Hi Ben, I'm trying to spend just enough time to transmit all of the key design principles and constraints to Alexey and his team - (and Anton and his team) - as they are first-in-line to then diseminate this knowledge to a wider group. Transmitting this information accurately is very key, very important at this point in time - mistakes here will be amplified over time.
On Mon, May 21, 2018 at 9:09 PM, Ben Goertzel <[email protected]> wrote: > > > I think it's better, if possible, to figure out a way to suitably > modify the core PM rather than > using a separate repo ... > Nothing I've heard so far requires any modification at all to the PM. > > However, I guess the PM tweaks would need to be done someone on your team, > as > Linas and Nil probably are too busy and we don't have a lot of others > who can rapidly > perform such changes... > Yes. Keeping up with email is almost full-time. > > I would personally be in favor of overloading stuff like TimesLink in > order to apply to both > NumberNodes and Values, because it seems to me that the Atom/Value > distinction is more of > an efficiency-driven implementation distinction rather than a > fundamental mathematical/conceptual > distinction... > :-/ Its both efficiency and fundamental. Square pegs and round holes. You cannot put a Value in the outgoing set of a Link. That means that you cannot overload a TimesLink. There are lots of other really neat things you can do, however. > > > Also, DNNs are trained on (mini-)batches. It is not too natural from an > > autonomous agent perspective, but efficient. > > Yes I see. Again maybe some new TensorValue construct will be needed, we > just > need to understand clearly what the requirements are in terms of any > special > indexing etc. > At this point, the natural progression would be to start writing some over-all design proposal. What's the input, what's the output, what's connected to what, what data is being generated, what data is being consumed. This will make it much more clear exactly what code needs to be written where. > > The difference between Atoms and Values is just an > implementation-efficiency > tactic... > Um, no, not really, They are really fundamentally very different beasts. This is key, and trying to blur the distinction is going to lead to trouble. Trust me, for now; if you want to come back to this in 1/2 a year, once all of the various team-members are up-to-speed, then we can have a rational discussion. But if we have that discussion now, it will crash and burn. -- Linas -- cassette tapes - analog TV - film cameras - you -- You received this message because you are subscribed to the Google Groups "opencog" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/opencog. To view this discussion on the web visit https://groups.google.com/d/msgid/opencog/CAHrUA36m1YG-4wOTn9qQKoHnS%3DtAk7M%2BizW35Ya_8kf%3Du9shkQ%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
