|
echinopsis posted:perhaps it’s unique amongst screenings in that it uses radiographic data rather than values 🤷♂️ its cuz nerds wanna see a tiddy
|
# ? Jan 5, 2020 21:07 |
|
|
# ? Apr 29, 2024 04:45 |
|
haunted bong posted:its cuz nerds wanna see a tiddy
|
# ? Jan 5, 2020 22:06 |
|
haunted bong posted:its cuz nerds wanna see a tiddy I don't think most nerds wanna see the *inside* of a tiddy
|
# ? Jan 6, 2020 00:15 |
|
animist posted:I don't think most nerds wanna see the *inside* of a tiddy when you’re desperate you’ll take what you can get
|
# ? Jan 6, 2020 02:36 |
|
catching up with the thread animist posted:one other way to think about the linear stuff is just to intuit that a linear operation followed by another linear operation is linear. so, stacking linear layers, you're really only training a single linear transformation. not having trained a machine ever, my cynical guess would be that if you clamp everything into a [0,1] range, then one part of a network learning a wild swing*(+500) followed by the next part learning a wild swing*(-100000) will only propagate so far and not render the whole network obviously meaningless this doesn’t mean that a network not utilizing these nonlinearities is a bad learner no, the problem is that the teacher teaching these wild swings is a bad teacher or to put it another way: animist posted:if we could specify exactly what operation we wanted the network to do, we wouldn't need a neural network, now would we? “if we can teach a machine to be intelligent, then we won’t have to figure out what intelligence is” I don’t think it’s going to work like that
|
# ? Jan 10, 2020 15:23 |
|
Artificial Intelligence Makes Bad Medicine Even Worse Here's a better article on the issues with Google's breast cancer screening technology. tl;dr is that medicine is hard and AI is attacking the easiest part, but it's not clear that there's any value in that.
|
# ? Jan 11, 2020 16:51 |
|
that's a p. good tweet.
|
# ? Jan 11, 2020 17:18 |
|
animist posted:I don't think most nerds wanna see the *inside* of a tiddy don't kinkshame
|
# ? Jan 11, 2020 19:26 |
|
https://twitter.com/StatsPapers/status/1223073098066350080 I'm genuinely surprised that this wasn't all known a long time ago.
|
# ? Jan 31, 2020 04:36 |
|
ultrafilter posted:https://twitter.com/StatsPapers/status/1223073098066350080 plz to be summarizing
|
# ? Jan 31, 2020 14:10 |
|
Captain Foo posted:plz to be summarizing thirty second takeaway is : k-fold cross validation is better than a single test/train split by a factor sqrt(k), more or less
|
# ? Jan 31, 2020 21:06 |
|
the whole thing is 61 pages and 274 numbered equations of inequality chasing and like gently caress im gonna read any more of it
|
# ? Jan 31, 2020 21:09 |
|
statisticians amirite?
|
# ? Jan 31, 2020 21:10 |
|
lancemantis posted:statisticians amirite? no
|
# ? Feb 1, 2020 00:56 |
|
fritz posted:thirty second takeaway is : k-fold cross validation is better than a single test/train split by a factor sqrt(k), more or less sure thank u for your service
|
# ? Feb 1, 2020 00:56 |
|
lancemantis posted:statisticians amirite? depends on your confidence interval
|
# ? Feb 1, 2020 07:21 |
|
I'll show you my confidence interval
|
# ? Feb 2, 2020 01:44 |
|
echinopsis posted:I'll show you my confidence interval planck time, huh
|
# ? Feb 2, 2020 11:05 |
|
|
# ? Feb 4, 2020 03:39 |
|
https://twitter.com/rchrdbyd/status/1227431038831468546
|
# ? Feb 14, 2020 01:05 |
|
does check off all 35 of the checkboxes, yes, when can you start??
|
# ? Feb 14, 2020 02:05 |
|
thanks for posting my resume
|
# ? Feb 14, 2020 02:47 |
|
saw spark, know he's alright
|
# ? Feb 14, 2020 12:15 |
|
CS472 Data science and AI for COVID-19 Description This project class investigates and models COVID-19 using tools from data science and machine learning. We will introduce the relevant background for the biology and epidemiology of the COVID-19 virus. Then we will critically examine current models that are used to predict infection rates in the population as well as models used to support various public health interventions (e.g. herd immunity and social distancing). The core of this class will be projects aimed to create tools that can assist in the ongoing global health efforts. Potential projects include data visualization and education platforms, improved modeling and predictions, social network and NLP analysis of the propagation of COVID-19 information, and tools to facilitate good health behavior, etc. The class is aimed toward students with experience in data science and AI, and will include guest lectures by biomedical experts. Course Format Class participation (20%) Scribing lectures (10%) Course project (70%) Prerequisites Background in machine learning and statistics (CS229, STATS216 or equivalent). Some biological background is helpful but not required. Staff Prof. James Zou (jamesz at stanford) Syllabus (tentative) Date Lecture Readings/notes April 10 Overview of COVID-19 and health systems during pandemic Nigam Shah guest April 17 Epidemiological predictions and modeling April 24 Infectious disease background of COVID-19 Michele Barry guest May 1 Project proposal May 8 ML for COVID-19 drugs Russ Altman guest May 15 COVID-19 genomic analysis Julia Palacios guest May 22 Project milestone presentation May 29 NLP analysis COVID-19 information on Twitter and FB June 5 Final demos
|
# ? Mar 26, 2020 21:26 |
|
I really hope they post lecture notes for that. Applying ML/AI to diseases is hard and I'd love to have pointers on how to do it right.
|
# ? Mar 26, 2020 22:20 |
|
ultrafilter posted:I really hope they post lecture notes for that. Applying ML/AI to diseases is hard and I'd love to have pointers on how to do it right. i wouldn’t hold my breath on getting that from a class some stanford prof threw together in his living room.
|
# ? Mar 26, 2020 22:24 |
|
ultrafilter posted:I really hope they post lecture notes for that. Applying ML/AI to diseases is hard and I'd love to have pointers on how to do it right. gonna be taking copious notes at the "how to use ML to develop a COVID drug" lecture
|
# ? Mar 26, 2020 22:47 |
|
yeah, uh, speaking as a college professor, that syllabus is garbage. maybe suitable for a graduate seminar where people already know what they're doing and all they need is a prompt. because that's all you're gonna get -- a prompt and "okay, uh, go make some stuff now." note that out of 9 class sessions, four are guest lecturers and three are student presentation days. perfect class for a professor who wants to get something with big name recognition on his CV and do minimal work of his own. okay class if you're already up to speed and just want a topic and a work period to fart around. poo poo class if you go into it expecting to receive some sort of useful training. e: and the two professor-led sessions are "NLP analysis" and "epidemiological modeling," both well-understood basic topics. what am quadratic curve Sagebrush fucked around with this message at 22:59 on Mar 26, 2020 |
# ? Mar 26, 2020 22:56 |
|
4xx at Stanford indicates an experimental graduate level class.
|
# ? Mar 26, 2020 23:08 |
|
yeah by the Stanford numbering scheme that's a graduate class e:fb quote:Numbering System
|
# ? Mar 26, 2020 23:16 |
|
sign me up for the experimental artificial language class or maybe the early graduate hardware systems class.
|
# ? Mar 26, 2020 23:41 |
|
i love advanced undergraduate analysis of algorithms
|
# ? Mar 27, 2020 00:38 |
|
ultrafilter posted:4xx at Stanford indicates an experimental graduate level class. ah well then it's probably exactly what it says on the box. "grad students, go make a thing, if you need more information here's a list of papers to read, i'll be writing grants. if you make anything cool i'm putting my name as first author"
|
# ? Mar 27, 2020 00:41 |
|
Sagebrush posted:ah well then it's probably exactly what it says on the box. "grad students, go make a thing, if you need more information here's a list of papers to read, i'll be writing grants. if you make anything cool i'm putting my name as first author" don’t give away the sausage factory!!
|
# ? Mar 27, 2020 01:25 |
|
Bloody posted:i love advanced undergraduate analysis of algorithms at some schools it’s like a weeder course for upperclassmen
|
# ? Mar 27, 2020 04:51 |
|
Bloody posted:i love advanced undergraduate analysis of algorithms pairs well with experimental artificial language
|
# ? Mar 27, 2020 04:59 |
|
I remember when my weeder class in my engineering program was 1st semester, freshman year I still feel like I made the right decision but for the wrong reason
|
# ? Mar 27, 2020 06:07 |
|
i went to music school and our weeder courses were ear training and music history
|
# ? Mar 27, 2020 16:25 |
|
my weeder course was having to show up
|
# ? Mar 27, 2020 17:18 |
|
|
# ? Apr 29, 2024 04:45 |
|
Sagebrush posted:ah well then it's probably exactly what it says on the box. "grad students, go make a thing, if you need more information here's a list of papers to read, i'll be writing grants. if you make anything cool i'm putting my name as first author" what kind of hosed up PI doesn't let the main grad student on the project be first author? how else will he ever graduate
|
# ? Mar 27, 2020 17:45 |