Agree on the pic, but when I saw it I was thinking would some deep learning analysis add any value at all? Maybe just for academic or conspiracist reasons. Like say the analysis determines that there was a 50% chance that the person in the pic is Earhart, but also 50% chance that the person in the pic was John Lennon. I know the analysis would catch an incongruous time period, but just sayin'...
>would some deep learning analysis add any value at all?
"Machine learning is like a deep-fat fryer. If you’ve never deep-fried something before, you think to yourself: 'This is amazing! I bet this would work on anything!'”
Not to veer too far off-topic - but what does deep-frying not work on? (ice cream, veggies, sweets, sea-food, beef/any meats... all delicious when deep-fried). Maybe fruit is the one exception? I have never tried but may give it a go. It seems like a poor comparison, as almost all foods are delectable when deep-fried. I know very little about machine learning though, I will concede. I also cannot open the link (blocked at work).
I say this mostly for the humor, but also for the truth: deep-frying is the best. And I deep-fry food quite often. In fact, I am headed home to deep-fry some Won-Tons soon.
You've probably had deep-fried bananas. I've had deep-fried watermelon. Like most deep-fried foods, neither is particularly memorable except for the kitsch, even if they do taste good.
Which I think supports the original point: you can deep-fry anything and enjoy tasty results, but that _still_ doesn't mean there's anything substantial in the sense of untapped value.
Food is highly sensitive to cultural and social context. Even with deep-fried fruits, the part that stuck with you the most was the "kitsch" rather than the flavor, texture, aroma, size, cost, or any other objective fact about the food item.
My observation offers no lessons on deep learning image recognition, but as long as we are already on a tangent...
Deep-fried battered pineapple rings were a staple of trips to the local chip shop when I was a child. Drizzled with syrup they make an excellent dessert when you've been running around outside all day.
Not what you'd call delicate or refined, though.
Maybe if you used tempura batter instead of chip shop batter.
I think the connection is the overuse/desire to apply deep-frying/machine learning constantly and to everything. While there are certainly (delicious) benefits, there are also drawbacks that make the non-stop use a net loss.
I had a similar thought as I typed - having more than one or two deep-fried dishes a week is unappealing, and having too many deep-fried courses together is unappealing (at least to me). And then potential health issues. When I get home, and after eating my fried won-tons, I'll check out the link.
