End-to-end deep learning solutions let you focus on what matters: the data itself, its annotations, the metrics you're optimizing for. Less tinkering, more problem-solving.
Show this thread
Deep learning, in many cases, enables engineers to replace complicated, heuristics-laden machine learning pipelines with a simple end-to-end model (plus a data collection & annotation process, typically easy to maintain).https://twitter.com/tymwol/status/1030163380022267904 …
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I like deep learning. But are deep neural network architectures, augmentation and optimization tricks not heuristics-laden?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Though, in reality, DL often involves different kinds of hidden complexity, including certain brittleness and fine tuning of the models specific to the training data.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
What are some the heuristic-laden in ML that is not in DL? Feature selections?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Agree, but "shallow" machine learning does the same, usually in less complex manner. Deep learning offers automatic feature engineering, but at cost of tuning the architecture instead. Maybe auto tuning would solve this, but at computational cost & time.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Do have a look at
@Labellerr1 very relevant to AI process innovation to optimize for accuracy, speed and costs!Thanks. Twitter will use this to make your timeline better. UndoUndo
-
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.