You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
The number one thing to keep in mind about machine learning is that performance is evaluated on samples from one dataset, but the model is used in production on samples that may not necessarily follow the same characteristics...
The finance industry has a saying for this: "past performance is no guarantee of future results". Your model scoring X on your test dataset doesn't mean it will perform at level X on the next N situations it encounters in the real world. The future may not be like the past.
So when asking the question, "would you rather use a model that was evaluated as 90% accurate, or a human that was evaluated as 80% accurate", the answer depends on whether your data is typical per the evaluation process. Humans are adaptable, models are not.
If significant uncertainty is involved, go with the human. They may have inferior pattern recognition capabilities (versus models trained on enormous amounts of data), but they understand what they do, they can reason about it, and they can improvise when faced with novelty
If every possible situation is known and you want to prioritize scalability and cost-reduction, go with the model. Models exist to encode and operationalize human cognition in well-understood situations.
("well understood" meaning either that it can be explicitly described by a programmer, or that you can amass a dataset that densely samples the distribution of possible situations -- which must be static)
My previously trained wetware learned that Owls sometimes have visible feet from this single new image:https://twitter.com/cassmarketos/status/1229473344480673792?s=20 …
Wait, are there people that don't know this? Fuzzing tests, input constraints etc.. all remain as important for production inference systems as they are for any piece of software. I have over the years recognized a lack of basic product competence for a lot of ML engineers
It's elementary, but many deep learning researchers are not aware of this, because in 100% of the work they've ever done, the model's score on the test set was the finality. Most DL research starts from an existing dataset and stops at a test set score.
This is so true, I see people saying there dl models are used in production but there is no way to compare how good these models are in production, this is also one reason finance doesnt use deep models in production.
Just finished a capstone project on patient similarity driven predictive modeling of clinical data, 22 papers with excellent model performance, but when I looked for a real world implementation...*crickets*
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
