week 01_Introduction_Philosophy.pdf

Cognitive Neuroscience for AI Developers (CNAID) Welcome to CNAID ! [email protected] [email protected] [email protected] @anki_xyz @Krauss_PK 2  Videos of the curr...

Cognitive Neuroscience for AI Developers (CNAID) Welcome to CNAID ! [email protected] [email protected] [email protected] @anki_xyz @Krauss_PK 2  Videos of the current lecture will be on StudOn (Old Video lectures on fau.tv (just add on)) https://www.fau.tv/course/id/3265 StudOn course https://www.studon.fau.de/studon/goto.php?target=crs_5533963 3 Results of end-of-semester test in Winter 2023/24 Exam SoSe 24 Exam will be a pure single-choice exam -> only single choice questions 90 mins duration Just pens are allowed (no calculators etc.) 4 Time table Blue: Lecture of Achim Schilling or Patrick Krauss Green: Lecture by Andreas Kist Lectures are presence lectures on Wednesdays: 10:15 a.m., room: H6 Q&A session: exercise sheets on zoom on Tuesday 14:15 p.m (zoom link will be shared on studon) 5 Exercise sheet example 6 Topics 7 Literature Cognitive Science, Friedenberg, J., Silverman, G., & Spivey, M. J. (2021). Cognitive science: an introduction to the study of mind. Sage Publications. Fourth Edition Cognitive Science, Bermúdez, J. L. (2023). Cognitive science: An introduction to the science of the mind. Cambridge University Press. Fourth Edition Cognitive Neuroscience, Gazzaniga, M. S., Ivry, R. B., & Mangun, G. R. (2014). Cognitive Neuroscience. The biology of the mind,(2014). Fourth Edition Cognitive Neuroscience, Ward, J. (2015). The student's guide to cognitive neuroscience. psychology press. Third edition Neurobiology, Kandel, E. R., Schwartz, J. H., Jessell, T. M., Siegelbaum, S., Hudspeth, A. J., & Mack, S. (2013). Principles of neural science. New York: McGraw-hill. Fifth Edition 8  Cognitive Neuroscience for AI Developers Introduction to Cognitive Science The big picture -> What is Cognitive Science? Cognitive Science: scientific, interdisciplinary study of the mind. Mind: „[…] the complex of faculties involved in perceiving, remembering, considering, evaluating, and deciding. Mind is in some sense reflected in such occurrences as sensations, perceptions, emotions, memory, desires, various types of reasoning, motives, choices, traits of personality, and the unconscious.” https://www.britannica.com/topic/mind Cognition: higher mental processes such as thinking, perceiving, imagining, speaking, acting, planning 10 Cognitive Science Highly interdisciplinary approach Main method: Scientific method (not philosophy and older disciplines) Scientific method: test hypothesis with experiment -> update hypothesis -> new Experiment (iterative process) Voit, E. O. (2019). Perspective: Dimensions of the scientific method. PLOS Computational Biology, 15(9), e1007279. 11 Cognitive Science Scientific method Occams razor: „given two explanations of the data, https://de.mathworks.com/discovery/overfitting.html all other things being equal, the simpler explanation is preferable.” “This principle is very much alive today in the emerging science of machine learning, whose expressed goal is often to discover the simplest hypothesis that is consistent with the sample data.” Blumer, A., Ehrenfeucht, A., Haussler, D., & Warmuth, M. K. (1987). Occam's razor. Information processing letters, 24(6), 377-380. D. Angluin and C.H. Smith, Inductive inference: Theory and methods, Comput. Surv. 15 (3) (1983) 327-369. 12 Multi-disciplinary perspective Philosophy Cognitive Biological disciplines disciplines Biology Psychology Neuroscience Linguistics Cognitive Science Cognitive is not the sum, but Neuroscience the intersection of (Neuropsychology) all these disciplines -> goal is to Artificial Intelligence understand the Computer Science mind Robotics Computational Physics disciplines 13 The blind men and the elephant Source: https://medium.com/betterism/the-blind-men-and-the-elephant-596ec8a72a7d 14 The blind men and the elephant To understand the mind interdisciplinary communication and cooperation necessary! Cognitive science: Not sum but interaction of disciplines Source: https://medium.com/betterism/the-blind-men-and-the-elephant-596ec8a72a7d 15 Questions to be answered: How does the human mind work? How does cognition work? How is cognition implemented in the brain? How can cognition be implemented in machines? Some of the hardest scientific problems 1. Brain is hard to observe, measure and manipulate 2. Brain is most complex entity in the known universe 16 The most complex entity in the known universe 100 billion (1011) neurons 10,000 synapses per neuron 1015 synapses in total 1022 possible connections Source: brainline.org only 1 of 10,000,000 possible connections is actually realized 17 The most complex entity in the known universe 11 Number of possible brain states > 210 22 22 Number of possible connectomes > 210 ~ 1010 Number of protons in the observable universe ~ 1080 18 How do we unravel how the brain processes information -> Problem of cognitive science and cognitive neuroscience Brown, J. W. (2014). The tale of the neuroscientists and the computer: why mechanistic theory matters. Frontiers in neuroscience, 8, 349. 19 The tale of the neuroscientist and the computer “Once upon a time, a group of neuroscientists happened upon a computer (Carandini, 2012). Not knowing how it worked, they each decided to find out how it sensed a variety of inputs and generated the sophisticated output seen on its display.” Brown, J. W. (2014). The tale of the neuroscientists and the computer: why mechanistic theory matters. Frontiers in neuroscience, 8, 349. 20 The tale of the neuroscientist and the computer “The EEG researcher quickly went to work, putting an EEG cap on the motherboard and measuring voltages at various points all over it, including on the outer case for a reference point. She found that when the hard disk was accessed, the disk controller showed higher voltages on average, and especially more power in the higher frequency bands. When there was a lot of computation, a lot of activity was seen around the CPU….” https://brainvision.com/products/livecap/ Brown, J. W. (2014). The tale of the neuroscientists and the computer: why mechanistic theory matters. Frontiers in neuroscience, 8, 349. 21 The tale of the neuroscientist and the computer “Next, the enterprising physicist and cognitive neuroscientist came along. “We don’t have enough spatial resolution to see inside the computer,” they said. So they developed a new imaging technique by which activity can be measured, called the Metabolic Radiation Imaging (MRI) camera, which now measures the heat (infrared) given off by each part of the computer in the course of its operations. At first, they found simply that lots of math operations lead to heat given off by certain parts of the CPU, and that memory storage involved the RAM, and that file operations engaged the Brown, J. W. (2014). The tale of the neuroscientists and the computer: why mechanistic hard disk…” theory matters. Frontiers in neuroscience, 8, 349. 22 The tale of the neuroscientist and the computer “…Finally the neuropsychologist comes along. She argues (quite reasonably) that despite all of these findings of network interactions and voltage signals, we cannot infer that a given region is necessary without lesion studies. The neuropsychologist then gathers a hundred computers that have had hammer blows to various parts of the motherboard, extension cards, and disks. After testing their abilities exten- sively, she carefully selects just the few that have a specific problem with the video output. She finds that among computers that don’t display video properly, there is an overlapping area of damage to the video card. This means of course that the video card is necessary for proper video monitor functioning. …” Brown, J. W. (2014). The tale of the neuroscientists and the computer: why mechanistic theory matters. Frontiers in neuroscience, 8, 349. 23 Moral of the story Cognitive neuroscience is still in an early stage phase We need a mechanistic theory to understand cognition in the brain We have to develop computer models that produce testable hypotheses We need a multi-disciplinary approach Neuroscience alone is not enough Brown, J. W. (2014). The tale of the neuroscientists and the computer: why mechanistic theory matters. Frontiers in neuroscience, 8, 349. Schilling, A., Sedley, W., Gerum, R., Metzner, C., Tziridis, K., Maier, A.,... & Krauss, P. (2022). Predictive coding and stochastic resonance: Towards a unified theory of auditory (phantom) perception. arXiv preprint arXiv:2204.03354. 24 Main ideas of cognitive science Boost by invention of computer 25  The cognitive revolution Started in 1950s Psychology, linguistics were redefining themselves (backlash against behaviorism) Computer science and neuroscience came up Personal computer novel brain imaging techniques boosted the development In 1960: start of the interdisciplinary field Different names: information-processing psychology, cognitive studies, cognitive science Miller, G. A. (2003). The cognitive revolution: a historical perspective. Trends in cognitive sciences, 7(3), 141-144. 26 Central ideas of Cognitive Science Cognition is equivalent to computation / information processing -> The mind/brain is an information processor. Information processors represent and transform information - Mental representations of information - Mental processes that act on and manipulate these representations called computations 27 Mind – Computer – Analogy Input Representation Processing / Output (binary) Computation Source: adapterwelt.net © P. Krauss Source: store.hp.com Source: ifixit.com Source: lifescience.com © P. Krauss © P. Krauss Source: wikipedia.org Source: brainline.org 28 Representations (something stands for something else) Traditional cognitive science view ! 4 types: Concepts (stands for entity) „apple“ Propositions (statements about the world) „Mary has black hair.“ Rules (relationship between propositions) „If it is raining, I will bring my umbrella.“ Analogies (comparisons between situations) „Life is a roller coaster.“ 29 Representations Traditional cognitive science view ! Features of representations: Symbolic: a symbol is a surrogate and refers to its referent The mind $ Representation (symbolic) Intentionality: relationship between representation and what it is about The world Referent (non-symbolic) 30 Representations Symbols can be assembled into physical symbol system (formal logical system) Formal logical system: symbols are combined to expressions Formal processes: manipulate expressions to create new expressions -> Formal logical systems can allow for intelligence and intelligent behavior (Physical symbol system hypothesis (Newell & Simon, 1976)) Hypothesis is often criticized: computers use symbols with no meaning as symbols are not connected to the environment (grounding problem) Computers do not perceive their environment (no bodies) -> thus symbols have no meaning 31 Representations + Computation Example: formal logic Traditional cognitive science view ! Symbols: all, mammals, …. Expression: all and only mammals nurse their young Processes: rules of deduction derive new, true expressions from known expressions Expression 1: all and only mammals nurse their young Expression 2: whales nurse their young New expression: whales are mammals -> Newell and Simon 1976 -> intelligence 32  Representations + Computation Expert systems Solves problems were you normally need a human specialist (expert) Uses set of if-then rules Example: MYCIN (Shortliffe) -> regarded as first expert system Diagnose of blood infections and meningitis Helped to choose antibiotics Simple inference machine with approx. 500 rules Asks questions to the physician (yes/ no) Duda, R. O., & Shortliffe, E. H. (1983). Expert systems research. Science, 220(4594), 261-268. Lacave, C., & Diez, F. J. (2004). A review of explanation methods for heuristic expert systems. The Knowledge Engineering Review, 19(2), 133-146. Duda, R. O., & Shortliffe, E. H. (1983). Expert systems research. Science, 220(4594), 261-268. Qiu, S., Sallak, M., Schön, W., & Ming, H. X. (2018). A valuation-based system approach for risk assessment of belief rule-based expert systems. Information Sciences, 466, 323-336. 33 Computation Traditional cognitive science view ! The mind performs computations on representations. e.g. language: putting a verb into past tense math: adding two numbers etc. -> endless list Define broad categories of mental operations according to: Type of operation Type of information that is processed -> Any information processing can be described at 3 different levels 34 Computation Tri-level hypothesis (Marr, 1982) Computational level (most abstract) Which problem is the system trying to solve? e.g. to sort a set of numbers 35 Computation Tri-level hypothesis (Marr, 1982) Computational level (most abstract) Which problem is the system trying to solve? e.g. to sort a set of numbers Algorithmic level How does the system solve this problem? Algorithm? Procedure? e.g. bubble sort, binary tree sort, quicksort 36 Computation Tri-level hypothesis (Marr, 1982) Computational level (most abstract) Which problem is the system trying to solve? e.g. to sort a set of numbers Algorithmic level How does the system solve this problem? Algorithm? Procedure? e.g. bubble sort, binary tree sort, quicksort Implementational level How is this algorithm implemented? Code? Physical? e.g. python code, assembler, logical gates, transistors, electron flow 37 Representations and Computation Classical view (early days of cognitive science) Representations are symbolic Computation in sequential steps Connectionist view (more modern) Representations are activation patterns spread over a neural network (Brain) Computations are parallel in the network From symbolic AI -> Deep Learning 38 Computation Tri-level hypothesis (Marr, 1982) Tri-Level hypothesis can be applied to classical information processors as well as neural networks (Dawson 1998) 39 Structural levels of analysis in the nervous system Brain Brain regions Neural circuits Source: brainline.org Source: operativeneurosurgery.com Source: Grillner et al., 2005 Neurons Synapses Molecules Source: wikipedia.org Source: www3.hhu.de Source: wikipedia.org 40 Metaphor Perhaps the algorithmic level is enough to build a general artificial intelligence It was enough to understand the physical principle behind bird flight! -> not necessary to rebuild wings in all details to build planes https://askabiologist.asu.edu/how-do-birds-fly https://askabiologist.asu.edu/how-do-birds-fly https://www.welt.de/regionales/berlin/article2 107579/Kassen-gegen-zweiten- Rettungshelikopter.html 41  Cognitive Neuroscience for AI Developers The Philosophical Approach The way towards the cognitive revolution -> back to the past 43 Philosophy…. The search for wisdom and knowledge …is the oldest discipline (ancient Greeks) …plays a critical role in cognitive science Source: wikipedia.org "The School of Athens" by Raffaello Sanzio da Urbino Not by producing results (theoretical not experimental), but… 44 Philosophy Defining problems Criticizing models Suggesting areas for future research Source: wikipedia.org "The School of Athens" Free to evaluate other disciplines by Raffaello Sanzio da Urbino Find criteria for intelligence etc. 45 Philosophy Primary method: reasoning Deductive reasoning: applying rules of logic to statements, in order to derive new statements („College students learn three hours a night“. „Mary is a college student“. -> „Mary learns three hours a night“) Inductive reasoning: draw conclusions based on several observations of specific instances of the world („Whiskers the cat has four legs.“ „Scruffy the cat has four legs.“ ->“Cats have four legs“) But the do not use the scientific method (systematic form of induction) 46 Philosophy 2 (out of several) Branches of Philosophy Metaphysics: What is the nature of reality? (First causes of things and the nature of being) -> mind body problem Source: wikipedia.org https://www.britannica.com/topic/metaphysics "The School of Athens" by Raffaello Sanzio da Urbino Epistemology is the study of knowledge What is knowledge? How is knowledge represented in the mind? How do we come to acquire knowledge? 47 The philosophical approach of Cognitive Science Metaphysics Mind-Body-Problem What is mind? Is mind something physical? Is body necessary to have a mind? (primarily metaphysical questions) Source: forbes.com Epistemology How do we come to know things? Are we born knowing certain things or is knowledge learned? How is mental knowledge organized? 48 The Mind-Body Problem: What is mind? How are psychological and mental processes related to physical properties? Brain: material and physical, measurable Mind: subjective conscious experiences Mind as non-physical entity inhabiting the brain „Ghost in the machine“ Source: wikipedia.org Two fundamental questions: René Descarte’s illustration of mind/body dualism. Is the mind physical or something else? What is the causal relationship between mind and brain? 49 Basic conceptions of the nature of mind Mind-Body Problem Functionalism Monism Dualism Substance Property Idealism Materialism Dualism Dualism 50 Basic conceptions of the nature of mind Monism only one kind of state or substance in the universe Aristotle (384-322 BC): mind and body form and matter Analogy: Different shapes of clay are different physical states of brain, no non-physical or spiritual substance 51 Basic conceptions of the nature of mind Monism Idealism: The complete universe is mental e.g. Simulation hypothesis https://de.wikipedia.org/w/index.php?curid=6184755 Materialism (physicalism): All things are made of atoms (Democritus ca. 460-370 BCE) mind is the brain mental states are physical states of the brain 52 Basic conceptions of the nature of mind Dualism Mental and physical substances are possible Plato (427-347 BC): mind and body exist in two separate worlds Mind: ideal world of forms, immaterial, eternal e.g. idea of an ideal circle Body: material world, extended, perishable e.g. concrete, physical circles 53 Basic conceptions of the nature of mind Substance Dualism (Descartes 1596-1650) There exist mental and physical substances Physical substances: world is made of atoms Mental substances: Unknown Theory: Minds can do e.g. pattern recognition. No physical substance can do pattern recognition. -> Minds are not physical. Criticism: 1) How do mental and physical substances interact? 2) Computers can do pattern recognition and much more! Property Dualism Mind and body are of the same substance but have different properties Mental states are non-physical properties of the brain https://en.wikipedia.org/wiki/Property_dualism 54 Basic conceptions of the nature of mind Critics on Dualism Dualism violates the principle of Occams razor: two different worlds that interact are needed. Not the simplest explanation! Another problem: Brain damage changes mental states Computer can do a lot of tasks assumed to be impossible (e.g. ChatGPT can write novels) 55 Conclusion: Philosophy Allows to ask much broader questions than those of other disciplines -> Shows the „bigger picture“ Gives key insights into the relationships between different research areas Plays a very important role in the interdisciplinary endeavor of Cognitive Science Non-empirical approach, in contrast to the scientific method Concepts validated based on logical reasoning and argument Philosophy is better suited to ask questions than to provide answers Close 2-way collaboration between philosophy and science is required 56  Functionalism – Are minds limited to (human) brains? Source: wikipedia.org Source: swx.it Source: wikipedia.org Central Assumption: Cognition is equivalent to computation Computationalism https://plato.stanford.edu/entries/functionalism/  Functionalism – Are minds limited to (human) brains? Source: wikipedia.org Source: swx.it Source: wikipedia.org “Functionalism is the doctrine that what makes something a thought, desire, pain (or any other type of mental state) depends not on its internal constitution, but solely on its function, or the role it plays, in the cognitive system of which it is a part.” https://plato.stanford.edu/entries/functionalism/ “For (an avowedly simplistic) example, a functionalist theory might characterize pain as the state that tends to be caused by bodily injury, to produce the belief that something is wrong with the body and the desire to be out of that state, to produce anxiety, and, in the absence of any stronger, conflicting desires, to cause wincing or moaning.” https://plato.stanford.edu/entries/functionalism/  Functionalism – Are minds limited to (human) brains? Source: wikipedia.org Source: swx.it Source: wikipedia.org Mind could be implemented in any physical system, artificial or natural, capable of supporting the appropriate computations The same mental state could be realized in quite different ways in two separate physical systems. -> Concept of multiple realizability 59  Functionalism – Are minds limited to (human) brains? Source: wikipedia.org Source: swx.it Source: wikipedia.org This could have some ethical implications: If mind can be realized in different systems at which point aliens or computers get human rights What about Large Language Models (LLMs) like ChatGPT? 60 61 Thank you for your attention !