A Listing of my Article Series in Deep Learning, Reinforcement Learning & Machine Learning
Welcome to my articles on Deep Learning, Reinforcement Learning, Meta-Learning, and Machine Learning. The purpose of this article is to index the series and articles I wrote so far. If you want to reach me or have non-technical comments regarding the articles, please read the F.A.Q. first for the most common questions and contact methods.
Deep Reinforcement Learning Series
This is one of the biggest Series that I want to tackle next: Deep Reinforcement Learning. This is an outline of the…
Speech Recognition Series
In this speech recognition series, we start our discussion with the basic phonetics. In the first article, we…
TensorFlow 2 Series
In this series, we cover major topics in deep learning coding with TensorFlow 2 — for computer vision, NLP, and…
AI Chips Series
In this series, we look into the technology landscape and trends of the AI chips.
AI Bias & Fairness Series
A good video lecture on an AI topic may be watched by hundreds of thousands of people. But a nice lecture on AI…
QC — Quantum Computing Series
This series covers the basics of quantum computing. The first 2 articles cover the basics followed by how it is built…
Not part of a series yet
Is the drawing above a portrait of a young lady or an old lady? Such optical illusion needs special skills to create in…
NLP — Word Embedding & GloVe
BERT is a major milestone in creating vector representations for sentences. But instead of telling the exact design of…
NLP — BERT & Transformer
Google published an article “Understanding searches better than ever before” and positioned BERT as one of the most…
How to scale the BERT Training with Nvidia GPUs?
In 2015, ResNet-50 and ResNet-100 were introduced with 23M and 45M parameters respectively. Fast forward to 2018, the…
Buy a Deep Learning Computer — David v.s. Goliath
MVP in the tech world is not “Most Valued Player”. Minimum Viable Product (MVP) means testing out hypotheses, finding…
Deep Learning Software Setup with GPU
Open-source platform rules DL. However, setting up open-source software can be a “Plug and Pray” experience. The lack…
Mixed Precision in Deep Learning
In our last article, we discuss how to purchase a deep learning computer. Insider the article, mixed-precision is…
First, I appreciate all the kind words from my readers. My blog serves two purposes for me. By explaining technologies in an approachable way, I open myself to perspectives that I did not see before. Second, I come back to these notes to refresh my memory. I am glad if it helps you in your own journey too.
How can I find your new articles?
I rarely promote my blog. I don’t put my articles behind the Medium paywall. So Medium does not categorize it or promote it. I respect that. So follow me in Medium may be the best way to do it.
How can I connect you professionally on Linkedin?
Feel free to connect me on Linkedin at Jonathan Hui. In avoiding my Linkedin newsfeed to be filled with non-technical posts, I usually don’t accept recruiters and professionals/students that are not in the same field. This year I realize my feeds are getting worse, therefore, I become a little bit more selective about accepting invites.
Some people try to reach me through e-mail. Usually, unless I know you well, I don’t respond to unsolicited emails. For specific questions on Medium, please leave a comment on the specific article. For non-technical questions, please message me through Linkedin.
Where is the source code and what tools you use for the blog?
I don’t supply source code. I find it impossible to support them. In particular, TensorFlow API is often backward incompatible. I suggest readers google it and pick the one with good community support or frequent updates.
Like to ask a technical question.
If you have technical questions, please leave a message in Medium in the related article. I find it too hard to answer them on Linkedin. Sometimes I may miss your question but sometimes, I need more clarification to answer them. I suggest readers reviewing their questions before posting them. Also, you can save the diagrams or equations that you have questions and then, drag and drop them into your response. This helps me a lot!
Troubleshooting is never a show-and-tell business. So I usually do not answer coding problems. This is too hard without reproducing the issue myself. I do suggest you google your problem first. If you cannot find the answer, there is a good chance that you are dealing with something special to you.
What is your professional work?
Once a while, people on Linkedin ask me about my professional projects. Because of confidential reasons, I never discuss specific projects. After experiencing a few odd experiences, I decide to apply a blanket policy of not discussing any professional works in email or Linkedin.
How often do you update your articles?
If the articles were published longer than 12 months, I strongly suggest readers be more cautious. This field changes very fast and unfortunately, like the fashion business — something popular last year becomes out of fashion now. Because of the limited bandwidth, I usually only add new information to an article when I get something important in my head that I want to write it down. :-) I also respond less to comments and questions on old articles.
Let’s have a coffee.
I appreciate your invitation. But I don’t drink coffee so this is not a good opening line for me. lol. I live in S.F. and ditch my car for Uber and Lyft. So it will not be easy if you live in South Bay. Because of my schedule, I don’t engage in promotional or marketing activities. So I rarely accept these requests. But I always enjoy nice conversations. So just be specific when you have something in mind. :-)
Can you be my mentor or recommend a research topic?
I am really flattered by your request. But I am sorry that I don’t have the bandwidth. I will focus on this blog that can reach or help a wider audience. There are many top universities that have their AI courses open online. They will be an excellent starting point.
How about sharing a startup idea?
Many people have reached out to me for the ideas they have. Most of them are amazing. I wish I have the bandwidth to follow up. If you want to discuss it, it will be nice to be specific. I usually respond badly to vague messages. I hope that I will allocate more time for it this year.
Can I use your diagrams?
I am open to sharing for non-commercial purposes. Unfortunately, misuses become more frequent than I expect. So, please use it only for education and non-commercial purposes. If it is third-party diagrams, I usually put a caption with an external link. If there are no captions, it will be likely created by me. Please add proper credits for both cases. If it is for a research paper, you may want to check with me with a LinkedIn message. I usually double-check the article to make sure the diagrams without captions are originated from me.
Credits and references
I try not to explain technology as a historian or overwhelm my readers with too many links. When everything pops, nothing pops. It also develops a habit for readers to ignore them. But I always try my best to give proper credits. For external images, I put an external link in the caption. If I miss the proper links, please let me know. I also try to beef up the reference section. But this will not be as strict as a published research paper. But many links do refer to the original research paper or important information, so if you are a researcher on a specific topic, please go through those links more thoroughly. This will be a continuous learning process. Let’s do it better together. :-)
But being said, some sites copy my whole article and give a fake link to the source. Sometimes, hi-tech people do things first and ask questions later. Do things at any cost is, unfortunately, happening too often. Gradually, it leads to decisions that they believe are right. Let’s not go into that path either. But feel free to translate the articles for educational purposes. I will appreciate it if you put back a link to the original article and leave a message on the original article to let people know your translations.
Hehe. I ban people from my blog. Every year, it is about 2–3 people. I really don’t need to do it but I just don’t want un-welcomed behavior un-noticed sometimes. I do not give out specific reasons. But this is extremely small comparing to the number of readers. Sometimes, people get self-entitlement, self-righteous, or judgemental. Social media trains our brains to give judgment in a second. I just want to create some awareness here. As President Obama said:
Achieving real change was more complex than being “as judgmental as possible”.
I personally find it interesting.
But again, I don’t give an explanation on why people get banned. Any response without knowing the person is futile. From the ML perspective, if your behavior is considered an outliner, your chance of being banned increases. Again, that is about 2–3 people out of many. Being said, I usually filter out the negative emotions and take comments seriously.
Do you want to write a book or review a book?
Sorry. I do not have the bandwidth now.