1
00:00:00,000 --> 00:00:04,540
of the real human data, or if that captures different answers.

2
00:00:04,540 --> 00:00:07,880
For example, if I say, instead of a sovereign German woman,

3
00:00:07,880 --> 00:00:11,820
you're a city English boy, I know,

4
00:00:11,820 --> 00:00:16,020
if you have explored those kind of things.

5
00:00:16,020 --> 00:00:20,920
Western politics of Hawaii

6
00:00:20,920 --> 00:00:25,480
kind of look at that as those democratic variables really

7
00:00:25,480 --> 00:00:28,880
matter in terms of that.

8
00:00:28,880 --> 00:00:34,020
Yeah, yeah, like the LLM was prompted

9
00:00:34,020 --> 00:00:37,060
to answer as a German woman, but then your sample

10
00:00:37,060 --> 00:00:40,600
is not all the German women, I guess, like the real data.

11
00:00:40,600 --> 00:00:43,380
So how do you account for those demographic variations

12
00:00:43,380 --> 00:00:46,380
when you prompt them all to answer just as one persona?

13
00:00:46,380 --> 00:00:46,880
You know?

14
00:00:46,880 --> 00:00:50,380
Yeah, yeah, yeah, that's what makes it so clear.

15
00:00:50,380 --> 00:00:55,260
So basically, he took like 200 persons for the survey,

16
00:00:55,260 --> 00:00:57,260
and all of those kind of .

17
00:00:58,880 --> 00:01:05,060
And did you explore like the difference between the personas

18
00:01:05,060 --> 00:01:05,720
that you made?

19
00:01:05,720 --> 00:01:10,700
Like did the LLM show some variability

20
00:01:10,700 --> 00:01:14,980
according to these different demographics?

21
00:01:14,980 --> 00:01:19,600
Yeah, yes, I had a lot of .

22
00:01:19,600 --> 00:01:22,900
I did run this analysis with the latest data.

23
00:01:22,900 --> 00:01:26,780
Previously, it was like, you could really kind of see that they kind of cast this kind of .

24
00:01:26,780 --> 00:01:27,060
They were like, I don't know, I'm not sure.

25
00:01:27,060 --> 00:01:27,320
I'm not sure.

26
00:01:27,320 --> 00:01:27,700
I'm not sure.

27
00:01:27,700 --> 00:01:27,860
I'm not sure.

28
00:01:27,860 --> 00:01:28,020
I'm not sure.

29
00:01:28,020 --> 00:01:28,280
I'm not sure.

30
00:01:28,280 --> 00:01:28,460
I'm not sure.

31
00:01:28,460 --> 00:01:28,860
I'm not sure.

32
00:01:29,400 --> 00:01:31,700
I don't keep them in a background.

33
00:01:31,700 --> 00:01:40,400
But they are like missing or some of that

34
00:01:40,400 --> 00:01:42,740
in contrast .

35
00:01:42,740 --> 00:01:44,900
Yes, there's some .

36
00:01:44,900 --> 00:01:46,200
There's some .

37
00:01:46,200 --> 00:01:48,260
Right. Well, thank you very much.

38
00:01:48,700 --> 00:01:54,540
The next presentation is bunun

39
00:01:54,540 --> 00:01:58,860
of theömni intensivpamc towards .

40
00:01:58,860 --> 00:02:02,360
in order to general intelligence with a causal temper approach

41
00:02:02,360 --> 00:02:04,800
to the cognitive execution substrate.

42
00:02:19,800 --> 00:02:20,280
Yes.

43
00:02:26,640 --> 00:02:28,400
Okay, thank you for this opportunity.

44
00:02:28,860 --> 00:02:37,000
Like my two main teams I'm interested in is how is cognitive science,

45
00:02:37,100 --> 00:02:43,080
is it possible for cognitive science to help to build AI systems?

46
00:02:45,300 --> 00:02:48,080
Or is it just an engineering task?

47
00:02:48,320 --> 00:02:53,840
And the other one is on the next slide,

48
00:02:53,840 --> 00:02:58,220
but I'm going to tractability.

49
00:02:58,860 --> 00:03:03,860
And then some questions related to these issues.

50
00:03:04,740 --> 00:03:07,860
And I finalize with the hypothesis that is kind of checklist

51
00:03:08,500 --> 00:03:14,860
that what we should know up to this point, maybe also later.

52
00:03:16,200 --> 00:03:27,620
So, my method is like all the participatory design, meaning that, okay, I set a goal that,

53
00:03:28,160 --> 00:03:28,740
like, unfiltered.

54
00:03:28,860 --> 00:03:29,620
I set a feasible goal.

55
00:03:30,520 --> 00:03:37,420
Let's build a robot which has AGI capabilities, meaning artificial general capabilities.

56
00:03:37,580 --> 00:03:39,620
Obviously, that's kind of unfeasible.

57
00:03:39,800 --> 00:03:49,820
But the purpose is that in the process we find problems and criteria that are, you know, on the way.

58
00:03:51,180 --> 00:03:58,040
So, and the other one good thing is that this kind of moves from the classical symbolic paradigm,

59
00:03:58,240 --> 00:03:58,740
symbolic paradigm to the AI.

60
00:03:58,740 --> 00:03:58,820
So, I'm going to talk about the AI.

61
00:03:58,820 --> 00:03:58,840
So, I'm going to talk about the AI.

62
00:03:59,000 --> 00:04:06,340
And it's a kind of manipulation paradigm to the embodied, embedded, and enacted and extended cognition,

63
00:04:06,920 --> 00:04:12,740
because it's a robot and it functions in a physical world.

64
00:04:13,400 --> 00:04:16,860
And this causal chamber is a device from physics.

65
00:04:17,640 --> 00:04:28,100
It's a box where you can measure and do, measure real physical phenomena and do experiments in that way,

66
00:04:28,220 --> 00:04:28,620
in real physical phenomena.

67
00:04:28,620 --> 00:04:37,920
real world instead of in simulation world so I have this thinking tool or

68
00:04:37,920 --> 00:04:48,600
concept of cognitive executive substrate meaning that everything that our head is

69
00:04:48,600 --> 00:04:54,700
the executive organ that interprets every language data just like in

70
00:04:54,700 --> 00:05:02,980
computer we have processor the processor executes certain kinds of programs so

71
00:05:02,980 --> 00:05:11,020
like we humans we execute philosophy we understand philosophy through this to

72
00:05:11,020 --> 00:05:16,660
our brain to the system all the knowledge we have and all the skills we

73
00:05:16,660 --> 00:05:24,480
have we need that in order to interpret that so and and the point is that

74
00:05:24,480 --> 00:05:24,680
that

75
00:05:24,680 --> 00:05:32,100
we should think always what is the system we are kinda what is the system

76
00:05:32,100 --> 00:05:38,760
that kinda executes the theories or anything or problems that we investigate

77
00:05:38,760 --> 00:05:42,800
instead of like in many times in philosophy you focus in the discourse

78
00:05:42,800 --> 00:05:49,460
what people say and what happens in this like conceptual realm

79
00:05:49,460 --> 00:05:54,460
it's like same thing like in you can have extended mathematics but

80
00:05:54,460 --> 00:06:01,240
only part of that is relevant for physics same thing here we should kind

81
00:06:01,240 --> 00:06:12,640
of ground our our theories in in cognitive executive substrate that's the

82
00:06:12,640 --> 00:06:16,720
conceptual tool so what is artificial general intelligence there are many

83
00:06:16,720 --> 00:06:24,700
definitions but but I like this one it's quite practical there's similarly

84
00:06:24,700 --> 00:06:30,100
like self-driving cars there's this levels of autonomy and then there is

85
00:06:30,100 --> 00:06:44,680
this is it the general system or is it very narrow narrow system and yeah and

86
00:06:44,680 --> 00:06:46,520
this image shows that

87
00:06:46,720 --> 00:06:58,660
resistance is futile that the crafts are going that direction where the in

88
00:06:58,660 --> 00:07:05,340
different tasks AI is able to surpass human performance so that's why I'm

89
00:07:05,340 --> 00:07:12,960
interested in in knowing why how would computer science help in going that

90
00:07:12,960 --> 00:07:15,220
direction

91
00:07:16,720 --> 00:07:23,880
but there's a problem there's this practical condition thesis that states

92
00:07:23,880 --> 00:07:32,020
that the AGI is not computable in practice

93
00:07:32,020 --> 00:07:37,540
5 minutes I have 5 minutes what left

94
00:07:37,540 --> 00:07:46,540
ok so but there's a problem

95
00:07:46,720 --> 00:07:54,280
they don't define that they define the problem problem space to broadly without

96
00:07:54,280 --> 00:08:01,480
any inductive biases and they think AI system as a monolithic agent instead of

97
00:08:01,480 --> 00:08:06,400
they should think it as a modular agent

98
00:08:06,400 --> 00:08:11,400
and

99
00:08:11,400 --> 00:08:14,400
yeah

100
00:08:16,720 --> 00:08:46,720


101
00:08:46,720 --> 00:09:16,720


102
00:09:16,720 --> 00:09:46,720


103
00:09:46,720 --> 00:10:16,720


104
00:10:16,720 --> 00:10:28,500
, that by its own perspective like what it like to cognize perspective without

105
00:10:28,500 --> 00:10:35,020
without minimal or suitable inductive biases that we throw into them how would

106
00:10:35,020 --> 00:10:46,680
that system kind of be able to create AGI so instead of having a thinking a

107
00:10:46,720 --> 00:10:54,580
model monolithic system we should have that set of modules where each module

108
00:10:54,580 --> 00:11:03,420
represents a cognitive capacity its module is tractable as an algorithm and

109
00:11:03,420 --> 00:11:13,560
also in this into input space the modules are organized in an architecture

110
00:11:13,560 --> 00:11:16,660
where each module is independent

111
00:11:16,720 --> 00:11:26,720
there's no cycles, and they are best effort broadcasters, meaning that they just function

112
00:11:26,720 --> 00:11:33,060
as their own and get modulated by others.

113
00:11:33,060 --> 00:11:43,240
Then these models should be, they can mimic sensitive periods and we can learn from, start

114
00:11:43,240 --> 00:11:56,420
from most digital levels like sensory levels and build on those and the whole system has

115
00:11:56,420 --> 00:12:01,800
different time loops where the lowest, the most digital levels are the most fastest and

116
00:12:01,800 --> 00:12:03,580
the highest levels are more fastest.

117
00:12:03,580 --> 00:12:12,100
So I woke up this morning and I saw, I realized in a dream there was something more, but I

118
00:12:12,100 --> 00:12:13,100
forgot it.

119
00:12:13,100 --> 00:12:23,220
I woke up so, so I invite you to like tell me what is missing here or what I need to

120
00:12:23,220 --> 00:12:26,940
consider more.

121
00:12:26,940 --> 00:12:31,980
Yes, any questions?

122
00:12:31,980 --> 00:12:40,280
I think there's time for one or two more questions, but it's probably great and then it's time

123
00:12:40,280 --> 00:12:41,280
for the next one.

124
00:12:41,280 --> 00:12:42,280
Great.

125
00:12:42,280 --> 00:12:42,780
Sorry.

126
00:12:43,100 --> 00:12:44,100
Thank you.

127
00:12:44,100 --> 00:12:45,100
Hi, Janne.

128
00:12:45,100 --> 00:12:46,100
Excellent.

129
00:12:46,100 --> 00:12:47,100
Thank you.

130
00:12:47,100 --> 00:12:48,100
Always nice to have discussions with you.

131
00:12:48,100 --> 00:12:49,100
You are researching on a very broad basis here.

132
00:12:49,100 --> 00:12:50,100
I have one very general question for you.

133
00:12:50,100 --> 00:12:55,100
So sorry to read it out from paper and sorry for making you to listen to it.

134
00:12:55,100 --> 00:13:06,100
But even if AGI, which we have, would be better in every single task than total analysis,

135
00:13:06,100 --> 00:13:09,940
what would you say would be the best solution for the data?

136
00:13:09,940 --> 00:13:10,940
So, I'm sorry.

137
00:13:10,940 --> 00:13:11,940
I'm sorry.

138
00:13:11,940 --> 00:13:12,940
I'm sorry.

139
00:13:12,940 --> 00:13:15,940
I'm sorry when I say this one's not the best solution.

140
00:13:15,940 --> 00:13:18,940
What I actually need to say is, that perhaps, creators andudio найти a superior intelligence

141
00:13:18,940 --> 00:13:19,980
to strawberry plums that you describe not merely as man, but as an even Senhor and you

142
00:13:19,980 --> 00:13:20,980
know, not only as this man top human, it might not mean that it has superior intelligence.

143
00:13:20,980 --> 00:13:24,500
This is rest injection on the position that intelligence is made of different common tasks.

144
00:13:24,500 --> 00:13:29,940
There's always a possibility about an entirely new task and a need for what we call shave

145
00:13:29,940 --> 00:13:30,940
on creativity.

146
00:13:30,940 --> 00:13:33,740
The long of human intelligence.

147
00:13:33,740 --> 00:13:38,940
So just coming up with different cognitive tasks in a machine does not create anything

148
00:13:38,940 --> 00:13:39,900
more than the superhuman and

149
00:13:39,900 --> 00:13:40,940
all is clear.

150
00:13:40,940 --> 00:13:41,940
Know?

151
00:13:41,940 --> 00:13:49,220
more and there's some of those parts it doesn't create suddenly something this is a promise and

152
00:13:49,220 --> 00:13:56,500
hype of or has been of ai that suddenly there will be this emergence of the general intelligence and

153
00:13:57,060 --> 00:14:05,300
doesn't seem to happen we still have to teach it different tasks i don't know uh my answer would be

154
00:14:05,300 --> 00:14:13,220
like uh using the other mind problem and the monolithic agent problem that if we are able if

155
00:14:13,220 --> 00:14:22,420
if we are afraid to that ai is replacing humans then then it certainly has certain kinds of

156
00:14:22,420 --> 00:14:33,460
capacities and the kind of also like if the tractable cognizant thesis is true then we can

157
00:14:33,460 --> 00:14:35,140
also they have to say that

158
00:14:35,860 --> 00:14:42,740
because they are critical everybody's afraid of ai doom so i would i would ask them that how do you

159
00:14:42,740 --> 00:14:49,540
prove that the ai doom is not intractable i would i would kind of throw it out there that it's

160
00:14:49,540 --> 00:14:56,260
intractable if we are in that school but if if we are in my school then

161
00:14:59,460 --> 00:15:02,740
it's tractable but i don't believe

162
00:15:04,740 --> 00:15:05,140
they do

163
00:15:05,300 --> 00:15:12,740
thank you

164
00:15:12,740 --> 00:15:25,860
now we have a coffee break of half an hour and 2 30 is the last keynote speech thanks again to all

165
00:15:25,860 --> 00:15:27,860
the great presentations

166
00:15:27,860 --> 00:15:31,860
yes

167
00:15:31,860 --> 00:15:33,860
yes

168
00:15:33,860 --> 00:15:35,140
yes

169
00:15:35,300 --> 00:16:05,300


170
00:16:05,620 --> 00:16:07,860
for you

171
00:16:07,860 --> 00:16:13,860
i

172
00:16:13,860 --> 00:16:14,660
thank

173
00:16:14,660 --> 00:16:14,740
you