Keep in Touch with Reality
- Use Human-like components (simulated neurons)
- Keep in touch with reality by
- Solve problems like humans do (cognitive models and architecture)
- Develop interesting systems (AI programs)
- Integrate Learning
- Take advantage of emergent algorithms
- Repeat
- The problem space is enormous and we need ways of
focusing.
- We've already back-tracked on neural reality because
it is really expensive to simulate.
- There are a number of ways to focus, including:
- computational limits (Turing completeness),
- neural properties (e.g. topology and behaviour), and
- solving known problems (e.g. stability-plasticity and
symbol grounding).
- I think Turing completeness is done, and we should
push towards neural properties and think
- developing cognitive models and AI systems will solve (or at
least address) these known problems.
