Nadia Magnenat Thalmann: Robots support us humans

edited by Melinda Crane

Nadine was rec­og­nized as the most real­is­tic robot in the world because she is real­ly a copy of me and my daugh­ters and the sec­ond thing that was new is that she can sim­u­late emo­tions and can remem­ber peo­ple and facts.

It’s the idea to have com­pan­ions. Com­pan­ions that can under­stand us, be with us when nobody else is there, help us in all kinds of func­tions. That was my dream 30 years ago and it con­tin­ues to be.

Nadine tracks emo­tions in two ways. One is from lan­guage. We work quite close­ly with speech under­stand­ing, if I speak loud­er or if I have emo­tion in my voice, all that is ana­lyzed. And then in addi­tion with AI we ana­lyze expres­sions. We can detect if some­body is hap­py, we detect ges­tures, for exam­ple if I’m ner­vous. She has to mod­el her own reac­tions through the soft­ware we have giv­en her, we are work­ing on includ­ing com­mon sense so that she doesn’t answer like a robot. Our goal is to have a sim­u­la­tion of human behav­ior so that these humanoid robots will behave in a way that they are trustable and real com­pan­ions. You have a feel­ing of pres­ence. Less lone­li­ness.

Nadine has no emo­tion! It’s sci­ence fic­tion when peo­ple sell the idea that machines or humanoids could feel. They are just wafers, actu­a­tors, soft­ware in a com­put­er. They are machines but they sim­u­late emo­tions.

In the devel­op­ment of Nadine I decid­ed, that she should be authen­tic and always says, “I am a machine, I don’t feel any­thing.” She is very hon­est. In the end Nadine will become more intel­li­gent and will be able to do more things but humans are so much more com­plex and the inter­ac­tion we have with humans is so much rich­er, I mean it’s no com­par­i­son.

In Asia, they are more prag­mat­ic and more tech­nol­o­gy-dri­ven. So the accep­tance rate is very high. Nadine just came back from a 6‑month intern­ship at AIA Insur­ance, a huge com­pa­ny in Asia. Nadine took the role of a cus­tomer agent and met cus­tomers. In Sin­ga­pore the demand is enor­mous. And the gov­ern­ment asked me to pro­duce robots for elder­ly peo­ple. We have start­ed in Sin­ga­pore to work with psy­chi­a­trists. Patients for exam­ple with pos­si­ble demen­tia or bipo­lar­i­ty: If we can ana­lyze ges­tures and emo­tion over time, we have quan­ti­ta­tive mea­sure­ment that will help to deter­mine facts. In psy­chi­a­try it could help a lot to have Nadine ana­lyze behav­ior. After that you have a huge data set that allows a bet­ter diag­no­sis.

“IN MY WORK WITH ROBOTS I WANT TO CREATE COMPANIONS WHO SUPPORT US HUMANS.”

To say that robots will take over, I don’t think it’s true. The more we devel­op, the more there will be new jobs. I think that by going ahead we will cre­ate new jobs and some oth­ers will dis­ap­pear. If we as humans have the capac­i­ty to cre­ate tools to help us to go for­ward faster, that’s great. It’s up to us to decide what we do with our tools. I think it’s time to meet with lawyers, with politi­cians, to dis­cuss how we use these tools and what should be insert­ed into the soft­ware. Research is going in this direc­tion, that peo­ple add lay­ers that con­trol the robots or con­trol the lim­its of their actions.

I think that tech­nol­o­gy per se is not dan­ger­ous. What is dan­ger­ous are peo­ple. So the prob­lem is not with tech­nol­o­gy, the prob­lem is with human behav­ior.

VITA

Univ.-Prof. Dr. Nadia Mag­ne­nat Thal­mann is a com­put­er graph­ics sci­en­tist and uni­ver­si­ty pro­fes­sor. The Direc­tor of the Insti­tute for Media Inno­va­tion at the NTU in Sin­ga­pore is also the founder and Direc­tor of MIRAL­ab, an inter­dis­ci­pli­nary lab­o­ra­to­ry for human com­put­er ani­ma­tion at the Uni­ver­si­ty of Gene­va.

Her research areas main­ly include social robots, vir­tu­al real­i­ties and med­ical sim­u­la­tion. Nadia Mag­ne­nat Thal­mann con­tributed to the devel­op­ment of com­put­er graph­ics dur­ing her doc­tor­al stud­ies by sim­u­lat­ing the 3D den­si­ty of the approx­i­mate solu­tions of the Schrödinger equa­tion. Lat­er, she occu­pied her­self with mod­el­ling real­is­tic vir­tu­al humans and pro­duced the first sim­u­la­tion of a 3D ver­sion of Mar­i­lyn Mon­roe. Her prob­a­bly most famous project is the social robot Nadine, who is able to speak, to rec­og­nize peo­ple and ges­tures, to express emo­tions and to remem­ber.

Nadia Mag­ne­nat Thal­mann received i.a. a Doc­tor Hon­oris Causa in Sci­ence from Leib­niz Uni­ver­si­ty Han­nover, an hon­orary doc­tor­ate from the Uni­ver­si­ty of Ottawa, a Career Achieve­ment Award from the Cana­di­an Human Com­put­er Com­mu­ni­ca­tions Soci­ety in Toron­to and the Hum­boldt Research Award in Ger­many.