Philosophy and Artificial Intelligence (Crash Course): Difference between revisions

From NCOR Wiki
Jump to navigationJump to search
 
(23 intermediate revisions by the same user not shown)
Line 23: Line 23:
::• the question of human identity – what does it mean to say that it would still be ''me'' that would ''live on''?
::• the question of human identity – what does it mean to say that it would still be ''me'' that would ''live on''?


:'''Saturday, October 2''' will begin with an introduction to how AI works. An AI application is a set of algorithms that can process information in a way that – in the typical case – partially emulates human behavior. Such algorithms presuppose in every case a certain model of the domain from which their input information is derived. We will explore how such models are built, dealing along the way with philosophical topics such as computability, explanation, and prediction.  
:'''Saturday, October 2''' will begin with an introduction to the parts of philosophy relevant to the understanding AI and the claims made on its behalf. An AI application is a set of algorithms that can process information in a way that – in the typical case – partially emulates human behavior. Such algorithms presuppose in every case a certain model of the domain from which their input information is derived. We will explore how such models are built, dealing along the way with philosophical topics such as computability, explanation, and prediction.  


:For some domains, we can create models based on physical laws or simple rules (such as the rules of chess). For most domains, however, the complexity of the relations involved prevents successful modeling. It is for this reason that we face difficulties when we try to use AI to predict, for example, who will win a football game or what will happen tomorrow on the financial markets.  
:For some domains, we can create models based on physical laws or simple rules (such as the rules of chess). For most domains, however, the complexity of the relations involved prevents successful (which means predictive) modeling. It is for this reason that we face difficulties when we try to use AI to predict, for example, who will win a football game or what will happen tomorrow on the financial markets.


:'''Sunday, October 3''' will use what we have learned on Day 1 to address the opportunities and limits of AI in modelling and emulating
:'''Sunday, October 3''' will use what we have learned on Day 1 to address the opportunities and limits of AI in modelling and emulating
Line 39: Line 39:
==Schedule==
==Schedule==
''Saturday''  
''Saturday''  
:09:00-10:45 What is Philosophy? An Introduction for Computer Scientists
 
::Searle's Chinese Room
09:00-10:45 '''What is Philosophy? An Introduction for Computer Scientists''' (BS)
::Human and Animal Intelligence
:Searle's Chinese Room
:11:00-12:30 Artificial Intelligence  
:Human and Animal Intelligence
::The Leg-Hutter Definition of 'Universal Intelligence'
:[https://buffalo.box.com/v/What-is-philosophy Video]
::Rodney Brooks and Insect Intelligence-Based Robotics
:[https://buffalo.box.com/v/Crash-Course-Introduction Slides]
:12:30-13:00 Lunch
:13:00-14:45 Logic and Complex Systems I
11:00-12:30 '''Philosophy and Artificial Intelligence ''' (JL)
::Models
:The Leg-Hutter Definition of 'Universal Intelligence'
::Logic Systems
:Rodney Brooks and Insect Intelligence-Based Robotics
::Complex Systems
:[https://buffalo.box.com/v/Philosophy-and-AI Video]
:15:00-16:45 Logic and Complex Systems II
:[https://buffalo.box.com/v/Intelligence Slides]
::Multivariate Distributions
 
::Non-predictive Models of Complex Systems
13:00-14:45 '''Logic and Complex Systems I''' (JL)
::Predictive Models of Complex Systems
:Models
:Logic Systems
:Complex Systems
:[https://buffalo.box.com/v/Logic-and-complex-systems-I Video I]
:[https://buffalo.box.com/v/ComplexSystems-I Slides I]
 
15:00-16:45 '''Logic and Complex Systems II''' (JL)
:Multivariate Distributions
:Non-predictive Models of Complex Systems
:Predictive Models of Complex Systems
:[https://buffalo.box.com/v/Logic-and-complex-systems-II Video II]
:[https://buffalo.box.com/v/Complexsystems-II Slides II]
:[https://buffalo.box.com/v/Logic-and-complex-systems-III Video III]
:[https://buffalo.box.com/v/Complexsystems-III Slides III]


''Sunday''
''Sunday''
:09:00-10:45 Transhumanism and Digital Immortality
 
::Why Digital Immortality is Impossible
09:00-10:45 '''Transhumanism and Digital Immortality''' (BS)
::The Biology of Transhumanism
:Why Digital Immortality is Impossible
:11:00-12:30 AI Ethics -- Why There Could Be No Robot Police
:The Biology of Transhumanism
:12:30-13:00 Lunch
:[https://buffalo.box.com/v/Digital-Immortality Video]
:13:00-14:45 How to Create Useful Artificial Intelligence
:[https://buffalo.box.com/v/Digital-Immortality Slides]
:15:00-16:45 Final Sum-Up and Question-Answer Session
 
11:00-12:30 '''AI Ethics -- Why There Could Be No Robot Police''' (BS)
:[https://buffalo.box.com/v/No-Robot-Cops Video]
:[https://buffalo.box.com/v/No-Robot-Caps Slides]
 
13:00-14:45 '''How to Create Useful Artificial Intelligence''' (JL)
:[https://buffalo.box.com/v/Creating-Useful-AI Video]
:[https://buffalo.box.com/v/Useful-Artificial-Intelligence Slides]
 
15:00-16:45 '''Final Sum-Up and Question-Answer Session''' (JL)
:[https://buffalo.box.com/v/Questions-and-Answers Video]


==Faculty==
==Faculty==

Latest revision as of 23:47, 12 December 2021

PHI598/498 Special Topics Philosophy and Artificial Intelligence (Crash Course) - 1 Credit Hour

Registration: Philosophy

Cross-listed with Computer Science and Engineering: Special Topics CSE 510/410 LEC SMIT

Registration: Computer Science

Schedule: Weekend of October 2-3, 2021

Faculty: Barry Smith and Jobst Landgrebe

Venue: 101 Davis, UB North Campus


Course outline

A range of traditional philosophical topics appear in a new light when addressed from the perspective of recent research in artificial intelligence. These include:
• consciousness (and self-consciousness) – can machines think? can they be aware that they are thinking?
• the nature of human and animal intelligence – might machines be one day more intelligent than human beings?
• the relation between brains and computers – is the brain a computer? could we enhance our brains by implanting chips inside our heads?
• the relation between mind and brain – might I, one day, be able to upload the contents of my brain and so live on, as an algorithm, in the cloud?
• the question of human identity – what does it mean to say that it would still be me that would live on?
Saturday, October 2 will begin with an introduction to the parts of philosophy relevant to the understanding AI and the claims made on its behalf. An AI application is a set of algorithms that can process information in a way that – in the typical case – partially emulates human behavior. Such algorithms presuppose in every case a certain model of the domain from which their input information is derived. We will explore how such models are built, dealing along the way with philosophical topics such as computability, explanation, and prediction.
For some domains, we can create models based on physical laws or simple rules (such as the rules of chess). For most domains, however, the complexity of the relations involved prevents successful (which means predictive) modeling. It is for this reason that we face difficulties when we try to use AI to predict, for example, who will win a football game or what will happen tomorrow on the financial markets.
Sunday, October 3 will use what we have learned on Day 1 to address the opportunities and limits of AI in modelling and emulating
• consciousness
• human language understanding and conversations (including discussion of the Chinese room argument and of the Turing test)
• social behaviour and ethics
• transhumanism, life extension, and digital immortality
• automation systems for business, engineering, and consumer applications
We will argue that AI will not bring cures for (most) deadly diseases, it will not replace human police with intelligent robots, and – except along certain narrow tracks, including game-playing and image recognition – it will not reach a level of intelligence that surpasses that of human beings. Along those narrow tracks, however, AI will continue to make impressive contributions to the future of humankind.

Schedule

Saturday

09:00-10:45 What is Philosophy? An Introduction for Computer Scientists (BS)

Searle's Chinese Room
Human and Animal Intelligence
Video
Slides

11:00-12:30 Philosophy and Artificial Intelligence (JL)

The Leg-Hutter Definition of 'Universal Intelligence'
Rodney Brooks and Insect Intelligence-Based Robotics
Video
Slides

13:00-14:45 Logic and Complex Systems I (JL)

Models
Logic Systems
Complex Systems
Video I
Slides I

15:00-16:45 Logic and Complex Systems II (JL)

Multivariate Distributions
Non-predictive Models of Complex Systems
Predictive Models of Complex Systems
Video II
Slides II
Video III
Slides III

Sunday

09:00-10:45 Transhumanism and Digital Immortality (BS)

Why Digital Immortality is Impossible
The Biology of Transhumanism
Video
Slides

11:00-12:30 AI Ethics -- Why There Could Be No Robot Police (BS)

Video
Slides

13:00-14:45 How to Create Useful Artificial Intelligence (JL)

Video
Slides

15:00-16:45 Final Sum-Up and Question-Answer Session (JL)

Video

Faculty

Jobst Landgrebe is a scientist and entrepreneur with a background in mathematics, neuroscience, neuroinformatics, and philosophy. Landgrebe is also the founder of Cognotekt, a German AI company which has since 2013 provided working systems used by companies in areas such as insurance claims management, real-estate management, and medical billing.

Barry Smith is Distinguished Professor of Philosophy with cross-appointments in the UB Departments of Biomedical Informatics, Computer Science and Engineering, and Neurology.


Background

Pre-requisites: This course is open to both philosophy and computer science and engineering students at both graduate and advanced undergraduate levels.

Grading: Grades will be assigned on the basis of class attendance and degree of active participation.

Optional preliminary reading:

Making AI Meaningful Again
There is no general AI