Artificial General Intelligence
Intro
What's AGI (or some sense of "general" intelligence)?
GoodAI
- https://arxiv.org/pdf/1912.01513.pdf
Pieter Abbeel
- Transformers as Universal Computing Engines: https://arxiv.org/abs/2103.05247
Andrey Karpathy
- https://twitter.com/karpathy/status/1472287874003333122
Pei Wang
- NARS: https://www.applied-nars.com/
Eric Jang
- Just ask for generalization from simple tasks supervised sponges, maybe in an offline RL model
- https://evjang.com/2021/10/23/generalization.html
- https://evjang.com/2021/12/17/lang-generalization.html
Jeff Dean
- Not totally on AGI, but idea of large multi-task and selective sparse-NNs are really a component
- PathWays: https://blog.google/technology/ai/introducing-pathways-next-generation-ai-architecture/
- PathNet: https://arxiv.org/abs/1701.08734
- TF https://github.com/jsikyoon/pathnet
- PT https://github.com/kimhc6028/pathnet-pytorch
- Extended PathNet: https://www.nature.com/articles/s41598-020-64165-3
George Hotz
- MuZero?
- https://github.com/werner-duvaud/muzero-general
- Solution for self driving?
Eliezer Yudkowsky
- https://intelligence.org/
Hutter
Luciano
https://www.researchgate.net/publication/334477701_Creativity_and_Complexity_CDT-12
David Deutsch
Any ability based on guessing must include means of correcting one’s guesses,
since
most guesses will be wrong at first. (There are always many more ways of being
wrong than
right.) Bayesian updating is inadequate, because it cannot generate novel
guesses about the
purpose of an action, only fine-tune—or, at best, choose among—existing ones.
Creativity is
needed. As the philosopher Karl Popper explained, creative criticism,
interleaved with creative
conjecture, is how humans learn one another’s behaviors, including language,
and extract
meaning from one another’s utterances.2 Those are also the processes by which
all new
knowledge is created: They are how we innovate, make progress, and create
abstract
Source: http://www.daviddeutsch.org.uk/wp-content/uploads/2019/07/PossibleMinds_Deutsch.pdf
Eric Beaulieu
Carl's contact. Modeled flies' visual cortex. Has a 10 years plan to AGI: http://www.e2027.com/
Josh Tenenbaum
Interesting talk: https://www.youtube.com/watch?v=7ROelYvo8f0&index=2&t=4130s&list=PLrAXtmErZgOdP_8GztsuKi9nrraNbKKp4
There are no real AI, there's only AI technologies
- None of those AIs have counsciusness, nor can learn from simple examples
What else we need?
- Intelligence is not just about pattern recognition
- It's about modeling the world
- Explaining and understanding what we see
- Imagining things we could see but haven't yet
- Problem solving and planning actions to make these things real
- Building new models as we learn more about the world
- Sharing our models, learning from others models
- https://arxiv.org/abs/1604.00289
Refs
- https://arxiv.org/pdf/1805.01109.pdf