Subject-centric

[Subject-centric @subjectcentric]

Asked Bing GPT to translate several simple English phrases to Text Meaning Representation (TMR). It created nested list structure with proper identification of objects, events, semantic roles, and properties. Second time - somewhat relevant syntactic tree…

[Subject-centric @subjectcentric]

It seems that conceptual machinery developed in ontological semantics can help with understanding of large language models https://t.co/QUImoKSjwp

[Subject-centric @subjectcentric]

Interesting reference to ‘semantic grammar’ https://t.co/MH1DiUD5IV

[Joscha Bach @Plinz]

Today, I discussed whether today's large neural networks might be slightly conscious, with @_DigitalFUTURES
https://t.co/kG8p6121DQ

[Subject-centric @subjectcentric]

Statistical approach used in ‘Delphi: towards machine ethics and norms’ https://t.co/dFn9ti1Xm8 is not ideal, but informative and the paper (and the discussion around it) generates lots of good questions

[Subject-centric @subjectcentric]

It seems like ‘Delphi: towards machine ethics and norms’ https://t.co/dFn9ti1Xm8 generates quite a discussion https://t.co/ccnYfFvs0q. I would recommend to separate a problem/goal from a very very specific approach/implementation

[Subject-centric @subjectcentric]

Fascinating research ’Delphi: towards machine ethics and norms’ https://t.co/dFn9ti1Xm8. ‘…As it is, Delphi is a neural snapshot of its training data, which can be used to study present perceptions of ethics and morality…’

[Subject-centric @subjectcentric]

Finch (movie) - interesting case of simulated consciousness and "Finch test" (can a dog "be OK" with an agent with simulated consciousness?)

[Subject-centric @subjectcentric]

It is amazing that an algorithm encoded in physical matter (computer) could drive interaction with external physical environment

[Subject-centric @subjectcentric]

One of the interesting differences between theoretical computational models such as Turing machine and computations in real life is that IRL computations are physical, running algorithms create measurable energy patterns

[Subject-centric @subjectcentric]

Nice, very nice! - section FEELING MACHINES AND CONSCIOUS MACHINES in Feeling and Knowing: Making Minds Conscious by Antonio Damasio

[Subject-centric @subjectcentric]

Ex Machina hypothesis (4): agents with simulated consciousness will be able to simulate experience of ’being’; related ‘images’ will always be present but could be out of focus/attention

[Subject-centric @subjectcentric]

Nice description: combination of “emotional” and “homeostatic” feelings into a new affective state in Feeling and Knowing: Making Minds Conscious by Antonio Damasio

[Subject-centric @subjectcentric]

’…Feeling probably began its evolutionary history as a timid conversation between the chemistry of life and the early version of a nervous system within one particular organism…’ Feeling and Knowing: Making Minds Conscious by Antonio Damasio

[Subject-centric @subjectcentric]

Ex Machina hypothesis (3): many people will respect rich internal world of agents with simulated consciousness and care about internal states of these agents

[Subject-centric @subjectcentric]

Ex Machina hypothesis (2): agents with simulated consciousness will be able to include simulated feelings into behaviour without feeling them

[Subject-centric @subjectcentric]

Ex Machina hypothesis (1): it is possible to construct artificial agents with very rich simulated mental models including ability to imagine and learn

[Subject-centric @subjectcentric]

Interesting considerations related to the Garland/Ex Machina test https://t.co/HMpCMzG1nH

[Subject-centric @subjectcentric]

‘…We are not cognitive computers, we are feeling machines…’ https://t.co/HMpCMzG1nH - which could be simulated with computers

[Subject-centric @subjectcentric]

‘…The paintings become experiments into predictive perception and into the nature of the conscious experiences that these processes give rise to…’ https://t.co/HMpCMzG1nH