Direction advice

Hi I am new to the AI community. I am in the healthcare field and I am interested in using AI as a research tool. I would like to create a neural network that would allow for evaluation of image clips. The purpose of this would be to 1) validate that this would work 2) use the model to answer clinical questions and 3) predict outcomes. I am obviously a beginner so I am wondering if this is possible. Also, would appreciate any direction that anyone out there can provide. Thx!

1 Like

Greetings Kloh1015,

I’m very interested in working in the healthcare field.

There are a few resources I can recommend, as far as courses:

  1. TensorFlow Developer - This course will teach you how to use TensorFlow open source library for deep learning. There are 3 sections for different types of input data, namely images, text and sequence models for time series forecasting.

I took the course and really enjoyed it. The instructor is great and the slides and notebooks are excellent. The only blocker I see is the code requires intermediate level skills for Python. I had about 5 months of experience with machine learning prior to taking the course. If you want to work on foundational skills first, you can take the machine learning specialization and then move to tensorflow after.

  1. AI for Medicine - I haven’t taken the course yet, but I viewed the curriculum and there are many interesting use cases to measure outcomes for images, text and tabular data. I would love to take this course at some point in the future.

Good luck and I hope you enjoy your pursuit of AI knowledge!


1 Like

Thank you so much for your advice! Really appreciate it. I think I really need to work on my foundation in Python since it seems like that would be integral for most platforms. Thx again!

1 Like

I use the GitHub copilot with VS code. The AI will help you find code fragments very fast. You still need to understand the code, but the AI just speeds things up with the learning process.



A couple of additional thoughts:

  • In general, avoid using VSCode for the programming assignments (at least for MLS and DLS). If you download the notebooks and work on them locally in VSCode (or Colab as well), and then upload them back to Coursera for grading, it can cause problems with the notebook metadata.

  • Avoid using coding tools like Copilot. One of the important tasks here is to learn how to code in Python. Subcontracting the coding to an AI tool is not going to give you the same level of comprehension.


“Avoid using coding tools like Copilot”

I think your view os fair. However, as you may view some tasks as ‘subcontracting’, what about the benefits of paired programming?

If you’re by yourself and you’re stuck, a code snippet my unblock an issue and save you significant time and stress. You can intuitively understand the logic and reuse it again in the future.

While code comprehension in general is important, there are certainly some instances where it’s not.

1 Like

There’s a difference between real-world work and taking a course.

The course Code of Conduct says you can only submit your own work.