automenta

wow i forgot all about this. havent touched it since. does anyone actually use moa?

thanks yes it seems worth more work. hopefully in the next day or so ill get the latest code and apply those gui changes to it including the other panels.

i updated the gui changes to the latest MOA code, and converted the first two panels (classification and regression) like the original classification panel changes. please let me know if...

anatomy of a configuration headache i ended up putting this overriding class in com.google.common.flogger.backend.system public final class SimpleBackendFactory extends BackendFactory { private static final BackendFactory INSTANCE = Slf4jBackendFactory.getInstance(); public static...

Narsese_AtomicTermIndex can only support a max # (ex: ~2^16 for the unsigned short) of unique "atoms" in the runtime of a NAR. you can argue that this is unlikely to...

basically: a "radix trie bag" which is a radix trie with priority accumulators for each leaf link in a node. these would get added and subtracted when leaves in the...

how were negated subterms supposed to work if compounds can only be 1-level deep? :skull: probably an immutable Term*[] array of pointers to subterms, like in Java, is inevitable. none...

``` typedef struct { bool hashed; HASH_TYPE hash; Atom atoms[COMPOUND_TERM_SIZE_MAX]; }Term; ``` So if this is supposed to be a Compound, and atoms[] are the subterms, how can a Compound...

i knew there must be some encoding that wasn't immediately obvious from the code. maybe add some comments to the Term structure explaining this 'atoms' array includes (non-atom) operators, and...

lets see some benchmarks to know if the concurrency support actually does anything, including if they make performance worse or better and how much on different #s of threads (1,...