Brain-Mind Relationship through Computational Cognitive Neuroscience
Explore how the brain creates the mind by leveraging computational cognitive neuroscience models that run on computers. Discover the unique features of human computation, connectionist models, and the Interactive Activation Model, shedding light on the underlying processes of thinking and perception.
Download Presentation
Please find below an Image/Link to download the presentation.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author. If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.
You are allowed to download the files provided on this website for personal or commercial use, subject to the condition that they are used lawfully. All files are the property of their respective owners.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author.
E N D
Presentation Transcript
How does the Brain make the Mind? Gary Cottrell Computer Science and Engineering Department Institute for Neural Computation Temporal Dynamics of Learning Center UCSD CSE 87 Lecture 2 3/20/2025 1
The main idea of last time One way to try to understand how this thing works is to build a workingmodel of it - using data from the other approaches. Nowadays these models can be run on computers - so that we can actually see them work. This is the role of cognitive science - or to put it more specifically: computational cognitive neuroscience 2 3/20/2025 CSE 87 Lecture 2
Last Time Why are people smarter than machines? - different hardware (or "wetware") lots of slow, dumb, noisy components that cooperate to solve problems: a massively parallel machine What are the outstanding features of human computation? -constraint satisfaction -noise tolerance -interaction of many types of knowledge What are the features of connectionist models? -simple processing units -connected via weighted links-spread activation and inhibition - relax to a solution 3 3/20/2025 CSE 87 Lecture 2
Last Time: The three axioms of cognitive science (a la Cottrell) 1. The mind is what the brain does There is no spooky stuff What the brain does, i.e., thinking, is a kind of computation Love at first sight is a computation The kind of computation the brain does is probabilistic - probability is the language of thought It must be, in order to deal with the uncertain nature of the world Neural networks are a good model of this kind of computation CSE 87 Lecture 2 2. 3. 4 3/20/2025
Your first neural net: The Interactive Activation Model: A model of reading from print Word level Letter level Feature level 5 3/20/2025 CSE 87 Lecture 2
The Interactive Activation Model This model demonstrated: context effects: top-down activation helps perception Rule-like behavior: acts as if it knows the orthographic rules of English (pseudoword effect) Noise tolerance: cleans up noisy inputs even for novel inputs! 6 3/20/2025 CSE 87 Lecture 2
Todays first simulation: The Necker Cube But first: Computer setup 1. Log in to windows (not Unix!) using your UCSD email as your login and your email password. 2. Open a browser window 3. Google Gary Cottrell and click on the first link. 4. Go down to the bottom of the page and click on the Courses link 5. Click on the CSE 87 link. Follow the instructions at the bottom, starting with step 3. 7 3/20/2025 CSE 87 Lecture 2
The Necker Cube 8 3/20/2025 CSE 87 Lecture 2
The left cube is the down-left-facing one The right cube is the up-right-facing one Each vertex is labeled by its value a hypothesisabout that vertex. E.g., FLL represents the hypothesis that this particular input is the front lower left vertex. The units are wired up according to constraints between the hypotheses. 3/20/2025 CSE 87 Lecture 2 9
The left cube is the down-left-facing one The right cube is the up-right-facing one Each vertex is labeled by its value a hypothesisabout that vertex. E.g., FLL represents the hypothesis that this particular input is the front lower left vertex. The units are wired up according to constraints between the hypotheses. 3/20/2025 CSE 87 Lecture 2 10
The units are wired up according to constraints between the hypotheses. 1. If FLL is on, then FLR should be on: Put in a positive link between them. CSE 87 Lecture 2 3/20/2025 11
The units are wired up according to constraints between the hypotheses. 1. If FLL is on, then FLR should be on: Put in a positive link between them 2. There can only be one FLL: Put a negative link between them CSE 87 Lecture 2 3/20/2025 12
The units are wired up according to constraints between the hypotheses. 1. If FLL is on, then FLR should be on: Put in a positive link between them 2. There can only be one FLL: Put a negative link between them 3. There can only be one interpretation of any vertex: Put a negative link between FLL and BLL CSE 87 Lecture 2 3/20/2025 13
CSE 87 Lecture 2 3/20/2025 14
CSE 87 Lecture 2 3/20/2025 15
CSE 87 Lecture 2 3/20/2025 16
Summary The Necker Cube simulation showed: 1. How a network can be hand-constructed by: Following the unit/value principle Wiring up the units according to the constraints 1. 2. 2. How spreading activation and inhibition can make a decision : The two networks competed to win via the negative links between them 17 3/20/2025 CSE 87 Lecture 2
The Jets & The Sharks Humans have a content-addressable memory I m thinking of someone. They are: a former actor Intelligent a former president old You probably retrieved the right person, even though part of the description is incorrect! 18 3/20/2025 CSE 87 Lecture 2
The Jets & The Sharks Computers access memories differently: To get an item out of memory, you need: an address (where it is in memory) the CPU gets the contents of that address puts it in a working memory (register) does something with (adds it to something else) 19 3/20/2025 CSE 87 Lecture 2
The Jets & The Sharks A memory model The Domain: Memory of the properties of two gangs The Model: Basically, the interactive activation model applied to memory 20 3/20/2025 CSE 87 Lecture 2
The Model 21 3/20/2025 CSE 87 Lecture 2
The Data 22 3/20/2025 CSE 87 Lecture 2
To run the model Open Matlab Navigate to pdptool/iac Type jets in the command window and hit <return> or <enter> 23 3/20/2025 CSE 87 Lecture 2
24 3/20/2025 CSE 87 Lecture 2
The properties of the J&S Model 1. Content addressable: - We can retrieve someone s name from their properties & vice-versa 2. Graceful degradation Works even with some misleading information: "best match search" 3. Default assignment Missing features can be "filled in" 4. Spontaneous generalization Even though only specific properties of individuals are encoded, we see prototype formation 25 3/20/2025 CSE 87 Lecture 2
END!! Any questions? 26 3/20/2025 CSE 87 Lecture 2
