Analytic Philosophy After Dark: The Blank Slate and Deep Learning
Join us at one of our favorite bars to casually discuss the connection between debates over the blank slate and deep learning.
How blank of a slate is the human mind? The “blank slate” is a catch-all term often used to describe the radical empiricist claim that our minds are born with no innate concepts. Most contemporary cognitive scientists reject so radical a proposal, but that raises the question: how many concepts are innate to the human mind and does the power of AI models trained on algorithms without many hard-coded rules shed any light on the possible answer?
Possible discussion questions
Do advancements in deep learning provide evidence that less innate concepts are needed for a theory of thought to work?
How many concepts do you think have to be innate for a system to be able to think as capably as a human in some domains?
Empiricists in the philosophy of mind view the learning of concepts as a type of hypothesis testing. One challenge for this view is that if we view the learning of concepts as hypothesis testing how can we learn a new concept without that concept being contained in the hypothesis itself?
Will future advanced AI systems use different deep learning algorithms (i.e. transformers, CNNs, etc.) as different parts of the architecture of higher cognition, similar to the idea of different mental faculties?
Optional content
https://philpapers.org/rec/FODTPS
https://drive.google.com/file/d/1zPMnIuwr_Gcj48rFpEakI5SpF9Z7xrga/view?usp=sharing
This is pretty long. Feel free to read none or just whatever parts seem interesting to you.
Subscribe to our Luma calendar: https://lu.ma/apad