Synthetic Data — A Novel Way to Augment Reality

The success of deep learning has also brought with it an insatiable hunger for data. Instead of solely collecting data from the real world, we can also generate it programmatically.

Pascal Marco Caversaccio
DAITA Technologies
Published in
7 min readMar 10, 2021

--

Original source

Let me get straight to the point:

“Machine learning (ML) can only be as good as the data you use to train it.”

But wait, why is this actually the case? Well, simply put, when you deal with ML algorithms, you need particular inputs to help your model understand things in its own way (yes, your model is extremely stupid at the very beginning; and many times later, too!). And training data is the only source you can draw on as input to your algorithms (as an example, you can think of pictures or video material as training data). Blatantly, they help your ML model extract useful information from the data and take some important decisions, just like human intelligence does. So far so good? If yes, let us move on to a specific class of algorithms used by ML methods.

Supervised ML requires additional input from labelled / annotated training data. And if your training data is not properly labelled, it is not suitable for supervised ML. The data…

--

--

Pascal Marco Caversaccio
DAITA Technologies

𝐖𝐨𝐫𝐤𝐢𝐧𝐠 𝐨𝐧 𝐰𝐡𝐚𝐭’𝐬 𝐧𝐞𝐱𝐭.