1/13
Vocabulary flashcards covering cognitive science definitions, critiques of education and anthropology, artificial neural network mechanics, generative AI, diffusion, and ethical/social concerns.
Name | Mastery | Learn | Test | Matching | Spaced | Call with Kai | Chat |
|---|
No analytics yet
Send a link to your students to track their progress
Cognitive Science
The study of minds and thinking, especially at the information processing level, defined methodologically as applying methodologies from multiple disciplines to multiple problems from those disciplines.
Cognitive Anthropology
A branch of cultural anthropology that studies how different cultures think, focusing on social organization, enculturation, cultural change and transmission, shared knowledge, distributed cognition, and situated cognition through qualitative methods without hypotheses.
Splitters
Researchers, such as anthropologies, who pick things apart by focusing on differences, as opposed to lumpers.
Lumpers
Cognitive scientists and researchers who put things together rather than picking them apart based on differences.
Critiques of Education Research
Criticisms noting that case studies are close to worthless and too applied to reveal basic cognitive processes, controlled studies are poorly done despite being expensive and hard, and research only addresses one part of cognition (learning).
Units
Variables within an artificial neural network layer that are said to fire like neurons whenever they hold a number higher than a given threshold.
Hidden Layer
A layer in an artificial neural network where each unit listens to specific input units, multiplies their activation by link weights, and fires if its activation reaches a threshold.
Supervised Learning
A process where neural network weights start randomly and, over 1000+ trials, weights that contributed most to wrong guesses are weakened so the network improves.
Diffusion
An image generation method where AI turns images into noise during training and remembers pixel changes, then generates new images by applying the reverse of those changes to random pixels when given a user label.
Prompt Engineering
A job task arising in a shrinking art market that involves writing text prompts and using software like Photoshop to edit AI-generated images.
Behind-the-Scenes Diversity Prompting
A technique where AI systems insert additional words into a user's prompt behind the scenes to generate diversity in outputs, which can also surface dataset biases.
Deepfakes
Fake videos, audio, sexual content, or false photos made about someone, creating legal and historical challenges and enabling claims that authentic media is fake.
LAION Dataset
A dataset of 5.6 billion images scraped from the internet without artist permission or compensation, used to train AI models.
Plagiarism Machines
A critique describing generative AIs trained on scraped art without consent, countered by arguments that humans also create art by studying and imitating other artists.