ChatGPT4 on the cat on the mat – Skywritings
Symbol Grounding: This refers to the concept of how words (symbols) get their meanings. Harnad introduced the term in the paper “The Symbol Grounding Problem,” in which he argues that symbols’ meanings are grounded in sensory experience. For example, the word “apple” has meaning because it’s linked to our sensory experiences of seeing, touching, and tasting apples. Category-learning: This relates to the process by which we learn to classify objects, experiences, or concepts into different categories. Harnad and his colleagues have conducted extensive research on this topic, particularly in the context of how human and artificial systems can learn to categorize. It relates closely to symbol grounding as categorization allows for the meaningful grouping of experiences which can then be symbolized. Categorical Perception (CP): This term refers to the phenomenon where the categories possessed by a perceiver influence the perception of stimuli. For instance, humans perceive colors as distin
SH: First, let’s set up a local context for this chat. Could you please tell me what you know about the following topics: 1. Symbol Grounding 2. Category-learning 3. Categorical perception 4. Language Origins and Evolution 5. Minimal Grounding Sets (Minimal Feedback Vertex States of Dictionary Graphs) ChatGPT: Certainly, let’s go over these topics one by one: Symbol Grounding: This refers to the concept of how words (symbols) get their meanings. Harnad introduced the term in the paper “The Symbol Grounding Problem,” in which he argues that symbols’ meanings are grounded
Explore this link on the map →related reading
- Vector Grounding? – Skywritingsgeneric.wordpress.soton.ac.uk
- Chatting with ChatGPT-4 About Symbol Grounding, Turing Testing, Zombies, Uncomplementable Categories, and Why the Hard Problem is Hard – Skywritingsgeneric.wordpress.soton.ac.uk
- What Is ChatGPT Doing … and Why Does It Work?-Stephen Wolfram Writingswritings.stephenwolfram.com
- Meaningful | Slate Star Codexslatestarcodex.com
- A global workspace in language models \ Anthropicanthropic.com
- Verbalizable Representations Form a Global Workspace in Language Modelstransformer-circuits.pub
- [2304.01481] The Vector Grounding Problemarxiv.org
- Ludwig.Wittgenstein.-.Philosophical.Investigations.pdfstatic1.squarespace.com
- When Models Manipulate Manifolds: The Geometry of a Counting Tasktransformer-circuits.pub
- Non-Human Words: On GPT-3 as a Philosophical Laboratory | American Academy of Arts and Sciencesamacad.org
- How An Algorithm Feels From Inside — LessWronglesswrong.com
- 2023 letter | Zhengdongzhengdongwang.com