Create accountLogin

Recent searches

No recent searches

Search options

Only available when logged in.
101010.pl is one of the many independent Mastodon servers you can use to participate in the fediverse.
101010.pl czyli najstarszy polski serwer Mastodon. Posiadamy wpisy do 2048 znaków.

Administered by:

Server stats:

509
active users

101010.pl: AboutStatusProfiles directoryPrivacy policy

Mastodon: AboutGet the appKeyboard shortcutsView source codev4.4.0-alpha.5

#aimodel

0 posts0 participants0 posts today
Continued thread
*
AMPEL-CDSS

-New and exciting were our appearances in national media, such as MDR, WDR, Süddeutsche, Spektrum, and Deutsches Ärzteblatt

Our #OpenSource publications:

- Multicenter #Dataset of over 5 million blood samples for #Sepsis detection -> github.com/ampel-leipzig/sbcda

- #CDS #AImodel for early sepsis detection based solely on a complete blood count -> github.com/ampel-leipzig/sbcmo

- Prospectively studied CDS models for detecting #RefeedingSyndrome and hypo- and #Hypercalcemia -> uniklinikum-leipzig.de/einrich

2/3

GitHubGitHub - ampel-leipzig/sbcdata: Laboratory Diagnostics from Septic and Non-septic Patients Used in the AMPEL ProjectLaboratory Diagnostics from Septic and Non-septic Patients Used in the AMPEL Project - ampel-leipzig/sbcdata
MYELECTRICSPARKS

Swiss startup FinalSpark has created a "living computer" using 16 mini brains made from human brain tissue, offering an alternative to traditional silicon-based computing.

Researchers can remotely access this biocomputer, called Neuroplatform, to study artificial intelligence, which usually demands a lot of resources.

MORE DETAILS: myelectricsparks.com/swedish-s

#computers#AI#aimodel
Lunatech

Observations about using #AI to help you write code - I am not a programmer, and I am old and my memory isn't what it used to be (and honestly it was never great - whatever the opposite of a photographic memory is, that's about what I have). But, on rare occasions I need a short program to do some specific task. #Python (specifically #python3 ) seems to be what everyone uses these days, but I never learned Python - the only languages I ever even partially learned were BASIC and Z-80 assembly language (that should tell you about how old I am!).

So in days past if I needed a short program I'd either try to write it as a #bash script (which is a little bit similar to BASIC) or if I felt I had to do it in Python, it would take me hours because I'd literally need to do multiple web searches to figure out the correct Python code for each step, and none of my code was really efficient because I don't know nor understand any of the advanced features of Python (what the hell's a "tuple", anyway?!). And also, Python is kind of a miserable language to work with because it is so picky about things like indentation and syntax. I realize if you write a lot of Python code you can use an #IDE to help you with things like that, but for the two or three times a year I need a script to do something, the learning curve for the IDE would be more effort than it's worth.

But now we have AI, and I find that trying to use it to help write code can be a very, um, interesting experience. For one thing, unlike some web sites which shall remain nameless, it never makes you feel like an idiot or chides you for not having searched hard enough before you asked a question, unlike some of the bullies and a-holes that inhabit those "question and answer" sites.

And AI is great for people like me who have lousy memories, because it remembers the basics of a language, and also a lot of the little tricks that can help make your code more efficient. It suggests things I would have never known nor remembered. It also comments much of the code it creates, so you can actually understand what it is doing (and you can ask it to explain why it used certain statements, so it really is kind of a learning experience).

But at the same time it makes really dumb mistakes, such as mis-matching parenthesis (a no-no in any language). Sometimes the mistakes are obvious, in other cases if you tell it what the problem was (the error message you received, or why the result was not as expected) it will fix the error, though that may take a few tries (and it may even repeat previously given incorrect code, which is frustrating). But what I find interesting is that it often gets the hard parts right, but totally fumbles the easy stuff.

But here are my questions: If you find an obvious error, and you tell the AI about it, does it learn from its mistake, or will is spit out the same bad code to the next person who gives it similar input? And also, is there a particular #aimodel that is great for generating code (especially Python code)?

#programming#coding#code
ExploreLive feeds
About