Life after Leo: engineering math

[Edward K. Ream]

Once again I am confronting my life after Leo. This time I have a clear plan. I shall study engineering math under the long-distance direction of Prof. Steve Brunton at the "other" UW, the University of Washington. In other words, I'll study his YouTube videos and his outstanding online course, Data Driven Science.

I started with a superb introductory video: Differential Equations Overview. The first 10 minutes:

- Explain the place of differential equations (diff-eqs) in engineering mathematics.

- Reveal the intimate connection between diff-eqs and linear algebra.

- Explain how eigenvectors help solve systems of diff-eqs.

- Show why mathematics from 200-300 years ago helps develop intuition that is relevant today.

Brunton assumes only that the viewer: has taken a course in calculus and is bright and motivated.

Specifically, Brunton does not assume the viewer remembers much calculus. The series contains refresher lectures for everything the viewer needs to know! Any motivated viewer will finish this video confident they can master this corner of engineering mathematics!

Summary

I no longer fear finishing Leo. My next project is to study all of Brunton's videos, starting with the differential equations.

Edward

P.S. Brunton's online course uses Jupyter Notebooks for exercises. Following his course may suggest new features for Leo.

EKR

[Edward K. Ream]

On Thu, Nov 28, 2024 at 9:22 AM Edward K. Ream <edre...@gmail.com> wrote

I shall study engineering math under the long-distance direction of Prof. Steve Brunton.

I first discovered Steve Brunton via his introduction to the Fourier Series. He is, by far, the best math teacher I've ever had.

Edward

[Ray wang]

wow, Edward,  you are my role model. 

[Thomas Passin]

On Thursday, November 28, 2024 at 10:29:27 AM UTC-5 Edward K. Ream wrote:

I first discovered Steve Brunton via his introduction to the Fourier Series. He is, by far, the best math teacher I've ever had.

This makes me want to check out his courses.  As for the FFT, my GF4 graphic calculator program makes it easy to get a good feel for the FFT.

[Edward K. Ream]

On Thu, Nov 28, 2024 at 6:38 PM Ray wang <wzy...@gmail.com> wrote:

I no longer fear finishing Leo. My next project is to study all of Brunton's videos, starting with the differential equations.

P.S. Brunton's online course uses Jupyter Notebooks for exercises. Following his course may suggest new features for Leo.

wow, Edward,  you are my role model. 

Thanks for the kind words. I'm glad you are inspired.

Edward

[HaveF HaveF]

My next project is to study all of Brunton's videos, starting with the differential equations.

This course looks good.

But, Edward, I would like to remind you that no matter what you are learning now, you must study with LLMs. Whether it is chatgpt or something else. 

One disadvantage of text or video courses is that they are not "alive". You can't discuss with him or translate it into your life experience to understand, you can only understand it according to the author's path.

The way I read some epub books now is, 1. Convert them to markdown documents first, 2. Then throw them to AI, let it summarize the key points chapter by chapter. When I find what details I want to know, I keep asking it in depth. I know that LLMs often have hallucinations, but it is much better than a rubber duck that doesn't talk back.

I even use it to read new papers, throw a pdf to AI and start a discussion, fearless of any subject (as long as I am interested)

 

[Edward K. Ream]

On Sat, Nov 30, 2024 at 9:46 AM HaveF HaveF <iamap...@gmail.com> wrote:

My next project is to study all of Brunton's videos, starting with the differential equations.

This course looks good.

But, Edward, I would like to remind you that no matter what you are learning now, you must study with LLMs.

Hmm. The videos are the gateway to the exercises. Working through the exercises is the only way to learn math.

Edward

[Edward K. Ream]

On Thursday, November 28, 2024 at 9:22:39 AM UTC-6 Edward K. Ream wrote:

I shall study engineering math under the long-distance direction of Prof. Steve Brunton at the "other" UW, the University of Washington. In other words, I'll study his YouTube videos and his outstanding online course, Data Driven Science.

The video Differential Equations Overview is the first lecture of the course ME 564: Mechanical Engineering Analysis.

This page will guide my study. At the top of the page is a list of homework assignments.

Be sure to scroll down to the syllabus. There are three parts: ODEs, Numerical calculus, and (!) Complex analysis. Each part contains a series of lectures. Each lecture contains lecture notes and accompanying .m (Matlab) files. Matlab is free if you use it for less than 20 hours per month.

Edward

[Thomas Passin]

Looking at the syllabus, you should have fun.  It's really racing over a lot of topics.

The first lecture has a matrix example built around weather. It includes both a Matplotlib page and a Jupyter Notebook for running the matrix system.  I imported it as jupytext into Leo and found that I could get the example running in VR3 - with a little tweaking. Doing this, I noticed an error in the Jupytext import code.

After the fix-up tweaks, I executed the notebook and got the attached graphical result.

[Thomas Passin]

Oops, posted too soon.  Here is the screen shot and an explanation of the tweaks and the import bug.

First, VR3 does not understand << named sections >> so I had to comment out the <<prefix>> line.  I could get VR3 to automatically ignore named section lines without much trouble. Since for @jupytext files the nodes all appear in the right order anyway, ignoring them won't hurt anything.

Second - this is the bug - the cell introducer line is supposed to include the language but it doesn't:

# %%

should be (I forget the exact syntax but it includes the language)

# %% python

So I had to add a header line to each node: @language python.

In the script that I wrote during the evolution of jupytext support, I translated the %% python to @language python. I could build that into VR3 for jupytext files.

Third, I had to add a line plt.show() at the end to generate the visible plot that you see in the screenshot.

So with a little bug fix to Leo's importer, and a little work on VR3, you could view and execute this kind of example right in Leo as you study.

[Viktor Ransmayr]

Hello Edward, hello Thomas,

Am So., 1. Dez. 2024 um 17:09 Uhr schrieb Thomas Passin <tbp1...@gmail.com>:

Oops, posted too soon.  Here is the screen shot and an explanation of the tweaks and the import bug.

...

On Sunday, December 1, 2024 at 8:23:05 AM UTC-5 Edward K. Ream wrote:

On Thursday, November 28, 2024 at 9:22:39 AM UTC-6 Edward K. Ream wrote:

I shall study engineering math under the long-distance direction of Prof. Steve Brunton at the "other" UW, the University of Washington. In other words, I'll study his YouTube videos and his outstanding online course, Data Driven Science.

...

Thanks a lot for those links, screenshot & tweaks !

With kind regards,

Viktor

[Edward K. Ream]

On Sun, Dec 1, 2024 at 11:51 AM Viktor Ransmayr wrote:

Hello Edward, hello Thomas,

Thanks a lot for those links, screenshot & tweaks !

You're welcome! The course is the work of a master.

Edward

[Edward K. Ream]

On Sun, Dec 1, 2024 at 10:09 AM Thomas Passin <tbp1...@gmail.com> wrote:

Oops, posted too soon.  Here is the screen shot and an explanation of the tweaks and the import bug.

Thanks for your work. My own experiments show that Leo is a decent platform for notebooks.

Edward

[Thomas Passin]

Ah, my mistake.  It's not python cells that have the language specified in the header line, it's markdown cells:

# %% [markdown]

So it isn't the import that has the bug but my memory.

[Edward K. Ream]

On Sun, Dec 1, 2024 at 12:31 PM Thomas Passin <tbp1...@gmail.com> wrote:

Ah, my mistake.  It's not python cells that have the language specified in the header line, it's markdown cells:

# %% [markdown]

So it isn't the import that has the bug but my memory.

Here's another factoid. @language jupytext isn't all that important. Why? Because jupytext converts markdown parts to python comments when importing. I'm not sure of the exact rules, and the jupytext docs don't seem clear on this point.

Anyway, my experiments show that jupytext does add comments. But that means that @language python would work!

Still, I'm glad I did all the recent colorizer work. It hasn't been wasted.

Edward

[Thomas Passin]

On Sunday, December 1, 2024 at 2:02:14 PM UTC-5 Edward K. Ream wrote:

On Sun, Dec 1, 2024 at 12:31 PM Thomas Passin <tbp1...@gmail.com> wrote:

Ah, my mistake.  It's not python cells that have the language specified in the header line, it's markdown cells:

# %% [markdown]

So it isn't the import that has the bug but my memory.

Here's another factoid. @language jupytext isn't all that important. Why? Because jupytext converts markdown parts to python comments when importing. I'm not sure of the exact rules, and the jupytext docs don't seem clear on this point.

Anyway, my experiments show that jupytext does add comments. But that means that @language python would work!

For VR3, there are two aspects here.  First, in general a jupytext file will have a mixture of code and markdown nodes. So VR3 can't assume that all nodes will be Python.  Second, VR3 cannot assume that a single @language python directive at the top of an at-file will mean that all child nodes will be python, again because there can be a mixture. Not a problem, VR3 is designed to handle a mixture.

[Edward K. Ream]

On Sun, Dec 1, 2024 at 3:46 PM Thomas Passin <tbp1...@gmail.com> wrote:

> On Sunday, December 1, 2024 at 2:02:14 PM UTC-5 Edward K. Ream wrote:

> ...my experiments show that jupytext does add comments. But that means that @language python would work!

> For VR3, there are two aspects here.  First, in general a jupytext file will have a mixture of code and markdown nodes.

True.

> So VR3 can't assume that all nodes will be Python.

That's only true when we first create the markdown section. As I explained above, after we reload the outline the jupytext library will convert the markdown code to python comments.

Edward

[Offray Vladimir Luna Cárdenas]

Happy exploration Edward,

This idea of exploring deep subjects assisted/extended by computational notebooks reminded me of two pretty recommended talks by Sam Ritchie:

1. "Emmy: Moldable Physics and Lispy Microworlds": https://www.youtube.com/watch?v=B9kqD8vBuwU
2. "Computational Physics, Beyond the Glass": https://www.youtube.com/watch?v=Jv2JgzAl5yU

I think that meta tools/systems like Clojure/Lips, Leo or Pharo/GToolkit can enable a virtuous feedback cycle when dealing with deep long processes/questions (that was one of the hypothesis in my PhD thesis while creating and putting metatools in grassroots communities).

Thanks for sharing this links and we will be waiting for your computational notes and notebooks along the process. They'll be welcomed as many of your writings and notes along these years.

Cheers,

Offray

--
You received this message because you are subscribed to the Google Groups "leo-editor" group.
To unsubscribe from this group and stop receiving emails from it, send an email to leo-editor+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/leo-editor/dd2ebc39-c139-47bc-9154-1a2901d1a21en%40googlegroups.com.

[Offray Vladimir Luna Cárdenas]

The computational notebooks for many of those advanced/deep topics is what make them alive beyond the LLM hype and even showcasing when is just oversized technology with pretty grounded rich alternatives.

What I would advice is more finding a community of practice to share the lessons, difficulties and learning and to keep the motivation. This is pretty old and more sustainable tech ;-).

Offray

--
You received this message because you are subscribed to the Google Groups "leo-editor" group.
To unsubscribe from this group and stop receiving emails from it, send an email to leo-editor+...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/leo-editor/bcdf58c2-079f-4728-a34b-939d3b1d8028n%40googlegroups.com.

[Edward K. Ream]

On Tue, Dec 3, 2024 at 1:07 PM Offray Vladimir Luna Cárdenas <off...@riseup.net> wrote:

Happy exploration Edward,

This idea of exploring deep subjects assisted/extended by computational notebooks reminded me of two pretty recommended talks by Sam Ritchie:

1. "Emmy: Moldable Physics and Lispy Microworlds": https://www.youtube.com/watch?v=B9kqD8vBuwU
2. "Computational Physics, Beyond the Glass": https://www.youtube.com/watch?v=Jv2JgzAl5yU

Thanks for the links :-)

Edward