Join devRant
Do all the things like
++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatar
Sign Up
Pipeless API
From the creators of devRant, Pipeless lets you power real-time personalized recommendations and activity feeds using a simple API
Learn More
-
I'll use this topic to segue into a related (lonely) story befitting my mood these past weeks.
This is entire story going to sound egotistical, especially this next part, but it's really not. (At least I don't think so?)
As I'm almost entirely self-taught, having another dev giving me good advice would have been nice. I've only known / worked with a few people who were better devs than I, and rarely ever received good advice from them.
One of those better devs was my first computer science teacher. Looking back, he was pretty average, but he held us to high standards and gave good advice. The two that really stuck with me were: 1) "save every time you've done something you don't want to redo," and 2) "printf is your best debugging friend; add it everywhere there's something you want to watch." Probably the best and most helpful advice I've ever received 😊
I've seen other people here posting advice like "never hardcode" or "modularity keeps your code clean" -- I had to discover these pretty simple concepts entirely on my own. School (and later college) were filled with terrible teachers and worse students, and so were almost entirely useless for learning anything new.
The only decent dev I knew had brilliant ideas (genetic algorithms, sandboxing, ...) before they were widely used, but could rarely implement them well because he was generally an idiot. (Idiot sevant, I think? Definitely the idiot part.) I couldn't stand him. Completely bypassing a ridiculously long story, I helped him on a project to build his own OS from scratch; we made very impressive progress, even to this day. Custom bootloader, hardware interfacing, memory management, (semi) sandboxed processes, gui, example programs ...; we were in highschool. I'm still surprised and impressed with what we accomplished.
But besides him, almost every other dev I met was mediocre. Even outside of school, I went so many years without having another competent dev to work with. I went through various jobs helping other dev(s) on their projects (or rewriting them), learning new languages/frameworks almost every time: php, pascal, perl, zend, js, vb, rails, node, .... I learned new concepts occasionally (which was wonderful) but overall it was just tedious and never paid well because I was too young to be taken seriously (and female, further exacerbating it). On the bright side, it didn't dwindle my love for coding, and I usually spent my evenings playing with projects of my own.
The second dev (and one one of the best I've ever met) went by Novo. His approach to a game engine reminded me of General Relativity: Everything was modular, had a rich inheritance tree, and could receive user input at any point along said tree. A user could attach their view/control to any object. (Computer control methods could be attached in this way as well.) UI would obviously change depending on how the user could interact and the number of objects; admins could view/monitor any of these. Almost every object / class of object could talk to almost everything else. It was beautiful. I learned so much from his designs. (Honestly, I don't remember the code at all, and that saddens me.) There were other things, too, but that one amazed me the most.
I havent met anyone like him ever again.
Anyway, I don't know if I can really answer this week's question. I definitely received some good advice while initially learning, but past that it's all been through discovering things on my own.
It's been lonely. ☹2 -
If you don't know how to explain about your software, but you want to be featured in Forbes (or other shitty sites) as quickly as possible, copy this:
I am proud that this software used high-tech technology and algorithms such as blockchain, AI (artificial intelligence), ANN (Artificial Neural Network), ML (machine learning), GAN (Generative Adversarial Network), CNN (Convolutional Neural Network), RNN (Recurrent Neural Network), DNN (Deep Neural Network), TA (text analysis), Adversarial Training, Sentiment Analysis, Entity Analysis, Syntatic Analysis, Entity Sentiment Analysis, Factor Analysis, SSML (Speech Synthesis Markup Language), SMT (Statistical Machine Translation), RBMT (Rule Based Machine Translation), Knowledge Discovery System, Decision Support System, Computational Intelligence, Fuzzy Logic, GA (Genetic Algorithm), EA (Evolutinary Algorithm), and CNTK (Computational Network Toolkit).
🤣 🤣 🤣 🤣 🤣3 -
After 3 days of pain, I finally got my first genetic algorithm and physics engine to work
MP4 version: https://chat.is-going-to-rickroll.me/...
9 -
Diving into genetic algorithms right now (its 4 am), am i just a rare specimen or is this not hard at all?2
-
I got a kindle paper as a hand me down gift. And I feel I'm reading so much more than before, now!
I'm starting a new small novel
A Wrinkle in Time, I'll be reading alongside my girlfriend.
I'm 52% done with a book called
Python Tricks: The Book
Literally coolest book I've touched. Contains a bunch of different tidbits about the language, granted most of them confirmed my understanding, but it's still neat to read and learn about them in a more rigorous setting.
I'm 13% done with another book called
How to Day Trade for a Living
I'm heading for the crypto currency exchange, but with a catch,
I'm reading another book called
Genetic Algorithms with Python
You can probably already guess where I'm heading.
I feel armed with more knowledge and I feel like this is a really great way to start the New Year off.7 -
http://ai-junkie.com is a brilliant website - it's finally allowed me to understand neural networks and genetic algorithms properly!4
-
Ok apparently I forgot rants can only be edited within the first 5 minutes, I thought it was 30, and you can't rant 2 times in 2 hours so I'll have to wait before posting this.
So, I'm doing a Genetic Algorithms class, something I liked since I was 15 yo and didn't know shit about coding, but I loved the carykh videos about it. (here is part 1: https://youtu.be/GOFws_hhZs8 )
The yearly class consisted of 3 little projects to be able to do the final exam and an investigation project to pass the subject without a final exam.
We had to make teams, and I got together with 5 more people.
I have a lot to say about these 5 people, but the only thing I'll say is that I was the most experienced programmer among the 6 of us, if they had any experience at all. Mind this is a third cycle class.
We were allowed to use any technology, as long as we wrote the important algorithms by hand, of course.
The development of the first project was such a mess, that one of the members left the subject.
While developing the second one, we were given the topic for the investigation project; fractals.
It took a lot for us to find an application of fractals where we could use genetic algorithms. Once we found it, fractal antennas, we had to learn about antennas, so we interviewed professionals, and such. We ended up learning to evaluate antennas.
We also found a site that used some parameters to generate fractals, we had the parameterization.
We just had to code it. It was July and we just had to code it by October.
We were 5 people, and "we" were so busy writing the little projects, we fucking couldn't finish the investigation project.
We just had to write the proper algorithms and GUI specifics, without even having to write boilerplate (we used the first project as a template), and they still took so much that we didn't have time for the important project.
That sucked, because I had been coding and investigating in many weekends, I spent countless hours on them, I had to pause development on other projects for these ones; and after all that we have to do the (very shitty) final exam.
Since May, the average people together "working" on the different projects was 2.6. And 100% of the time, I was one of them.
We tried to speed up things in the last months but even with the deadline on us and the project not even started, there was no time we all got to work together.
Dude projects don't just get made, someone has to develop them.
It's so sad we had the project ready to be made and 5 people couldn't finish it. There was so little to do to pass and yet these people couldn't.
I guess it's my bad too. I wish I could rush the project in a couple of weeks, but unfortunately the guy with a job and 8 other subjects can't.
You can find the project in my GitHub. I'll do a requiem of what it was to be one of these days, after I catch up with all I left aside for this subject...3 -
I had the idea that part of the problem of NN and ML research is we all use the same standard loss and nonlinear functions. In theory most NN architectures are universal aproximators. But theres a big gap between symbolic and numeric computation.
But some of our bigger leaps in improvement weren't just from new architectures, but entire new approaches to how data is transformed, and how we calculate loss, for example KL divergence.
And it occured to me all we really need is training/test/validation data and with the right approach we can let the system discover the architecture (been done before), but also the nonlinear and loss functions itself, and see what pops out the other side as a result.
If a network can instrument its own code as it were, maybe it'd find new and useful nonlinear functions and losses. Networks wouldn't just specificy a conv layer here, or a maxpool there, but derive implementations of these all on their own.
More importantly with a little pruning, we could even use successful examples for bootstrapping smaller more efficient algorithms, all within the graph itself, and use genetic algorithms to mix and match nodes at training time to discover what works or doesn't, or do training, testing, and validation in batches, to anneal a network in the correct direction.
By generating variations of successful nodes and graphs, and using substitution, we can use comparison to minimize error (for some measure of error over accuracy and precision), and select the best graph variations, without strictly having to do much point mutation within any given node, minimizing deleterious effects, sort of like how gene expression leads to unexpected but fitness-improving results for an entire organism, while point-mutations typically cause disease.
It might seem like this wouldn't work out the gate, just on the basis of intuition, but I think the benefit of working through node substitutions or entire subgraph substitution, is that we can check test/validation loss before training is even complete.
If we train a network to specify a known loss, we can even have that evaluate the networks themselves, and run variations on our network loss node to find better losses during training time, and at some point let nodes refer to these same loss calculation graphs, within themselves, switching between them dynamically..via variation and substitution.
I could even invision probabilistic lists of jump addresses, or mappings of value ranges to jump addresses, or having await() style opcodes on some nodes that upon being encountered, queue-up ticks from upstream nodes whose calculations the await()ed node relies on, to do things like emergent convolution.
I've written all the classes and started on the interpreter itself, just a few things that need fleshed out now.
Heres my shitty little partial sketch of the opcodes and ideas.
https://pastebin.com/5yDTaApS
I think I'll teach it to do convolution, color recognition, maybe try mnist, or teach it step by step how to do sequence masking and prediction, dunno yet.6 -
MY GENETIC ALGORITHMS INVESTIGATION PROJECT. I WANTED TO RANT ABOUT THAT SO THIS WEEKLY RANT IS PERFECT. I can't write the whole rant rn, stay tuned.5
-
- Finish "Introduction to algorithms"
- Learn some genetic algorithms
- Get my hands dirty on reinforcement learning
- Learn more about data streaming application (My currently app is still using plain stupid REST to transport image). I don't know, maybe Kafka and RabbitMQ.
- Learn to implement some distributed system prototypes to get fitter at this topic. There must be more than REST for communicating between components.
- Implementing a searching module for my app with elastic search.
- Employ redis at sometime for background tasks.
- Get my handy dirty on some operating system concepts (Interprocess Communication, I am looking at you)
- Take a look at Assembly (I dont want to do much with Assembly, maybe just want to implement one or two programs to know how things work)
- Learn a bit of parallel computing with CUDA to know what the hell Tensorflow is doing with my graphic card.
- Maybe finishing my first research paper
- Pass my electrical engineering exam (I suck at EE)1
Top Tags
Weekly Rant
View