We have come a long way by Statistical methods. Last night, I went into a rabbit hole of Cognitive Architectures like ACT-R[http://act-r.psy.cmu.edu/] and OpenCog[https://wiki.opencog.org/w/The_Open_Cognition_Project]. The intro ofcourse was not as documentation friendly as the width of excellent docs on Deep Learning.Are the Symbolic methods abandoned because they are proven to not work? Also, are any new researchers working on this?