I’m working on OCR to recognize the scores on pinball machines. I have about a quarter million photos encompassing every model of pinball machine in the world but I just don’t have the know-how to accommodate all the font styles.
Ask HN: What problem are you close to solving and how can we help?
151–160 of 486 posts
Re: Ask HN: What problem are you close to solving and how can we help?
#152I could use some help with some heuristics for Machine Learning, like how much data do I need to make a workable model, what framework/approach makes more sense given my ultimate goals. Here's an example: there's a lot of ML tutorials on doing image identification. Like you have a series of images: picture one might have an apple and a pear in it, picture 2 might have an apple, orange, and a banana in it. Where I'm s…
Define workable model. Do you care more about recall (how many of the images with label X will be labeled by your model with label X) or precision (how many of the images where the model says have label X are actually with label X)?
It is a good fit for ML, but you need to be clear on what the results will be. If you expect 100% accuracy, that won't happen. Even 90%+ accuracy would be require a lot of effort.
Re: Ask HN: What problem are you close to solving and how can we help?
#153What are the resources to look at for designing the architecture of a scheduling system (for coroutines/threads)? Would I first look at how Operating systems implement it? Or how VMs like Erland implement it?
Same question for effects systems, where would I look to understand how they're designed and the trade offs for their design decisions?
Re: Ask HN: What problem are you close to solving and how can we help?
#154I could really use some help with adversarial attacks. If there's someone trying to use CV to recognize stuff and we are trying to prevent it (basically, a black box situation) - is it viable to use adversarial attacks at all? Will it work long term? Can they overcome our AA by downsampling and adding noise? Can we make another AA that would still prevent them using CV? If it's better to chat elsewhere - my Twitter i…
If you have access to their model it's possible, and adversarial attacks can be made to survive different sorts of processing. But like all piracy prevention tools it's a game of cat and mouse so if you're working against an intelligent adversary it'll come down to how many resources each of you can contribute to this fight.
(Btw your twitter is not in your profile. Not that I have anything to add in private, just thought you should know if you're relying on it.)
Re: Ask HN: What problem are you close to solving and how can we help?
#155I am a bit ashamed to ask about such a trivial topic on HN, but I am not really sure how AJAX in Wordpress plugins works. I have a plugin that exports some WooCommerce orders into XLS. I would like to add a progress bar via AJAX, because the export may take very long for thousands of orders. But I am not really sure how to use AJAX in context of Wordpress specifically. I would love to see a minimal functional example…
First of all, you'll need to make PHP to display "progress", probably you'll need to override ob_start() or something like that, and find a format that let you append new progress on the response on the fly.
I guess you already have an URL on your Wordpress setup that triggers this export. Let's call it {url}/export.
Wordpress already has jQuery by default included. So'll you need to call that URL using jQuery $.post and then, accordingly to the response, update your progress bar.
There is nothing specifically about Wordpress on this, besides the fact that you need to setup your own URL on Wordpress to do this, and then include your own JS after jQuery. That's all.
If you find this too-complicated, a quick-hack is to create a page on WP Admin called Export Tool, and then on your theme create page-export-tool.php. That .php will be called when visited that Export Tool page.
Re: Ask HN: What problem are you close to solving and how can we help?
#156I want to bring back old school distributed forum communities but modernise them in a way that respects attention and isn’t a notification factory. Mastodon is a pretty inspirational project but the Twitter influence shows, I miss the long form writing that was encouraged before our attention spans were eroded. Not at all close to solving it, but it’s been on my mind for a long time. Would love to hear if there are o…
Be sure to add cryptographic signatures on the postings and up/downvotes (also with signatures). Then many people can develop content ranking and blocking, and who knows someone may get it right
Re: Ask HN: What problem are you close to solving and how can we help?
#157What are the resources to look at for designing the architecture of a scheduling system (for coroutines/threads)? Would I first look at how Operating systems implement it? Or how VMs like Erland implement it? Same question for effects systems, where would I look to understand how they're designed and the trade offs for their design decisions?
Everyone starts with callbacks, moves on to coroutines, and eventually ends up writing polling loops. Can take decades for each person to work through it.
Erlang's messaging may have potential as another alternative
Re: Ask HN: What problem are you close to solving and how can we help?
#158I’m working on OCR to recognize the scores on pinball machines. I have about a quarter million photos encompassing every model of pinball machine in the world but I just don’t have the know-how to accommodate all the font styles.
Have you tried an off-the-shelf solution like Tesseract? It works quite well if you do the recommended preprocessing.
Re: Ask HN: What problem are you close to solving and how can we help?
#159How to find motivation/energy to do a long-term creative project when having a full time jobs + other responsibilities?
The brain hates to start things and loves to finish things. This can be hacked in that a working session should always leave something unfinished.
Say you're writing a novel. Every writing session (but the first, obviously) should cover the end of the last scene and the beginning of a new one, AND THEN STOP, i.e. not finish the new scene.
Your brain will want to come back to the work to finish it, which overcomes the friction of "starting" something new every time.
It's easier said than done. It's surprisingly difficult to leave something unfinished at the end of each work session. But that's the trick.
Re: Ask HN: What problem are you close to solving and how can we help?
#160Stateful, exaclty once, event processing without the operational capacity to run a proper Flink cluster. This thing needs to be dead simple, pragmatic and cheap/simple to operate and update. The only stateful part in our infra at the moment is a PG database. We are going to start work on this in a weeks, so I'm looking for some insights/shortcuts/existing projects that will make our lives easier. The goals is to proc…
(EDIT: just realised that you specifically mentioned stateful event processing, while what I describe below are two approaches for stateless, exactly-once event processing)
Having had a few cracks at this problem, in my opinion using locks is the wrong approach.
What you will want is:
* split all input data in batches (eg batches of 10k records, or periodic heartbeats every X seconds, etc)
* assign each batch a unique identifier
* when writing data to the output store, store the batch id along the data;
* when retransmitting a batch for whatever reason, reuse the same batch id and overwrite any data in the output store that matches this batch id.
Obviously this becomes more tricky when you’re dealing with eg window functions or more complex aggregations.
In this situation, I believe that an approach such as “asynchronous barrier snapshotting” works best. Every X seconds, you increment an epoch. While incrementing, you stop ingestion. Then you first tell the output source to create a checkpoint, then the input source to create a checkpoint, and once both have been checkpointed, you can continue streaming data again.
Anyway, these are two approaches I’ve used over the years that work well. Explicit locks don’t work well in distributed processing, imho.