interaction at Paddington Campus. She has made significant contributions in the area of social robotics, Meteor app, and human machine interface design. I've had the pleasure of collaborating with Mary on a number of projects over many years, and her work is truly outstanding.
Mari holds adjunct positions at the Australian Centre of Field Robotics at the University of Sydney, and it was set at a University in Japan. Mari will now give a presentation for around 40 minutes followed by 10 minutes of Q & A. Please join me in welcoming Mari.
Learn@Lunch presentation with Professor Mari Velonaki:
Thank you so much and welcome. It's good to see everyone here. You can hear me with this microphone. Yes? It's the last presentation of the year and may be for you it's the last talk that you attend this year. This talk was scheduled the beginning of the year, but that was very sick and I lost my voice. For someone as talkative as me that was a disaster not to be able to talk. I mean everyone was, they have been our household. Now it's the end of the year, it's quite of symbolical and us have been giving for an unknown to me reason, many presentations over the last three weeks and I feel a little bit exhausted.
This is a very special time for me because I feel this can be a little bit more intimate, and not my standards. A little bit more personal because I feel when I was putting my slides together, let's talk about some things from the past and some things that really matter. Instead of my usual more academic, it's going to be a bit of that too. It was about designing social robots, and Maurice my wonderful friend and collaborator of many years, Maurice Pagnucco, talk about learning. I think when we start talking about social robotics, no one is an expert and we can learn together and we can all contribute. No one is an expert, because first of all, social robotics is a new field. It's a multidisciplinary field by definition, because we have the world social, society, right?
We'll talk about systems the everyday life, people, machines, robots and systems that they operate outside of the safety of the lab or the safety of expertise personnel, and hopefully that can enhance our experience if we need them, if we interact with them. It's multidisciplinary, it requires expertise or interest research interest for people from psychology and the social science to obviously art and design for interaction design, behavioural design, aesthetics, AI mechatronics, ethics very important and most importantly, the general public. Society needs to engage in these conversations because we talk about systems that can be implemented two very, very different social scenarios.
I will start with cultural, it could be cultural artefacts, it could be assistive or it could be responsive, interactive tools for learning, but we need to include people in society in these conversations because sometimes, not all the times, technology and technological developments are good for us. I'm talking about all these things and here you have a small humanoid track, with my one and only humanoid track. That's Maurice. That's me. That's me again. Great. This is me and an Android robot. Again, this is Hiroshi Ishuguro's robot, I've worked with this robot in 2006, as designing a breeding project, and this is the Gemini that was the first time in the facility ran a cross cultural experiment between people in Tokyo and people in Sydney. The research question was how people approach and accept a social and Android robot.
An Android robot it's a robot that tries to imitate humans, it looks very human like, in Australia and Japan. Do you know why it's in my first slide? It's not my first slide because over working to Android robot, I believe that Android science is the way to go. I have a lot of respect for Hiroshi Ishuguro and his research from a research point of view. As a matter of fact, I believe the very opposite that they Robots I would like to create they shouldn't be Android like, but I have it as a first slide to remind me how wrong I am and how can learn from it. Because for years and in many presentations like today, I used to say for example, an Android robot will never be acceptable in Australia, people in Australia they will never, never, never like even to engage they will be completely paused for something like that. That's been recorded and usually I'm quite passionate, you work it out by now but my ideas and I know this is not possible.
Then we have these experiments with 115 people in Sydney, 115 people in Tokyo the results. The Japanese people trust the robot more, but the Sydneysiders wanted to interact with the robot more and they like it more. Contrary to my belief of this will never be in Sydney and it was. Anyway, I was wrong. That's why it's there. How it all started, the deal started from cultural robotics. A little bit of my background, Maurice said I'm a Professor of Social Robotics. Okay, my background I moved in the Auspicious 1994 from Switzerland to Australia. I'm half Greek, half is French to start the interactive media because it was very happening in Australia at the time, artificial ages, media. My first degree and my background was a by-product of an art culture, art and technology culture in the early days. Of course, if you read different articles it's about these exotic thing about art and science working together and an artist makes robots but the truth is I was always an artist was putting boards together electronic system and for years I've learned how to set things on fire.
Finally, I moved to robotics lab that other people taught me that it's okay sometimes to set things on fire because it's a working prototype or a prototype that you aim to make it work. It all started from media art and interaction art and something in social robotics that it hasn't been recorded as a big contribution as it should be. As social robotics grew, I think it needs to be. It's how interactive art and especially Australian interactive art has contributed to robotics. Because Australia has a very long history about technology, people in spaces, and how people respond to different technological artefacts. I would like to pay my respects, and that's how it started to move to developing different sensory interfaces. My PhD was in New Interfaces and Human Machine Interaction, and in 2003 designed my first robot but when I moved to the Australian Centre Field Robotics that I'm still adjunct.
The robots were working at the lab at the moment started from cultural robotics and this is the Fish-Bird robot. They appear like wheelchairs but they're robots. That's my first robotic work in 2003. It was a project supported by the Australian Research Council. That was a big, big opportunity, because that's... and that says something about the Australian Research Culture. Because someone relatively young than you would get the support to build a robot with robot since I decided the interface design the original concept. You will ask me, you see two wheelchairs that they don't look like robots but Fish-Bird robots. The research question about Fish-Bird, because you'll tell me it an artwork, why it's research. You have to remember it's 2003. In 2003, we had two schools of robotics cute robot, pet like, or humanoid robots. We wanted to talk about engagement with robots, but also how to enhance engagement and the question was appearance versus behaviour.
Choosing wheelchairs you start from a very, very difficult object. Because as we all know, a wheelchair is a social sensitive object, but also it's very strict. It's obviously a device, it's an assistive device assist people to perform tasks, but for me I had many positive attributes to be a good robotic project. One is wheelchair was I think, if I'm not mistaken, invented by the Assyrians as an object to support and help the human body. It was a historically the first object designed to assist the human body. Therefore, this kinesis enabling movement and its relationship to the human body was important to me. Secondly, was an assistive device that helps to tell a story. This is a fairy tale, my favourite fairy tale, contemporary, it's Fish-Bird for the ones that don't know. They're in love, but they cannot be together due to technical difficulties, right?
In their share isolation, Fish-Bird use this small thermal printers and the right this messages, they have different handwriting's to each other, but also to the visiting people in the audience who disrupt this perfect relationship that sometimes is not perfect. They also fight. We have 10,000 scenarios. They get jealous, they have behaviours. This work was developed in 2003, and the reason I'm showing it to you is because 16 years later, it still runs around the world each year, just came back from Spain. It's been accepted in 13 countries and we develop the biggest data connection between human and robots. We have over 600,000 recorded interactions. What we're recording in this museums and why it's relevant to robotics. It's how much time people spend with Fish-Bird, one of the conditions is you have two visitors at any given time in the gallery space, on the museum space. The body language of the people who visit.
We do skeletal analysis of how the body language change, how close they're to the robots. What messages? What's the correlation between the time they spent and the nature of the communication they receive between the two robots? Fish-Bird to date, for us, for my group, it's one of the most important projects because it influence every other robots that we design, and from culture to assistive care to defence. There are some principles that we've learned through an artwork installed in different museums. Also, I must say what's very attractive about museums is you get a very different diverse population and you can have comparative studies about how people respond, let's say London versus China versus Australia versus US versus Japan. I'll show you a brief video. I have to double click they told me, you see, I didn't design. I don't know all these things with interface. Every time I give a presentation, I make all these mistakes.
That was the first time that was presented as part of us electronica in Australia. The biggest challenge for us with Fish-Bird was to make them not look like robots. To string that technology, to make silent models that they can be... they're embedded within this. They're all custom made, the only thing that we bought was actually the tyres. To hide all the small computational processing within the seats. Then from 2003 until now, the biggest problem we have in robotics is the computational power. Now we're using the cloud but back then, what we did is they communicate with each other wirelessly, but also they communicated to a much bigger network that does all the big processing. We use vision, the space and different lasers and greater technology to get accurate data, that they go to the robots much faster because if you had all the computational power it would be really, really heavy.
The other thing that was introduced was poetic language, although, I'm not trashing. I love Anna Akhmatova's poetry, there was a lot of texts from Akhmatova. There were text composed by myself, but also for three years people donated love letters to the Fish-Bird robot from around the world. They were sending different messages. They also have 15% ability to write their own text. They have different types of texts that they compose, anything with that it keeps growing as their behaviour. Every time they have travel to different countries, we give them the geographic locations, where they are, and we connect them to the local newspaper, so they print something in the local language to be connected. Also, they have all these Google Maps about where's the nearest park or forest or sea and if they trust you enough they send you messages of how to set them free. In one location I'm not going to say where someone was arrested because they were trying to set Fish free and take it out of the museum. It doesn't happen very often but it did happen once.
Yes, I'll move to the next one. After Fish-Bird, I was asked because I was on the humanoid robots, especially after Android robots to design a robot that is a humanoid robot. Again, that one wouldn't have that it's only application would be to be museums and just simply test how this new morphology, this new aesthetic proposition, how people would accept it. We went from one extreme to another from a wheelchair to something like Diamandini. Diamandini that's the name, it was inspired by a Southern European bust that was developed straight after the Second World War. When I look at it like it was heavily inspired by this, it was a phase that to me it was interesting because it could be someone could be 18 years old or 30 or 35. It had the look of someone who sees a lot.
I did want to go for something huge. This is the Diamandini, she's 155 centimetres tall. She's that high. Unrealistically slender, she's very, very slim. She doesn't look like a real human, she looks like a sculpture. We made the finish to look homage to sculpture, have this porcelain finished, that actually can crack and can age like a sculpture. This is at Victoria and Albert Museum in London and it's wonderful because it's in the... I'm sure you're familiar in the Mediaeval and Renaissance Gallery. It looked like part of the sculptures there only when she was moving obviously she wasn't. What I didn't tell you, she's very heavy. She's 120 kilos because of batteries and motors. That simply means if you're not careful she can kill someone because she can move very fast.
When I started working with US Air Force, and there were people talking to me about how to improve performance with Maurice. We're operating in real time because people can kill. I know exactly what you mean because in my installations people can kill. You don't have special personnel you have a robot that it has to operate as an autonomous system, seven days a week, more than 10 hours because big museums are open seven days a week. In remote locations, people pay you to instal this work, they hire the work for six months or for three months. Once you're there and you leave, it has to operate. This is what we have in common. That was from the first time in 2012 and also the first time because Diamandini... Yes, it's me talking you don't want me... You're here to listen to me but not twice. You can see it.
Why am I showing Diamandini? What we've learned from Diamandini that, she's a sculpture robot that people look. What we found the Diamandini also exhibit here in different countries, it's something that for the first time in robotics, we discovered that the cultural difference from country to country are not that significant, especially in young age groups 18 to 35, they're very similar. England was very similar to China or Australia or Japan, but what was very different was the socio economic background of the people visiting, which is something that... when people talk about cultural difference, the biggest gap was in social economical difference not in cultural difference between people and how they behave.
For us, we work with Victoria and Albert in not only to exhibit the work but to collect data and run a workshop at the Sackler Centre. We've been there for a couple of months, just working. We installed first of all Diamandini in the super prestigious gallery that we thought if you break anything, you'll never get to work in this town again, the Mediaeval and Renaissance and then in the Sackler Centre. Within the same museum and under the same roof, we get extremely different responses. Because the Sackler Centre, it's a hands on centre that you have the young kids and people from nursing homes, and they know that it's okay to touch and explore. They were dancing with her, they were pushing her, some were pushing too hard. You put something into the public you can edit. While here, all of a sudden the value, the cultural value of the robot was very different to be here must be good. People not only wouldn't touch but they were just taking photos and just stand next to her.
The situational context is extremely important when you design something that's why I think it's very important for us to consider for other robotic applications. For not only is it a museum or a nursing home or a hospital, which hospital? Which nursing home? In which area? In which country? Is it Alzheimer's or just people that they're old and they need a bit more support. It was very, very intriguing. Then we thought, "Okay, robots are interesting in the lab, but maybe not everything needs to be robotic. Maybe we can do small devices that they connect to the robots." We work on this small project, they are fragile balances, that it's small, autonomous cubes that you can pick up and they enact. It doesn't matter where Fish-Bird are in the world, the message is by holding and picking them up very gently. If we get the messages around the boxes and we enable the communication. They act as advertisers, as agents.
It's the anti interactive work because when you hold it you have to be very careful in order to... it has accelerometers and different sensors too. You have to be very careful when you turn them and you'll see in order to be able to read the writing. If you do it very fast or if it's if some object movement, it's eligible and eventually will. Also, Fish-Bird we developed different handwriting's for them. In comparison, between Diamandini and Fish-Bird, people are still interested in the Fish-Bird because of behaviour, not aesthetical decisions versus Diamandini. Diamandini they just want to take photos also she's responsive to touch, we wanted to see where people are comfortable to touch human or robot and it's usually the hands and the head. You pick it up that was the charging station and you'll see all the messages. You see it coming.
Why I'm showing this year shorter, it's a hopefully interesting interactive work is we had an aha moment in the early 2000 with a sensei. Can it be an assistive device? Can we learn this was something else, something like that outside the space? You see, it's moving and you can... We discovered that it's a great tool to engage people that they rehabilitate and that's how we started working with rehabilitation. People that had an accidental or they want to regain stability. Instead of giving them exercise, you give them a story and a narrative and sense making. Then we also instal this work in a couple of nursing homes that people work on stability.
This is the journey for us in the lab that started saying its creative robotics labs because we've learned a lot from the cultural space, we want to transfer these and moving to other areas like education, age, care, hospitals. Look, culture doesn't need justification, we don't need to justify why something is important in a cultural sense. We need to justify why a technology that is used by people in an age care facility or with people with dementia should not be designed with creative principles, why creativity needs to be the privilege of an art context. This is an area that I think is extremely important for us.
In the lab, we do the socialist assistive work. This is a project that it's a mnemonic device that is currently on trial in different nursing homes in Sydney. It's a watch that just gives you notification, it has a photo when your loved one will visit again. As we speak, we also develop with a component in three dimensional object that it looks like 3D screen, to have the same image of a person and you can send notification. I send notifications to my mom, she's a nursing home, she suffers from dementia, about when I'm going to visit again. It's a very simple project and it doesn't require robotics, right? It's just a device, but it was hard. We work with neuroscientists and people from psychology to discover how simple the notification should be with someone who suffers from dementia.
When we developed the watch, it was okay for the gentleman in the nursing homes but the ladies, although it was a cold design process. I was talking to them including my mom, they said that they were interested to experiment, but then when they put it on default, it was too bulky, too big, too ugly, not like jewellery and they refused to wear it. Yes. The gentlemen were very sport and they still like it, but people like my mom was a worse, she says, "No way, I'm not wearing this." Again, with a nursing home personal we created this beautiful object that it looks like holographic but it's not moving because people get confused when things move, it's still one image. We put it on the dresser or side table in different areas.
This is a project by Dr. Scott Brown, who's in our lab. He works with autism, autism and play, and how to create environments that can engage children and their families. It's not about only learning and improving but how you engage to society with all these. This is interactive tent, it's a tent that you can work in and it could be two children together or a therapist and the child to play and interact in 3D environment versus a digital environment. Up to that point, most of the things with learning and autism was screen based. One of the biggest problem is to engage with super intelligent children with special abilities into a very different world not so much and to create another screen another barrier. In a physical world, even if you use technology that you can play and you can engage with families with other children and with a therapist and be installed in their schools.
Now we have two robots, two other robots that we use in the lab. This is Paro that you know, it was developed in Japan. It's a seal robot that makes interesting sounds and response. It's been used in many nursing homes around the world and we were invited to do an evaluation study for it. To me Paro, it's an interesting starting point it's under the category robots for psychological enrichment, but we can do so much more. Instead of having a toy, it's a great platform. I don't criticise it because it's easy to criticise how to improve something and difficult to build. It has many merits, to me first of all is like this creature that it doesn't look realistic it's not a dog or a cat it's a fairy tale.
No one will think it's a real one especially people that some of their senses are not quite accurate. They cannot see very well or they're in a more advanced dementia state. It doesn't have any senses that they make any measurements about the person, we cannot learn the condition about the person interacting with Paro and would it make any difference if it was another fluffy toy? Or is this specific aspects of Paro that hopefully reduce anxiety and stress? I think it's a great platform but we'd like to do more working with Paro.
Kaspar, what do you think about Kaspar? When I show you this photo what do you think? Do you like it? You have to be honest, come on. Do you think it's attractive? No. No. Kaspar was developed in Hertfordshire by Professor Kerstin, one of our colleagues. It's the most successful robot in autism, she has the biggest data collection in the world. We have in HRI she has with people with autism. We work with her and David Silver who was my boss Doc and PhD student now he's in Cicero still adjunct in the lab. Dr. Brown they work together into the next generation of Kaspar. I always put Kaspar in there because it's a conversation starter when it comes to social robots. I'll show you Fish-Bird and Diamandini and she's made out of porcelain.
Why Kaspar? Kaspar looks to me like a horror movie, right? It's a bit that mask. I mean, my daughter... I have a teenage daughter who watches this horrible things with horrible dolls. Remember what I was saying about situational context, this is for this autistic kids that face was chosen by the children, right? They're much more comfortable with faces that are not super realistic and it doesn't have movement. It's a good example to say my personal preference, my personal aesthetic preference needs to be parked somewhere. Because it doesn't matter what I like, it matters what this robot is designed for who. Who's going to use it? First, the element of core design, right? That we need to really, really spend time about who are our users? Why we're doing this work?
You can see more things about psychological enrichment, Kaspar, Paro. Also, Paro is sometimes used by children in hospitals. I'm here, how do we aim to create technology better? That's our vision. It's to enhance and improve the human experience when interact with technological systems and to test is it necessary to have new technological systems? I'll tell you what, there are many things that there are some morning that are, "Every time I work with industry it's so much easier than working with the university." There're times that I say that, right? It's great, I love working with industry. There are other times that I say, "You know what, what I develop this time is not the product." Even if it's a disaster, I can still write the paper, it's great I mean the university and say, "That doesn't work, we had bad result, right?" I can still get brownie points by saying, "We shouldn't be doing this because people are not comfortable. It's too noisy or it's too realistic or the kids actually get bored after the first week." That's important.
This is what I was telling you about, how do we create social robotics discipline? I think we covered that at the beginning. We need everyone plus society, right? Ethics, art and design, artificial intelligent, mechatronics, social science. We're talking about the future now, where we want to go. Experiential approach robot design, it has to be experiential, it needs to be tested. We are at the moment in history when all this anecdotal project I show you are just the beginning, because it's very different to design things to view in a lab or in a gallery, and it's different to start designing for long term interaction. Long term interaction has very different principles to something that you can be impressed, entertained or use something for rehabilitation for few weeks or a month, and then it's the end of it.
How do we design? How is that some framework that we design robots for long term interaction? That's really difficult. We need to consider obviously, the robot morphology how this robot looks, right? Because, it depends on the use case, we have to use different design principles. The behaviour needs to be complex, because otherwise we will get bored very easily. It's going to be like your old model of your phone that you want the new model and we cannot afford to waste taxpayer's money, resources to create systems that become gadget, especially in social robotics. If it's entertainment, it's a different thing.
Real time responsiveness, it has to work and it has to be responsive and you can't guesswork. The communication and the responsiveness is extremely important. You don't respect something that doesn't work. It's like training Siri at the beginning, "Siri. Hey, Siri. Hey, Siri." Then it's like, "It doesn't work for me, it works for everyone else." Actually, it even response to me at this time. Learning and adaptation, if a system doesn't learn it stop being useful, right? Adopting, it needs to be able to read the human context and the situational context of the environment. I think these are important principles and these are big research questions that we need everyone to work towards. Every project you feel... I mean, what's great about the field? Every new project, you feel like a newbie, you feel like you know nothing, because you have to consider all of these things.
What needs to be reconsidered? Obviously they look, I don't think we've nailed it I think we have a lot to learn how a robot should look. How it should move? How robot moves, it's still one of the biggest gaps in the field. How should a robot communicate? Representational versus non representational humanoid or Android versus zone morphic or something completely technological. I think there're some cases that it's very important that the robot just looks like a robot and has a fantastic technological aesthetic, but it doesn't. Textures and materials, what we come in contact with and size. Because if you have something in a nursing home, you don't want it to be big. We want something small or something in a hospital. Size, materials and presentation or not is important.
The other thing about modes of communication that was testing the geminoid in Sydney in Japan, the effective interaction explicit versus implicit. This is a big area because most of the work up to date is in we do some work with that, but it's voice natural speech, text. I think we need to consider a very different things like effective touch, versus touch. We do a lot of working for the robot understand macro and micro expressions about the human state. We do work with emotional speech analysis versus speech recognition. What is more important, it's what the tonality of my voice communicates versus because it's very different when she say, "Come here." Versus, "Come here." Two different ways, there's a lot of things we need to consider. Sound is completely underdeveloped, how Sonic communication, even things that they're noises. It's still primary school level, including our own robots.
I mean, these are things that we need, how Sonic movement could be a mode of communication, how we can create some visual excitement and entertainment in circumstances that are not necessarily designed for that. How do you make people happy if they need to use something? If they have no choice? It's the new level of the robotic wheelchair that is coming. How things like that can implement it in autonomous cars. Elon Musk, already with all the Tesla's in few years from now, the new project is to be able to hire like keepers and be autonomous. How does someone like my aunt who 75 years old trust to be what's the interface to enter in that car and trust that she's going to be comfortable to go safely home or wherever she goes? All these things need to be considered. The quality of movement, it's something that I've mentioned before, how robots moves and stops.
What we have in robotics of the moment it's start and stop and perform and stop. We don't want to create lifelike things, but presence is important. With Fish-Bird for example, even when they need to preserve battery life and they're there in the corner, they need to move a little bit or if someone walks into the room and they need to turn and look or facing a person. There's more things that we need to implement in all robot design other than the on off button, other than its own and it's off if you want to have presence and engagement. This all it's completely, completely because do you know why all these years or A to B because it was simply navigation button, it was optimal reaching your goal. We used to design robots especially field robotics when I was in field robotics, to go from point A to point B, but sometimes they may be the more efficient way. The fastest way to reach somewhere it's not the best way when you interact with people, when you interact with children or fail people when you need to avoid things, when you need to communicate that you are safe and aware that something is happening there.
These are some of the things that we started to explore with the Fuji Xerox in Japan, we had the three year innovation contract with Yokohama their centre in Japan. We started coming up with propositions, how do we do a robot for them that can actually be co-owned for different members of the company? This is in a prototype stage and we wanted something that it could be humanoid, but not a humanoid. It's very clear that is not something human. We wanted to develop this body language that the robot not only develop a soft body language, but with different turning parts, but start understanding the body language and the habitual modes of the employees that it needs to share space with. Hopefully, next time I can show you more things on that.
To sum up how much time to have someone was telling me? Four? Perfect. Towards engaging human robot interaction, human accepting of technological systems influenced by the morphology and behaviour, trust and engagement, system competency, they really need to work. Intriguing into a system learning adaptation and evolvement. Intriguing interest, when I was working Fuji Xerox with the employees, you can imagine it's almost like a paradox. You have an exotic researcher from Australia with an exotic accent with other people from the lab, going to Japan to design a system for a Japanese company and trying to persuade the workers, the researchers there. It's an innovation R & D Centre to work with me. It was a very hard exercise. Everyone needed to be involved, right? I don't think I'm capable of doing that.
It was a co design process and it was from our questionnaires that the whole company was engaged, to the team of people that they're going to engage with a robot to be with us for three years, regular meetings, discussions. Some of the things that they were very clear about, about intrinsic interest. "I don't want something that is predictable all the time. Predictable is not always good." We're talking about something in the workspace. Can you have a physical form in the workspace and actually what the robot will do, which I didn't tell you, it's to read the situation when people are in very stressful conditions, make sure that they take enough breaks.
It's an oxymoron that we need to design, a sculpture looking robot that people could design it to be persuaded that they need to take more breaks and it's okay because when they used to receive text messages on their screen it was, "Someone is monitoring this. Do I look too tired?" There's a social stigma and I know that we work with companies in the Bay Area in the States, not only Japan and Australia I'm sure, about performance and us now getting tired. How, "I'm so fit. I'm so healthy. I can do 10 hours." No, I can't. Social stigma, if I tell someone my lab, like, "Why don't you have a break?" "She's not going to renew my contract." "No, no, no I don't need a break. I'm fine."
How do you do a system? How do you develop I cannot reveal the personalised language, individualised language, that what that robot communicates to you is just for you? It's not for Maurice, it's not for the lovely gentleman over there? It's not the same thing because then you again, create a social stigma, right? You create the language that the robot is going to see Maurice is not doing well today. We have different activities that actually they're fun activities that they break the day and it's something that people agree. What is fun for them, not for me again. In the behaviour they wanted to have an element of surprise, a robot that will do something a little bit naughty. Not dangerous, but something I don't predict. An element of surprise, because we're so complex. We don't want the same thing, right?
Understanding human behaviour, how can we ever? How hard is that? How do we understand? It's one of the hardest thing. How we've extensive testing and experimental evaluation of new interactive technologies. We do that with open experimentation in museums, as I said around the world, but I can't lie to you and I know you're not going to believe me, not everything is designed to be museum exhibit material. It simply impossible to start testing everything, some projects are great, but not everything. Lab testing, we have different issues with lab testing, and Maurice you remember that, I don't know a few testings in labs every time you have a different experiment, you start adjusting the cameras and no one walks in and you have the light on.
One of our dreams was to create the state-of-the-art facility that we can get really, really, really complex data sets about human effect and intent and we can always provide technology evaluation new system. We created the national facility for human robot interaction research. Its national, but the main partners that they were part of the ARC Grant, it's the University of New South Wales, the School of Art and Design, the School of Computer Science and Engineering, School of Psychology, I'm very happy for it. Then from the old family of Australia, University of Sydney, the Australian Centre for Field Robotics, The School of Information Technologies, the University of Technology, the Centre for Autonomous Systems, and of course from St. Vincent's, private the Department of Rehabilitation Medicine, which is really, really important.
This is the space and we call it the Tiger Room, because it's a crazy yellow and has the stripes but what it's hidden in there we have over 200 discrete sensors. The space is long enough to be able to have robot running on a big system and it is the only one in the world. We get a lot of external contracts, not only for robots but for system evaluation like a ground truth about how to violate the system, like maybe water different wearable designers or smaller projects. It's a wheelchair accessible, also has a functioning kitchen that is wheelchair accessible and people from St. Vincent use it. Post trauma, post rehabilitation to monitor how a person prepares a meal before he goes back home. We have a kitchen that it's linked, it's on this item. I'm sorry, we didn't have a photograph of that.
It's something I'm very proud of and we involve more at the moment like the projects that we've never believed that we're going to engage other than robots. We have a project of the Brain Institute for early predictions of science of Alzheimer's. We have the project of defence, which is about measurement and creating an autonomous system for predicting cognitive load. To us, again is in the heart of improving the human condition, understanding more about people. Our last project that we are part of is the Ageing Futures, as I'm now officially middle age, and all the people in this group we start talking about how do we want age. The agent Futures Institute is a UNSW Institute. It started, we got the funding last year, six and a half million for five years. This is a dedicated space for the first time 50 experts in ageing from seven faculties of the university and we have also these partnerships with government, industry, academic, community collaborations. Of course, social robotics is part of it and I'm so happy that that can lead the programme on people, technology and society.
To me, it just gives us more strength and a better understanding of where we want to go. The main thing about ageing people forget that ageing needs a life course approach. From children at any given age and the principles of useful technology, engage in technology, do we need technology? To how to apply technology needs to be considered from many, many different experts and I hopefully we can provide a small contribution towards positive ageing. I have vested interest in that. On that note, please ask me any questions that comes to your mind. I was trying to squeeze as many things as possible. Thank you.
Audience Q&A
Maurice. P:
We have about 10 minutes for questions, if anyone has a question.
Speaker 1:
I just want to ask a simple question, in my involvement with artificial intelligence in the '70s we talked about robots and things like that, that would reduce the repetitiveness of human endeavours and things like that. I don't know, the last 30, 40 years has this concept change? What do you bring into this new world about your robotics dealing with Android and artificial intelligence? How would it enhance our lives in the future?
Mari Velonaki:
It's a great question. It's a very difficult question, to be honest with you is I don't know. Will it enhance our life in the future? With the Androids and humanoids, I don't think they will enhance our lives. I think we need to have that's why I think that for me if I would prioritise will be the smaller projects. Project that is not so much about repetition, but it's about system, it's about predictions and well being. Repetition is an issue but everyone... I know there's a lot of work about how do you cut down on labour and repetitive labour? There're some jobs for example, some robots I have, that most of our nurses have lower back pains. Most of our nurses and not only in Australia, in other countries too. They're people and it's a social service, we all know that they're not paid enough for what they do and for their services.
A big percentage have lower back pain because how they lift people or how they move people, you can have all the holes and all these things is not effective. I see them even in nursing homes, or when my mom was intensive care how hard it is to do all this work. For example, a robot that I like very much that is a prototype in Japan, it lifts the patient it could be for hospital or nursing home, but it doesn't replace the human. To me something that I think is very important with all these technologies is where's the human keep the human in the loop. Right? The human is there to do what humans do best whole Mr. Smith had to tell her before she goes to surgery or before she moves into the... "Everything is going to be okay." Because the nurse will have the time to hold the hand and keep the icon that because all the nurses and all the people I've seen they work so hard and their main constraint is how to move this body who cannot move there, how to use the horse, how to engage them. For example that I think the things like that, that would be fantastic to have a robot that it takes over a job that is not good for humans. Also, it doesn't remove the human from the picture.
To me, I think it's going to be more the cognitive load predictive things safety, how people can stay home, my dream how people can stay home longer. How you have an environment that it doesn't feel like the big brother setup in terms of monitoring, how do we create district centres and the robots just become smaller agents in there. It doesn't have to be that help their network to other things. Right? Your house could be a robot, with some agents as a reference point for entertainment, for exchange if you have. To me, I think the safety independence, security, the new generation of walking frames or body walking frames that they could help people to balance. I say why? For example, when you do that in at least in Japan, imagine instead of give you directions it plays that your favourite tunes to encourage you or little things, or the colours can change. I think it's not going to be so much like the '70s, and as a kid in the '70s what I was dreaming of a robot to be.
There was a wonderful document, "Whatever happened to my robot. They disappoint me, my robots can do nothing of the things I wanted them to do." I think it's going to be again, a bit of wisdom and not so much the sci fi Fish-Bird that I wanted, we wanted to have. Great, and it's good to have the Android as an experiment to say do you like this? No, hopefully not. I think would be very bad to have Androids around us trying to be human or just to typing things. To me, it would be about how they can assist in different environments social stigma, cognitive load, safety.
Speaker 2:
You talked a little bit about how the advances in computation have made robotics a bit easier. Thinking about general purpose robots, it's always the thing that said that humans consider the hard problems for humans are like playing games or talking and things like that. The things that we don't even think about the easy things walking and sitting and breathing, it's reversed for robots, it's much harder to do all of the work in the interacting with the environment and the higher function is almost easier to programme. In terms of getting towards that general purpose robot, the one that can help us around the house and that thing. What advances are we seeing in the interacting with the world and making that easier to do for robots?
Mari Velonaki:
That's an excellent question, because in the last weeks, there's this AI initiative in Australia there was the announcements of the Old Parliament House that Minister of innovation, the Go8 had the discussions about AI in Australia and be AI ready. I always remind people that we have two gears, we have AI in digital things that everything moves so fast. Then you have embodied AI, robotics, everything moves so slow, and very slow. It's still very, very slow. One of the reasons that we need to rethink the design of robots, we all know, I mean, preaching to the converted here for robot to move from here, to have legs to do arms, do all these things to do the human things. We don't need to do this, and hopefully we shouldn't be doing this other than as an experiment.
We have to look at forms, and first of all, why you need to have a robot why embodiment? We have to be very, very conscious about why choosing to have something embodied when you can read much faster. There's something about presence that it does, it does make a difference, but that presence shouldn't compete with a human presence. That's why I said including our work, some of the things. For example, they wheelchairs move much better than the Diamandini and it was much also cheaper and easier to build. Instead of having this on a directional platform. The robots like Atlas, the big robots that defence used in the US that they move, it was years and years and millions and billions of dollars to build, and still they fall.
I think we have to look, to reconsider, when I talk about designs, and I put these things up is not because I have the answers, sorry. Because they're really big things that nobody knows, but obviously something is not quite right. Okay, we have the entertainment section, the news and the sci fi but we know yes it's a promise but maybe we have to go back to the drawing board here. I would agree with you. We have the robots at home, we have the training place for the robots at home and we are about to sign up some a contract that cannot talk about the robot that kills with housework. I can reveal that it's not going to be humanoid and it's not going to walk around and sing for you. It's going to be very mechanical. That's so far. Thank you.
Mari Velonaki:
One about the German company they just started, I think the idea with this thing it has to be affordable in five years from now. The smaller devices, things for nursing home, the work that we do for defence or how to have interfaces, they should be very cheap and very affordable. The same thing, because we cannot afford ideologically to make these things expensive.
Speaker 3:
How many thousands? I'm just wondering.
Mari Velonaki:
Something for nursing home could be 7000 to 10,000. That's not expensive. Some of the devices that we build from mnemonic device, it's probably less than 1000 as a prototype. This is when the robot like Kaspar that the schools use I mean, we give it for free, we don't. It's a prototype, it doesn't cost more than 6000 to build. That's prototype level, it's not industrial. That's the other thing with social robotics, yes, the things for Diamandini are very expensive, because of the finish. It's a museum piece that goes around that we can charge people to for this, but I think when it goes to education or health, or sensors at home, or safety, it's unethical to be expensive. Thank you.
Speaker 4:
My question is a little bit different from so many others. I'm interested in the response that your work's head by artists. You mentioned obviously, you had a lot of exceptions from a lot of prestigious museums. I mean, I can definitely see the practical application and see the artistic value as well. I wonder, and you mentioned the fact that there's a heritage in Australia, of artists that work with that makes this between humanity and technology.
Mari Velonaki:
Absolutely.
Speaker 4:
I'm just interested to know what response you've had from artists.
Mari Velonaki:
Well, I guess I wouldn't say they positive, it is positive because the artists are part of it. To me, they don't even need to justify with a part of it. In my lab, the Creative Robotics Lab started six years ago, and I was in a different amazing robotics lab it's still my family, but it was field robotics. For 11 years, I was the only woman academic, and I never had a PhD female student. Now, I must say that 60% of my lab are women and there's so many artists, in my lab we have one of the most successful PhDs that we had in Social Robotics were from artist to be honest, and they're they're the ones that the companies pay because that's what they were missing the most. I must say like in the lab, we have people from sculpture, interaction design, pure design, working alongside with people with AI robotics, hardcore mechatronics.
You know why does it work together? Because artists, especially in this field, always, including myself when I was starting out, we've learned how to work collaboratively. Engineers also learn how to work collaboratively because putting a robot together you have systems, architecture, design, system engineer software, hardware, finishing design. What it makes it easy with this project, you need everyone to achieve this shared goal. It's like a puzzle, you need all these disciplines together. It's not, "I'm better than you." Because, well, if the platform doesn't move, then you don't have the robot. If you don't have an interesting concept and people are not interested in how this looks and behaves. It didn't need to have an acceptance because they were in the guts of the lab. I must say from industry every time because you have an exchange programme with industries, we started with Japan and they always want someone from design or interaction design because they have less of these people.
That's something that has changed. Maurice you remember this 2003 when I was started to talk about social robotics in 2003, 1995 I did my first interface I'm old, but 2003 I was super exotic, media art and robotics, I built a career of being exotic. It isn't it great that these artists don't need to be exotic? They don't need to justify why they have a space in that field that it's equal. I'm very about that. Thank you. Sorry, I'm a bit passionate. I said it's going to be more personal today.
Maurice. P:
There's two questions there.
Mari Velonaki:
That's a great audience actually.
Maurice. P:
I have three questions provided, Maurice can give short answers.
Mari Velonaki:
You need to monitor, that's not my stronger attribute here. Yes, sir.
Speaker 5:
Hi, my name is Calvin from IBM. Thank you for the series of lunch and learn.
Mari Velonaki:
Thank you.
Speaker 5:
To what extent do you do get involved in your research with industrial partners that the University have and what role do they play in helping you in your research?
Mari Velonaki:
Okay, the industrial partnerships with... we're a lab that we build things. We love to write papers, but for us, we need to develop something, test it and then write the paper. We can exist without the industry. Also, in terms of, to be brutally honest with you, my lab is postgraduate only, we don't have any undergraduate students. The facilities are testing space. We rely on industry and competitive grants, because we all know that research is soft money. Industry is essential, we wouldn't be running without industry.
We have to contracts, one is share IP, if it's like Fuji Xerox, because it's a new idea and they also contributed a lot in the research. We didn't do that ourselves and they do so many other things that we couldn't do. Simply, some other projects of the future fantastic and would like to get our hands on and create positions for us that they the IP, its contract research, should the IP stays with the company? I think is essential and for us industries it's beyond funding. We will be successful when all of these things have a space in society and industry can help us with that. When they become useful products, but also industry can say, "This is not useful." Other than what we validate in that case, is there a market for it? How it can be distributed. It's very important to us here.
Maurice. P:
I'm sorry. I'll take the last question from here. It has to be a quick answer Mari.
Mari Velonaki:
You see what they did to me. They invite me and then it's all this editing.
Speaker 6:
Thank you very much for a wonderful talk.
Mari Velonaki:
Thank you.
Speaker 6:
I've been very interested in the collaborative ethical approach that you've a nuanced as well. It that a big ask of corporations given how unethical they can sometimes behave?
Mari Velonaki:
It's not. It's not. Because, everyone has the ability and capability to be unethical. I think, we're still in a good position to, why you're here, why you want to work with us. I'll try to be brief Maurice, our secret is we're a niche lab, we say it is a niche. Now the niche is to sound a bit better, right? We're special interest because we're a small lab, we can't afford to do things that are not in the heart of our research. If I had 20,000 people, that wouldn't be a problem, we could do it. At this stage, it has something that it means something in our own research priorities. With the companies that we work with, it's a collaboration. Why you come to us and what can we do with you and why this is a project that it has to be priority because we get the 10% in different projects, 10% here 10% there, you have to be very careful. I think it's a very valid question and you have to be honest to yourself where you want to go with this.
Maurice. P:
All right.
Mari Velonaki:
Thank you.
Maurice. P:
First of all, thank you to all of you for joining us. It's our last learn at lunch session this year, but please join us again next year. Let's just thank Mari one more time.
Mari Velonaki:
Thank you.
