start end text 0 4540 of the real human data, or if that captures different answers. 4540 7880 For example, if I say, instead of a sovereign German woman, 7880 11820 you're a city English boy, I know, 11820 16020 if you have explored those kind of things. 16020 20920 Western politics of Hawaii 20920 25480 kind of look at that as those democratic variables really 25480 28880 matter in terms of that. 28880 34020 Yeah, yeah, like the LLM was prompted 34020 37060 to answer as a German woman, but then your sample 37060 40600 is not all the German women, I guess, like the real data. 40600 43380 So how do you account for those demographic variations 43380 46380 when you prompt them all to answer just as one persona? 46380 46880 You know? 46880 50380 Yeah, yeah, yeah, that's what makes it so clear. 50380 55260 So basically, he took like 200 persons for the survey, 55260 57260 and all of those kind of . 58880 65060 And did you explore like the difference between the personas 65060 65720 that you made? 65720 70700 Like did the LLM show some variability 70700 74980 according to these different demographics? 74980 79600 Yeah, yes, I had a lot of . 79600 82900 I did run this analysis with the latest data. 82900 86780 Previously, it was like, you could really kind of see that they kind of cast this kind of . 86780 87060 They were like, I don't know, I'm not sure. 87060 87320 I'm not sure. 87320 87700 I'm not sure. 87700 87860 I'm not sure. 87860 88020 I'm not sure. 88020 88280 I'm not sure. 88280 88460 I'm not sure. 88460 88860 I'm not sure. 89400 91700 I don't keep them in a background. 91700 100400 But they are like missing or some of that 100400 102740 in contrast . 102740 104900 Yes, there's some . 104900 106200 There's some . 106200 108260 Right. Well, thank you very much. 108700 114540 The next presentation is bunun 114540 118860 of theömni intensivpamc towards . 118860 122360 in order to general intelligence with a causal temper approach 122360 124800 to the cognitive execution substrate. 139800 140280 Yes. 146640 148400 Okay, thank you for this opportunity. 148860 157000 Like my two main teams I'm interested in is how is cognitive science, 157100 163080 is it possible for cognitive science to help to build AI systems? 165300 168080 Or is it just an engineering task? 168320 173840 And the other one is on the next slide, 173840 178220 but I'm going to tractability. 178860 183860 And then some questions related to these issues. 184740 187860 And I finalize with the hypothesis that is kind of checklist 188500 194860 that what we should know up to this point, maybe also later. 196200 207620 So, my method is like all the participatory design, meaning that, okay, I set a goal that, 208160 208740 like, unfiltered. 208860 209620 I set a feasible goal. 210520 217420 Let's build a robot which has AGI capabilities, meaning artificial general capabilities. 217580 219620 Obviously, that's kind of unfeasible. 219800 229820 But the purpose is that in the process we find problems and criteria that are, you know, on the way. 231180 238040 So, and the other one good thing is that this kind of moves from the classical symbolic paradigm, 238240 238740 symbolic paradigm to the AI. 238740 238820 So, I'm going to talk about the AI. 238820 238840 So, I'm going to talk about the AI. 239000 246340 And it's a kind of manipulation paradigm to the embodied, embedded, and enacted and extended cognition, 246920 252740 because it's a robot and it functions in a physical world. 253400 256860 And this causal chamber is a device from physics. 257640 268100 It's a box where you can measure and do, measure real physical phenomena and do experiments in that way, 268220 268620 in real physical phenomena. 268620 277920 real world instead of in simulation world so I have this thinking tool or 277920 288600 concept of cognitive executive substrate meaning that everything that our head is 288600 294700 the executive organ that interprets every language data just like in 294700 302980 computer we have processor the processor executes certain kinds of programs so 302980 311020 like we humans we execute philosophy we understand philosophy through this to 311020 316660 our brain to the system all the knowledge we have and all the skills we 316660 324480 have we need that in order to interpret that so and and the point is that 324480 324680 that 324680 332100 we should think always what is the system we are kinda what is the system 332100 338760 that kinda executes the theories or anything or problems that we investigate 338760 342800 instead of like in many times in philosophy you focus in the discourse 342800 349460 what people say and what happens in this like conceptual realm 349460 354460 it's like same thing like in you can have extended mathematics but 354460 361240 only part of that is relevant for physics same thing here we should kind 361240 372640 of ground our our theories in in cognitive executive substrate that's the 372640 376720 conceptual tool so what is artificial general intelligence there are many 376720 384700 definitions but but I like this one it's quite practical there's similarly 384700 390100 like self-driving cars there's this levels of autonomy and then there is 390100 404680 this is it the general system or is it very narrow narrow system and yeah and 404680 406520 this image shows that 406720 418660 resistance is futile that the crafts are going that direction where the in 418660 425340 different tasks AI is able to surpass human performance so that's why I'm 425340 432960 interested in in knowing why how would computer science help in going that 432960 435220 direction 436720 443880 but there's a problem there's this practical condition thesis that states 443880 452020 that the AGI is not computable in practice 452020 457540 5 minutes I have 5 minutes what left 457540 466540 ok so but there's a problem 466720 474280 they don't define that they define the problem problem space to broadly without 474280 481480 any inductive biases and they think AI system as a monolithic agent instead of 481480 486400 they should think it as a modular agent 486400 491400 and 491400 494400 yeah 496720 526720 526720 556720 556720 586720 586720 616720 616720 628500 , that by its own perspective like what it like to cognize perspective without 628500 635020 without minimal or suitable inductive biases that we throw into them how would 635020 646680 that system kind of be able to create AGI so instead of having a thinking a 646720 654580 model monolithic system we should have that set of modules where each module 654580 663420 represents a cognitive capacity its module is tractable as an algorithm and 663420 673560 also in this into input space the modules are organized in an architecture 673560 676660 where each module is independent 676720 686720 there's no cycles, and they are best effort broadcasters, meaning that they just function 686720 693060 as their own and get modulated by others. 693060 703240 Then these models should be, they can mimic sensitive periods and we can learn from, start 703240 716420 from most digital levels like sensory levels and build on those and the whole system has 716420 721800 different time loops where the lowest, the most digital levels are the most fastest and 721800 723580 the highest levels are more fastest. 723580 732100 So I woke up this morning and I saw, I realized in a dream there was something more, but I 732100 733100 forgot it. 733100 743220 I woke up so, so I invite you to like tell me what is missing here or what I need to 743220 746940 consider more. 746940 751980 Yes, any questions? 751980 760280 I think there's time for one or two more questions, but it's probably great and then it's time 760280 761280 for the next one. 761280 762280 Great. 762280 762780 Sorry. 763100 764100 Thank you. 764100 765100 Hi, Janne. 765100 766100 Excellent. 766100 767100 Thank you. 767100 768100 Always nice to have discussions with you. 768100 769100 You are researching on a very broad basis here. 769100 770100 I have one very general question for you. 770100 775100 So sorry to read it out from paper and sorry for making you to listen to it. 775100 786100 But even if AGI, which we have, would be better in every single task than total analysis, 786100 789940 what would you say would be the best solution for the data? 789940 790940 So, I'm sorry. 790940 791940 I'm sorry. 791940 792940 I'm sorry. 792940 795940 I'm sorry when I say this one's not the best solution. 795940 798940 What I actually need to say is, that perhaps, creators andudio найти a superior intelligence 798940 799980 to strawberry plums that you describe not merely as man, but as an even Senhor and you 799980 800980 know, not only as this man top human, it might not mean that it has superior intelligence. 800980 804500 This is rest injection on the position that intelligence is made of different common tasks. 804500 809940 There's always a possibility about an entirely new task and a need for what we call shave 809940 810940 on creativity. 810940 813740 The long of human intelligence. 813740 818940 So just coming up with different cognitive tasks in a machine does not create anything 818940 819900 more than the superhuman and 819900 820940 all is clear. 820940 821940 Know? 821940 829220 more and there's some of those parts it doesn't create suddenly something this is a promise and 829220 836500 hype of or has been of ai that suddenly there will be this emergence of the general intelligence and 837060 845300 doesn't seem to happen we still have to teach it different tasks i don't know uh my answer would be 845300 853220 like uh using the other mind problem and the monolithic agent problem that if we are able if 853220 862420 if we are afraid to that ai is replacing humans then then it certainly has certain kinds of 862420 873460 capacities and the kind of also like if the tractable cognizant thesis is true then we can 873460 875140 also they have to say that 875860 882740 because they are critical everybody's afraid of ai doom so i would i would ask them that how do you 882740 889540 prove that the ai doom is not intractable i would i would kind of throw it out there that it's 889540 896260 intractable if we are in that school but if if we are in my school then 899460 902740 it's tractable but i don't believe 904740 905140 they do 905300 912740 thank you 912740 925860 now we have a coffee break of half an hour and 2 30 is the last keynote speech thanks again to all 925860 927860 the great presentations 927860 931860 yes 931860 933860 yes 933860 935140 yes 935300 965300 965620 967860 for you 967860 973860 i 973860 974660 thank 974660 974740 you