I recently wanted to do a deeper dive into some of the AI areas that we use in our technology stack. Specifically we do a number of things with machine learning, natural language processing, parsing, and use of chatbots and voice assistants. The first three items are more on the validation and processing side where the last two are more on the output/consumption side.
I spoke with a good friend who has spent a lot of time working on ML/AI at FB and he recommended checking out some high level coursework on Coursera and some other services. I have to admit, I was pretty skeptical. I can't recall the last time I had set out to do something of an academic nature.
Throughout the various companies I have been involved at I would routinely go deep on specific books or articles around particular topics that were relevant to work. In looking at trying to understand some of the deeper AI topics, I would have to do some fundamental learning of Python as well as some of the overall topical information around AI, Machine Learning and Deep Learning. While I knew specifically what I wanted to explore, I would have to lay some foundations to grasp the potential to apply some of my theories to these toolsets.
I started out doing some overview classes on Coursera by Andrew Ng and at the same time started taking an introductory class on Python. Both of the classes were well done and I found myself quickly going through them as I have had some long time ago experience programming and the topical stuff from Andrew Ng was very well presented.
These classes only left me wanting to go deeper and I am now working through some coursework using various tools to data mine Twitter and doing analysis as well as doing an overview of TensorFlow through a Udacity Course.
What has amazed me in the few weeks I have been grinding on this is that there are some really powerful resources available to take you very deep in a very short period of time. In addition to the high quality video lectures, there are Jupyter Notebooks with sample code you can modify and various Python scripts that can get you going very quickly. I actually now find myself at the point where I am trying to not figure out if what I want to do is possible but rather what is the best tool or tools for me to test my theories out as fast as possible.
I started out very skeptical about the value of taking the "online education" path to gain new knowledge and I have to say that right now I am pretty blown away by what is available. Stay tuned for more as I go a little deeper in the coming weeks.
I spoke with a good friend who has spent a lot of time working on ML/AI at FB and he recommended checking out some high level coursework on Coursera and some other services. I have to admit, I was pretty skeptical. I can't recall the last time I had set out to do something of an academic nature.
Throughout the various companies I have been involved at I would routinely go deep on specific books or articles around particular topics that were relevant to work. In looking at trying to understand some of the deeper AI topics, I would have to do some fundamental learning of Python as well as some of the overall topical information around AI, Machine Learning and Deep Learning. While I knew specifically what I wanted to explore, I would have to lay some foundations to grasp the potential to apply some of my theories to these toolsets.
I started out doing some overview classes on Coursera by Andrew Ng and at the same time started taking an introductory class on Python. Both of the classes were well done and I found myself quickly going through them as I have had some long time ago experience programming and the topical stuff from Andrew Ng was very well presented.
These classes only left me wanting to go deeper and I am now working through some coursework using various tools to data mine Twitter and doing analysis as well as doing an overview of TensorFlow through a Udacity Course.
What has amazed me in the few weeks I have been grinding on this is that there are some really powerful resources available to take you very deep in a very short period of time. In addition to the high quality video lectures, there are Jupyter Notebooks with sample code you can modify and various Python scripts that can get you going very quickly. I actually now find myself at the point where I am trying to not figure out if what I want to do is possible but rather what is the best tool or tools for me to test my theories out as fast as possible.
I started out very skeptical about the value of taking the "online education" path to gain new knowledge and I have to say that right now I am pretty blown away by what is available. Stay tuned for more as I go a little deeper in the coming weeks.
No comments:
Post a Comment