Hacker Newsnew | past | comments | ask | show | jobs | submit | vanboxel's commentslogin

I'm a very technical person, but I never feel dumber than when doing mathematics. Self-teaching it is especially challenging. In many fields, one can learn to recognize general patterns and quickly apply them elsewhere. In math, I find that general pattern matching can be helpful for getting ideas, but there's so shortcut to working through the logic required in a rigorous mathematical proof. It's almost a different kind of thinking, and it takes a ton of effort.

If you're interested in self-teaching, I recommend an excellent Abstract Algebra book designed explicitly for that: https://today.williams.edu/books/abstract-algebra-a-student-.... This was the book that got me into rigorous mathematics and started me on the journey of doing a math PhD (defending this semester!). It only requires high-school algebra, and honestly, even that isn't really needed.


Several versions ago, there was a mod (maybe called "homeworld") that had a sink of you sending various resources through a portal back "home". It started simple with N iron plates, but then they wanted 10N oil barrels, and eventually red circuits. It also added some farming resources (which were affected by local pollution, so you had to plan where to put them). It was fun to have little goals, but I don't know that it's been updated recently.


The "Math Misery" blog has several problems like this, for example http://mathmisery.com/wp/2017/04/02/cryptarithmetic-puzzle-8...


That story sounds exactly like Ooka Tadasuke. https://en.wikipedia.org/wiki/%C5%8Coka_Tadasuke#Famous_case...


Interestingly, wiki links back to a Paul Graham article referencing Ooka http://www.paulgraham.com/property.html


From that article:

> In 1956, an illustrated book was created by I.G. Edmonds, an American military officer. Published by the Pacific Stars & Stripes, it was called Solomon in Kimono: Tales of Ooka, a Wise Judge of Old Yedo.[4] Edmonds' work was published in 1961 as Ooka the Wise, and then in 1966 renamed The Case of the Marble Monster and Other Stories and made widely available to American schoolchildren by the Scholastic publishing company.

Perhaps GP read the story thanks to Edmonds.


That’s Ooka for sure. I did a PowerPoint on it for my criminal law class.


I don't often self-promote, but my channel, DanDoesData, hosts the long-form type of series you may be looking for. Every video is a hour of live-stream where I research, design, and implement the model on the fly, explaining my thought process along the way.

One of my longest and most "real-world" projects was a (relatively simple) self-driving car based on single camera input. I used this model for a self-driving PowerWheels race. It later served as a starting point for a small autonomous RC car.

Here's the start of that playlist, but I've put together several projects. Working on a GAN for fake startup names right now.

https://www.youtube.com/watch?v=meU1fXXIAOQ&list=PLFxrZqbLoj...


Can you describe some of your methodology for detecting the fakes? Presumably you have trained a detector on a large data set; did you generate this yourself using several popular deepfake tools?


I can't go into depth on our models or data sets, because this is a cat and mouse game between creation/detection. We create deepfakes using all known tools, currently non-public methods as well as our own methods.


I was a little bummed to see no mention of `slimux`. It's a fantastic way to execute commands in a real shell (another tmux pane) from the comfort of your ViM session. By combining this with say IPython running in the shell, it's a powerful way to do data analysis. I wrote an ebook on this kind of workflow at http://dvbuntu.github.io/compute/


I really enjoy vim tabs for multiple files. Especially if I don't need to see all of them at once, but just quickly switch between them. `:tabnew foo.txt` to open another file, then `gt` to go to the next tab and `gT` to go back.

If you work interactively with data, you might also enjoy Slimux (https://github.com/epeli/slimux), a vim plugin that allows you to send lines from vim to an arbitrary tmux pane. I usually have IPython running in such a pane so I never have to copy+paste. I've got an ebook I've been kicking around describing this workflow in more detail here (http://dvbuntu.github.io/compute/posts/2050/01/01/workflow.h...).


What's the difference between buffers and tabs? I am genuinely curious.


Here's a nice explanation of how they are different in practice, rather than just in theory:

https://stackoverflow.com/questions/26708822/why-do-vim-expe...


I created a video course aimed at ML enthusiasts with software experience. It introduces Google's TensorFlow library, walks through code examples, and tries intuitively explain some of the statistics. https://www.packtpub.com/big-data-and-business-intelligence/...

But I also agree with other posters that you should study fundamental statistics first.


I actually studied exactly that! The optimal calculations for headway distance get a bit involved, but I coded up a modest algorithm and simple simulator. https://github.com/dvbuntu/oddities


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: