To be fair, I don't think anyone sensible was claiming we're anywhere near solving NLU with DL (or any other method). It's just an effective tool for very simple/constrained NLU problems.
an another important wake up call for AI & deep learning: fieldwide “spectacular fail” on important new natural language understanding benchmarks, developed by @sleepinyourhat and others https://gluebenchmark.com/
-
-
-
Also, we haven't failed yet! I doubt anything based *completely* on existing methods will get to human performance on the benchmark, but there hasn't been a serious effort yet, just our own deliberately-simple baselines.
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.