The Coca-Cola Company using TensorFlow for digital marketing campaigns (TensorFlow Meets) digital marketing campaign

by adminhuahinsunvilla3sd



Ever wonder how The Coca-Cola Company uses machine learning?

In this episode of TensorFlow Meets, Laurence sits down with Patrick Brandt, who is part of the Marketing Operations team at Coca-Cola North America. Learn about how TensorFlow was used to run one of Coca-Cola’s biggest digital marketing campaigns, which involved image classification using deep neural networks for coca-cola bottle caps. Check our Patrick’s talk at TensorFlow Dev Summit ‘18 and subscribe to the channel to stay up to date with the latest in ML and AI.

TensorFlow Dev Summit ’18 Coca-Cola talk →
Applied AI at Coca-Cola blog article →

TensorFlow Meets playlist →
Subscribe to the TensorFlow channel → .

Images related to the topic digital marketing campaign

The Coca-Cola Company using TensorFlow for digital marketing campaigns (TensorFlow Meets)

The Coca-Cola Company using TensorFlow for digital marketing campaigns (TensorFlow Meets)

Search related to the topic The Coca-Cola Company using TensorFlow for digital marketing campaigns (TensorFlow Meets)

#CocaCola #Company #TensorFlow #digital #marketing #campaigns #TensorFlow #Meets
The Coca-Cola Company using TensorFlow for digital marketing campaigns (TensorFlow Meets)
digital marketing campaign
You can see a lot of useful information here: See more here
You can see a lot of useful information here: See more here

Related Posts

9 comments

Joe Wong 27/02/2022 - 5:47 Chiều

Why does Coke or any company for that matter still use dot-matrix printers anymore? your just asking for trouble with the unclear old school fonts???

Reply
aitube.io 27/02/2022 - 5:47 Chiều

Cool!

Reply
William Dye 27/02/2022 - 5:47 Chiều

How did they generate the synthetic data for the initial training set (4:21)? I'm surprised that they only achieved 50% accuracy before moving on to noisy real-world data. In my current project I'm deliberately overfitting (100%) to the synthetic initial training data, then tweaking the data/model/code to reduce overfitting, and only then moving on to real-world data. Yes, yes, I realize that overfitting is bad, I'm just saying that it seems easier to start with an overfit model than start with a model that can only hit 50%. Am I wrong?

Reply
Aron Einhorn 27/02/2022 - 5:47 Chiều

Wondering who this one dislike is on this video.
Must be the CEO of Pepsi himself. 😉

Reply
Paolo Leopardi 27/02/2022 - 5:47 Chiều

Hi Laurence, regarding mobile apps: any plans for official support of TensorFlow Lite on Flutter?

Reply
Aron Einhorn 27/02/2022 - 5:47 Chiều

Very interesting straightforward use.
Patric is a cool dude.
BTW, Laurence, do you ever get to sip from those coffee mugs? It seems to always be empty…

Reply
Saj Cloud 27/02/2022 - 5:47 Chiều

Interesting use!

Reply
shubham srivastava 27/02/2022 - 5:47 Chiều

Great…happy to see that

Reply
Simo's mood 27/02/2022 - 5:47 Chiều

thank you so much for your time

Reply

Leave a Comment