most data science jobs require CQL knowledge to my understanding
like I was able to get a internship at the Federal Reserve as a data analyst with just 3 years done of economics.
upvote
it’s basically this, but instead of fnagling with protobuf’s that aren’t suited to describe data with rules, constraints, models, whatever, you use something that is: context-free grammars (you know, like an actual programming language).
there’s a mutational engine that mutates the data as it’s represented as an AST; i.e you’d swap nodes, change parent/child, add/delete nodes.
the idea of having a fuzzer that mutates an AST or some tree-like data structure instead of viewing everything as a bytestring to be bitflipped excites me greatly.
reader’s note: this is an original idea, do not steal.
I'm doing a NLP project for medical records, trying to take an existing github project that used an attention model called CAML and apply BERT onto it so it becomes bidirectional
Here's an article explaining it.
i like the cut of your jib
wish I was in a phd program
will compose my thoughts to PDF and notify this thread in the next few days.
I’ve made more money with music than I made with my Computer Science degree.
update: it’s already been done
“The smaller the screen, the smaller the artifacts. I tried on my smartphone”
- old guy on digital foundry shilling for the new google games-in-the-cloud service (aka dead-on-arrival downgrade technology that nobody will ever use, just like every other google product)
+frag
this is pretty much all of my good ideas
Interesting thread.
x_x
im just late to the party
just going to have to ramp it up to inferring a grammar structure from a set of binary inputs