We have all heard of the story of the hare and the tortoise.
Once a hare and tortoise lived in a forest. The hare was very proud of his speed. He made fun of the tortoise for his slow speed. The tortoise challenged him to have a race with him. The hare gladly accepted the challenge
The race started. Crow was the referee. The hare ran fast, tortoise was left far behind. The hare stopped to take rest under a tree. He fell asleep.
The tortoise passed him to the winning post. The hare woke up and ran as fast as he could. He saw that the tortoise was already at the winning post. Tortoise won the race.
Moral of the story : “Slow and steady wins the race”
Now, imagine those numerous occasions where this is relevant. The context might be different, in that, it might be applied to our work or studies or general behavior. If you are to guide someone or you want to give some lessons, you would not go on narrating the whole story. You would just mention the moral. Why? Because, narrating the whole story takes time and interpretation is generally left to the listener. Hence, using moral of the stories are better.
AI Models are like moral of the story.
Your data and analytics typically tell a story, an AI model is the moral of that story. AI models will apply the learning of your own data and give you inference. These AI Models can be built using any technique viz. deep learning, traditional ML or even simple rule engines. But, they are the essence of your data.
AI Models are more computationally efficient. A good model would generalize well and pick only those features that are relevant. This gives a significant advantage over the standard ETL and analytics pipeline.
You can have different models cater to specific tasks using the same data. In our hare & tortoise story, “slow and steady wins the race” is the standard moral. But you can also say “never be overconfident and haughty”, “never compare yourself and complain about the situation”. In a similar fashion, your data can give multiple AI models. Application of these models can be based on specific cases.
At Streamoid, we have a no-code platform that can create models from the data available in organizations. The platform is tuned to fashion retail and is completely data driven.