Data Scientist answers Your DEEPEST Questions | Data Science AMA

Just answering some of your questions! Part two to come soon, see ya on the live stream.

Oh, and don't forget to connect with me!

Happy coding!
Nick

P.s. Let me know how you go and drop a comment if you need a hand!
#machinelearning #ai #python

[nicom]

Great vid and thanks for answering my question! It's a relief to know I'm not the only one coming into the Data Science field with an Economics degree haha. Fortunately I had my fair share of maths and stats as you so the transition was smooth. Currently finishing up my DS Masters and your vids were an amazing help to get me closer to some ML domains I wasn't too familiar with. Keep up the good work!

[venomboss]

Heyo Nick, Ive been a viewer of the channel for a while now, following tutorials, learning about different techniques and applications of ML, and creating models from your really awesome tutorials. I have been going through the RL tutorials recently and attempted to do create some models on my own, but after attempt after attempt after attempt, the models don't really seem to be doing much more than the random action models. My question for you is how do I utilize the tutorials in the best way possible? What's the best way to approach applying what I have learned?

Of course, I have lots of other questions, so Ill wait for the next AMA to ask them. Until then, love the content, keep up the awesome, awesome work, and thank you so much for taking the time out of your presumably extremely busy schedule to produce these super helpful videos.

[innocentntuli]

Thanks Nick any possibility of doing a Lip Reading Algorithm for the hearing impaired I feel like that field it’s still in it’s infancy and the only one that has reached a accuracy of 50% if SRAVI developed by Liopa

[prashantsai]

Hey Nicholas, a big thanks for your efforts I am just thinking that if there is any way to make a model that can learn new things without erasing the previous things it has learnt. I am making a object detection model which can recognise 5 objects, so after few days I want to add another 5 new objects to the model, for this I need to train model for all the 10 objects from starting again. But is there any way that we can train the model to recognise these 5 new objects without erasing the learning of previous 5 objects. So that we don't need to train from starting😅

[reviewramble]

Thanks Nicholas for answering my questions. I will apply your advice and use optuna

[soumanabettoboubacar]

You're a nice guy. I like your role.

[rovshansharifli]

Hey Nick, thanks for answering my question:) I mean sure mathematical concepts are important to understand how does the deep learning work. But I wanted to know what math exactly do I apply, lets say if I simply want to build a model to do classification on MNIST dataset?

[innocentntuli]

If someone were to build a CNN model and wanted it to work on mobile or someone were to do a machine learning algorithm how would they deploy it on mobile in such a way that it would have a user front end interface on mobile such as your sign up sign in etc

[rakeshkumarrout]

hey nick i missed you video.Can you help me find out.What ScaleAI does.how do they manage to do all these very fast.Tech behind it.kindly help me find out.Yhank you