The 2nd week of the fastai course
It’s been 2 weeks since I started learning the fastai course. I have not built my model from scratch yet. But I’m close to that as we are about to build our first image recognition models next week.
It surprised me, so far, that there is not that much code really needed here (or it looks like only).
That’s a cool way to start (I believe) as it allows you to understand the ML usage's general concept, and then if you need, you can go deeper and deeper by yourself.
2 weeks ago, I would not understand this news correctly: “Chinese e-commerce giant Alibaba (NYSE: BABA) plans to fully open its conversation artificial intelligence (AI) to other companies, especially those in the smartphone and home appliance sectors, the company confirmed with KrASIA on Friday.”
Simon, from 2 weeks ago, would think that Alibaba opened something for someone. Ok, probably, cool.
Simon from yesterday was like, “Wow, interesting. So, we are talking about transfer learning in this case. Let’s keep that in mind. Maybe, there will be some opportunities where we can use Alibaba’s models and add some “heads.”
Also, today I can understand the joke below 😂:
I also talked with one of my friends about the transfer learning provided by a company called Scopito. But that was not just a talk about their business; we talked about technology and data collection in this project.
As I mentioned in the first reflection’s article, I still see a lot of opportunities in mixing AI and marketing. AI technologies will allow us to deliver more relevant content on different platforms.
What obstacles do I see in this field? I would point to mindset.
Today we had an internal webinar with Sanyam Bhutani organized by the Unpack AI course team.
I asked him whether there is a big difference between a non-tech background and a tech background person when you start learning. He answered that all of the obstacles are in your head. There are plenty of examples when people started at 0 and became one of the top data scientists.
Anyway, it’s not that easy. You just need to go through the questionnaire in the Deep Learning for Coders book to feel that your brain is about to heat up.
Looking forward to the next week to finally build my first something LOL.
And I’ll leave a screen of the Drivetrain Approach principle to remember and for the article’s head photo =)
