Hacker News Comments on
Hacker News Stories and CommentsAll the comments and stories posted to Hacker News that reference this url.
Andrew Ng's machine learning course on Coursera from almost 10 years changed the trajectory of my entire career.
Statistical rethinking by Richard McElreath also helped me to understand bayesian analysis and simulation -- possibly the best hands on bayesian analysis book for beginners
edit: added links
⬐ throw1234651234That course was absolute trash from my perspective - nothing clicked. I am not sure why it's so highly praised, despite covering fundamentals. Pacing and everything was off, and this is from someone with a ton of coding experience and stats background.⬐ thatchercI had a real "click" moment with Statistic Rethinking (2nd Edition) too! I never really "got" Bayes' Theorem before (could use it to solve homework problems and stuff but never appreciated it) but somewhere in SR Chapter 2 the whole point of it really snapped into focus. Good book!⬐ bmitcWould you recommend his original Coursera course or his new course here: https://www.deeplearning.ai/courses/machine-learning-special...?
> Statistical rethinking by Richard McElreath
Thank you for that reference! I'm definitely going to check it out.
If you are able and want to, are you able to elaborate on how you changed the trajectory of your career? I'm a software engineer who was originally wanting to be a mathematician, so I've always been wanting to get back to something a bit more analytical and quantitative.⬐ jerrygenserI would do the first and classic one on Coursera first and then newer deep learning. The original one gives you better grounding in basic ML stuff that is not as "sexy" these days but is still fundamental and used for solving many problems.
I was in a similar boat but on the analytics side wrt to maths. I also took courses at a local university in Math to build up my knowledge. Learning pure math at the college level, like learning to do proofs in front of people on a chalkboard -- I think is very difficult to do unless you take it in the context of a class.
I think you should still give it a go, it's never too late!
edit: answer original question⬐ jerrygenser> If you are able and want to, are you able to elaborate on how you changed the trajectory of your career?
I was in data analytics at a health insurance company in more of a BI and data pulling role. Taking his course gave me the confidence to start doing more machine learning related projects in my company eventually becoming the first person in the company to be called "Data Scientist" after developing multiple models that were making an impact on the business.
edit: answer original question
My advice is to get 'ML' on your resume, some way, somehow, along with some selection of: TypeScript, React, Python, C#, Java, Linux, and Kubernetes.
Start with a good online free course like https://www.coursera.org/specializations/machine-learning-in...
You don't have to be a full-bore data scientist to benefit from the surging interest in ML -- for example, I've made my current focus 'UX for data scientists', I focus on trying to make interfaces that are pleasing to people working in ML, and this, I am pleased to say, just got my ass hired.
It's also salient to know that, over the next decade, while tech salaries will likely trend down (esp. as CoPilot, nocode/lowcode, etc continue to make our lives easier and hence less lucrative ;D) the *amount of addressable work* in ML specifically will almost certainly offset this change.
YMMV in other tech specializations, but I'd still rather be in our industry than most others. Again, there's plenty of work in MLOps that does not require being a data scientist.
Finally, if you can physically live in a geolocation with lower cost of living (or better yet, a currency that is weaker than USD) you can be the lower-cost option that everyone ditches Fancy McSanFransiscoPants for in 2023. Be sure to blog and show your work on GH ;D
⬐ uxcolumboGood points.
I'd like to hear more about UX for data scientists.
Do you mean creating custom user interfaces similar to the ones used in Tableau or PowerBI etc to explore large datasets?
Or just in general following good UI principles when showing data, e.g. Tufte, Few, etc?
Which interfaces / apps do you think do a good job in this regard?⬐ undowareThat's just it -- I don't think there are any great interfaces, yet. Jupyter scratches only the surface of what literate experimental coding could be.
I also have a back-of-napkin argument that the field is due for an invention as revolutionary to ML as the beizer handle was to vector art, or the spreadsheet was to money management.
I'm interested conceptually in eventually leveraging eg VR and sonification for aid in exploring very-high-dimensional datasets.
Meanwhile, yes, there is a lot of just nuts/bolts Good Design to worry about. :)⬐ uxcolumboVR… great example. I’m curious to know whether and what kind of impact the immersiveness of VR will have on exploring large data sets.
I wonder what kind of UIs Deepmind created for their data scientist… similar to Jupyter or better.
Because you're looking for fundamentals and concepts rather than coding, I suggest Andrew Ng's courses , especially . His teaching is beginner-friendly.
⬐ samwestdevHow beginner friendly is this course? Any math prerequisites?⬐ WalterGRI'm quite looking forward to checking them out.
Here's some discussion prior to the release (which was two days ago) of this new specialization (3 courses rather than Ng's original single course):
Andrew Ng updates his Machine Learning course (deeplearning.ai)
Submitted by carlosgg | 29 days ago | 328 points | 124 comments
⬐ lajamerrExcert from the email about the differences between the previous course.
>The Machine Learning Specialization is designed to be accessible for first-time learners and includes:
>An expanded list of topics that focus on the most important machine learning concepts (such as modern deep learning algorithms, and decision trees) and tools (such as TensorFlow)
>Assignments and lectures built using Python -- the programming language of choice for machine learning developers
>New ungraded code notebooks with sample code and interactive graphs to help you visualize what an algorithm is doing and make it easier to complete programming exercises
>A practical advice section on applying machine learning which has been updated significantly based on emerging best practices from the last decade⬐ WalterGRIt's quite important to highlight that what was Ng's single course is now a specialization containing 3 courses.
The site describes the specialization this way:
> It provides a broad introduction to modern machine learning, including supervised learning (multiple linear regression, logistic regression, neural networks, and decision trees), unsupervised learning (clustering, dimensionality reduction, recommender systems), and some of the best practices used in Silicon Valley for artificial intelligence and machine learning innovation (evaluating and tuning models, taking a data-centric approach to improving performance, and more.)
From the email -
> An expanded list of topics that focus on the most important machine learning concepts (such as modern deep learning algorithms, and decision trees)
How much deep learning is in the original course? I got the impression - perhaps wrongly - that it was mostly about older approaches.
Does it make sense for a learner to jump directly into deep learning?
(Obviously the quote above doesn't say "directly into". And "Deep learning" isn't mentioned in the course overviews until Course 3, "Unsupervised Learning, Recommenders, Reinforcement Learning”… But I’m still curious.)