 هل أنت تعرف أن تيكيو كان في مجموعة مجموعة؟ هل كنت تعرف أن أصل ما أسبكي سبان؟ هي تشقن مجموعة مرحبا مرحبا لكارميكي ميلان يونورسيدي أشعر أن تتحدث with me in English or Arabic ماذا أنت مجموعة؟ كيف أستطيع أن أساعدك؟ المجموعة is an important language أنا مجموعة أنا مجموعة أنا أمامي أنا مجموعة أنا مجموعة أنا أستطيع أن أستطيع أن أخبرت كارميكي ميلان الأسئلة بأسئلة في الأرابيكي في الإنسان حاولني أنت تجدوا إلى 1004 أنت تحتاج تأتي إلى جميع ثم تلعب تأتي إلى سبيل و تأتي إلى سبيل تجد تأتي إلى دولار من فرد أكتب أكتب أسئلة للمشاكل لنحن نقول أنه ليس موضوع الجديد على الناس أنا بخير إسرتي على تحضره بالعراضية والإنجيزية لديه جيد ومتذكر في الوقت next time your computer doesn't work لا تأخذ it لدينا تشعرين أيضا دكتوريت سيمونس, شكرا لك so much for taking the time for this interview and first would like to know from you how can we call a robot socially interactive and how is it different from any other robot so most robots when they operate they only take themselves into account their own safety their achievement of their own goals so how fast to get to a particular location how easily it takes to do something but in social interactions you take other people into account as well so you have to understand what they're trying to do and what you can do to help both of you do something smoothly so like when you're getting on to an elevator if you barge into the elevator when people are trying to get out you block the elevator and no one can move so the people who are getting on until the people get off and then they can get on and by doing so by understanding that there is a someone else who is trying to achieve a goal to try and do something then you can get both people acting very smoothly and they won't interfere with each other so that's kind of the basis of social interaction and socially interactive robots are robots that not only take their own goals into account but also the goals of people around them so how does a social interactive robot really work? so basically the main difference is that what I said that they have to understand what the people around them are trying to do and act accordingly so one of the things I like to say is that most of us think that the social rule for getting on and off to elevators is that you wait for everyone to get off the elevator before you get on but that's actually not the correct rule the correct rule is that you wait for everyone who intends to get off the elevator before you get on because some of the people aren't going to get off they're just waiting to go to another floor and so when we as people and the elevator opens when we as people decide whether we want to get on immediately or wait we have to look at the people in the elevator and try and understand their intention try and understand whether they want to get off or whether they want to stay and that understanding of the other people's intention is what is critical for making social interactive robots work they have to not only know what they want to do themselves but they have to perceive the people around them and try and understand what those people are trying to do and people don't have signs on them that says I'm getting off on floor 3 I'm getting off on floor 4 the robots have to look and try and understand cues such as gaze so when the elevator opens people who aren't getting off tend to look down people who are getting off tend to look up want to make sure it's the right floor some of them they start moving you have to detect them moving forward versus just rocking back and forth and those types of simple cues are things that are very difficult for robots to do so we pick up those cues very quickly because we've been trained from birth to pick up these social cues but robots have to be programmed to understand those cues and it turns out to be a very difficult thing to understand what those cues are so what techniques is CMU doing to enable robots to act and socially accept them so we're looking at basically two different types of social interaction one is what I call conversational interaction which is basically what we're doing right now so while I'm talking to you you're looking at me and you're nodding at me to indicate that you understand what I'm talking about and if you were to start looking off over to the side there I would know that I had lost you and I need to say something in order to get you back into the conversation so that's kind of conversational interaction and Hala is an example of a conversational social interactive robot so it uses gaze to try and bring people in to engage them it watches them as they type on the keyboard and it knows enough about how conversations run to respond appropriately someone says hello the robot will say hello someone says salam and so it does the appropriate thing in the right context the other aspect that we're looking at is what I call navigational interaction and that is how robots can navigate through space in a socially interactive way so the getting on and off elevators in a socially interactive way passing in corridors so when people are coming towards each other in a corridor one person moves to the right the other person moves to the right and they pass without really having to think very hard about how to do that in fact it's not only a social rule it's a cultural rule so in places like England and Japan where they drive on the left hand side they don't pass on the right they pass on the left so if you were to go to England and not understand that then you would start bumping into people so you have to know what the social rule is in order to work well in the society okay so how do you see the international efforts in the area of social interactive robots and where do you see the cutter fits in so when we started this work we were focused on social interaction and the idea was that there were these kind of universal social rules that people obeyed and it turns out that that was simplistic that it actually is not only social rules but cultural rules as well so the way that people interact in different parts of the world differ somewhat according to the culture so a middle eastern culture has somewhat different social rules than a western culture and so what we're looking at in Qatar especially with Hala is to try to understand those cultural differences and be able to develop a system that can be social and culturally appropriate in Qatar and at the same time if you were to take it to the United States it would be socially and culturally appropriate in the United States by changing a few a few variables a few parameters and right now what we're doing is just trying to explore the various ways that these robots can differ but one example in navigational interaction is that there is this idea of personal space how close people will come to one another when they're talking and in different cultures the personal space is larger or smaller and we can characterize that we can say if you're in Mediterranean culture personal space is much smaller and we can tell the robot this is the size of the personal space when you're in Italy for example versus the United States where it's much larger and then the robot can behave like an Italian or like an American so what do we mean when we speak about the gap between the robotic behavior what do we mean by this gap so again the basic gap is understanding other people it's this idea that the robot is not operating in isolation but it's operating in a people environment and it has to understand not only what it is trying to do but what the people around it are trying to do and to behave is that facilitate not only its own actions its own goals but the goals of the other people so when you when people are typically when you talk about people being antisocial you say he's very rude he's very antisocial what they typically mean is the person only thinks about himself and doesn't think about anyone else and that's the same thing for the robots if they behave only thinking about themselves then they tend to look antisocial and when they start considering other people then they look more social so the main thing that we have to look at is how we can get them to understand that there are people around and what those people are interested in doing whether they want to have a conversation with the robot or be left alone or be escorted somewhere or that they have certain other needs all those things need to be understood by the robot in order to be looked at by the social so what future uses can social interactive robots be used so my personal philosophy is that social interaction can be applied anywhere that we have technology so not just robots so you might think of the robot receptionist is a very good example service robots so a robot that is coming into clean floors it needs to know that there are other people around there maybe it shouldn't come into this room because there are people working it goes to some other room first and then comes back to this room later so it doesn't disturb the people maybe when it comes in it asks very politely is it okay to clean the room here rather than just barging in and cleaning the floors but other technology as well we use these days that are electronic cell phones, microwaves DVRs, television stereo all of those have very unsocial interfaces buttons and knobs they don't understand what people are trying to do think about a world in which your microwave knows how you like your tea to be heated in the morning the temperature of the water and when you come in you don't have to press all these buttons it just recognizes here's someone with a tea cup I know that he likes it at this particular temperature and you just put it in and it says oh good morning nice to see you again I'm going to heat your tea up like you like it and you can say no I'd like it a little warmer today than usual okay I'll do that for you that's social interaction you can think that's how you would do it if two people are couple living together you don't have to you don't have to press buttons you just talk to them and they understand you and they know what you like and those types of things that type of social interaction can be used anywhere that we have technology it would make interacting with our technology so much more useful when we have a social interactive robot how can we evaluate it and make sure that how can we assess its effectiveness so we're very lucky in that respect because psychologists have been doing experiments on humans how humans interact for many many years and so there's a understood methodology for understanding human-human interaction and measuring it and measuring its effectiveness so basically we use the same technologies we basically use the same techniques that psychologists use in order to evaluate and measure so we do a lot of user studies we bring people into the lab and ask them to perform certain tasks with the robot and then we look at how well they perform the tasks we ask them to rate their experiences with the robots and through that we understand which techniques work better which techniques work worse with HALA it's a slightly different thing because HALA is out there we don't invite people in but when people interact with HALA all the interactions are recorded and we can do experiments so one week we can run HALA with a certain program and the next week we run it with a slightly different program and we can compare how people interact with them how long they interact whether they say thank you to HALA when she answers their questions how frequently they do that how frequently what types of questions they ask and we can use that to try to understand whether changes that we make to HALA's behavior actually improves the interaction or worsens it and based on that we can go further in the research and develop other techniques so my final question tell us about the lecture you gave about social interactive robot at CMU on April 19th and how did you find the audience response I think it went rather well it's always hard for me to judge how my own talks go but there seemed to be a lot of interest in this idea of social robots actually what was interesting to me is there was a lot of discussion about why you would want to have a social robot so one of the things that people might ask is why would you want a robot that behaves like a human rather than a robot that behaves like a robot some people are afraid that if robots start behaving too much like humans that it will lessen the difference between robots and humans but so there was a lot of discussion about that I think though that in general you do want robots to act more like humans because we know how to interact with other humans we don't know how to interact with robots right now so if we have a robot that behaves differently than anything we've ever seen before we have to learn how to interact with it that takes a lot of effort so you put a robot in let's say a facility where the elderly are and have the robots try and help the elderly get around older people they're not going to want to learn how to interact with this very complex machine they know how to interact with people they know how to interact with the aids in the facility you just want them to the robots to interact in the same way so people don't have to learn new ways of interacting with this very complex technology so I want to thank Dr. Reed Simons for this interview and we wish you the best of luck thanks a lot