I'm one of those people who, at any one time, is always dabbling in about 100 things. But as far as what is getting the bulk of my attention the past few weeks:
1. OpenCog[1]
2. SOAR[2][3]
3. Computational Linguistics[4][5]
why?
The unifying element to most of the things I'm learning / researching these days is my interest in AGI. The interest in linguistics is driven by a belief that I hold, that says that language and cognition are tightly coupled in interesting ways, to the point that it makes sense to work on language and other aspects of cognition "hand in hand".
Not to suggest that I think that use of natural language is required for (some forms of) intelligence... but rather I just think language (in terms of internal dialogue / sub-vocalization) is used for some really important parts of human intelligence... Couple that with my notion that - if you want to "exercise" an (purportedly) artificially intelligent system, it makes sense to do it by attempting to converse with it - and you can see how it all leads me to a place where working on language understanding in parallel with other aspects of intelligence makes a lot of sense.
I make no real claim that this is a novel insight, that it is in fact an "insight" at all, or even that any of this is actually correct. It's just where I personally am at with things these days.
1. OpenCog[1]
2. SOAR[2][3]
3. Computational Linguistics[4][5]
why?
The unifying element to most of the things I'm learning / researching these days is my interest in AGI. The interest in linguistics is driven by a belief that I hold, that says that language and cognition are tightly coupled in interesting ways, to the point that it makes sense to work on language and other aspects of cognition "hand in hand".
Not to suggest that I think that use of natural language is required for (some forms of) intelligence... but rather I just think language (in terms of internal dialogue / sub-vocalization) is used for some really important parts of human intelligence... Couple that with my notion that - if you want to "exercise" an (purportedly) artificially intelligent system, it makes sense to do it by attempting to converse with it - and you can see how it all leads me to a place where working on language understanding in parallel with other aspects of intelligence makes a lot of sense.
I make no real claim that this is a novel insight, that it is in fact an "insight" at all, or even that any of this is actually correct. It's just where I personally am at with things these days.
[1]: https://opencog.org/
[2]: https://soar.eecs.umich.edu/
[3]: https://en.wikipedia.org/wiki/Soar_(cognitive_architecture)
[4]: https://en.wikipedia.org/wiki/Computational_linguistics
[5]: https://www.amazon.com/Foundations-Computational-Linguistics...