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
-
Probably the biggest one in my life.
TL:DR at the bottom
A client wanted to create an online retirement calculator, sounds easy enough , i said sure.
Few days later i get an email with an excel file saying the online version has to work exactly like this and they're on a tight deadline
Having a little experience with excel, i thought eh, what could possibly go wrong, if anything i can take off the calculations from the excel file
I WAS WRONG !!!
17 Sheets, Linking each other, Passing data to each sheet to make the calculation
( Sure they had lot of stuff to calculate, like age, gender, financial group etc etc )
First thing i said to my self was, WHAT THE FREAKING FUCK IS THIS ?, WHAT YEAR IS THIS ?
After messing with it for couple of hours just to get one calculation out of it, i gave up
Thought about making a mysql database with the cell data and making the calculations, but NOOOO.
Whoever made it decided to put each cell a excel calculation ( so even if i manage to get it into a database and recode all the calculations it would be wayyy pass the deadline )
Then i had an epiphany
"What if i could just parse the excel file and get the data ?"
Did a bit of research sure enough there's a php project
( But i think it was outdated and takes about 15-25 seconds to parse, and makes a copy of the original file )
But this seemed like the best option at the time.
So downloaded the library, finished the whole thing, wrote a cron job to delete temporary files, and added a loading spinner for that delay, so people know something is happening
( and had few days to spare )
Sent the demo link to client, they were very happy with it, cause it worked same as their cute little excel file and gave the same result,
It's been live on their website for almost a year now, lot of submissions, no complains
I was feeling bit guilty just after finishing it, cause i could've done better, but not anymore
Sorry for making it so long, to understand the whole thing, you need to know the full story
TL:DR - Replicated the functionality of a 17 sheet excel calculator in php hack-ishly.8 -
Rage..
Boss: 'I've got a PowerPoint presentation with 45 charts in it. Oh, and a huge excel sheet with the data for it. Please build some of the charts in our own software, with the given data.'
Easy, I thought.
Yea, thanks to the person that gave my boss the data.. The half of the important columns were removed (privacy stuff).
And.. Excel? Oh, and his calculations are nowhere documented nor consistent.
I converted excel to postgres, easy.
It took me 2 hours to fkn research what he calculated in one line chart, just to implement it in like 10 minutes.
2 hours, man I could made awesome stuff in that time!
I guess I should write this in CAPSLOCK to make it more interesting. I'm just raging in my head 😂1 -
We are devs right?
We have cpus and gpus lying around right?
We are still alive... right? 🤔
How about we do our part and utilise our PCs for helping with COVID-19 research.
I've stumbled across this little tool that not only keeps me warm at night but helps researchers with several diseases.
https://foldingathome.org/iamoneina...
It's like a a bitcoin miner but for research purposes, no it's not a dodgy bitcoin miner.
Oh and feel free to keep yourself anonymous as there are stats that will identify your username - when they work.
There are installers for windows, Mac, and linux distros so everyone can get involved.29 -
Imagine if you will, a fictional world outside our own.
In this world, the requirement for getting a drivers licenses is 4 years of research into quantum mechanics.
- Was it interesting? Yeah.
- Did I learn it because I had to? Yup.
- Will I use the harmonic oscillation calculations of a particle when driving my car. Fuck no!
- Did it cost me an ungodly amount of money? It sure did!
- Will some dumb people still say it was useful because it is the minimum (fictional) barrier to entry for driving a car. You bet your sweet ass they will!!!
It was about as useful as any made up requirement, make-work, self-funding, circle-jerking, waste or time and money to feed the pockets of people who are too scared to do actual work so they teach, can be.
I paid all that money to be taught technology that was old when my mother was in school.
In the first year out of school, with only a $300 subscription to PluralSight some uDemy courses and hard work, I learned 100X as much as everything they put in front of me in school.
-------
School has its place.
Children who don't understand the importance of learning and need their hand help.
Adult children (some of which on on their 3rd or 4th degree) who also need their hand held.
People too afraid to enter the real world.
Doctors.
-------
I would do it again because it is the minimum requirement of entry, but thats nothing more than a bullshit make-work project.
Play their game as long as you need to. Keep your own game in mind. Don't drink the koolaid, just fake a sip. Then when the time is right, play by your own rules.
Peace4 -
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
Top Tags
Weekly Rant
View