Same, because it's not left-handed people who are more creative, it's lefties in politics who are more creative.
Its a psychology trait of openness that governs curiosity, imagination, and well creativity. This trait also highly correlates with leaning left politically.
There's lots of posts on HN for developments and companies doing OCR and Document Extraction. It's a classic CV problem but still has come a long way in the past couple years
Yeah, this is a very well-traveled road, but LLMs have made some big improvements. If you asked me (the guy who wrote the original piece linked above) what I'd use if accuracy alone was the goal, probably would be AWS Textract. But accuracy and structure? Gemini.
Around a month ago there was a PINN post[1] on here and there was a healthy amount of skepticism in the comments. Even in the toxic positivity of LinkedIn, commentors say they're overhyped when a ML "Influencer" posts that one GIF with a MLP and PINN fitting to an oscillator. I would be interested to see what they're actively being used for.
I fully agree with the comments in that post. I started studying them because, well, they sound really cool, but my first impression was definitely that this sounded like a lot of effort (computationally) to solve a single equation.
But then I only tested on simple equations where traditional solvers have no difficulties.
In my previous position, we studied the behaviour of black holes with exotic geometries, and we never could make our solvers work with the added time dependence. I would be very curious to see how a PINN would have fared on this (given enough compute time of course).
I can't speak to attitudes, but when I was in grad school years ago, Koreans were very well represented in my program. There were fewer Korean students than Chinese but more than Indian. Generally they do well in both coursework and research so there's likely at least some opportunity to stay in the states, assuming they don't have obligations back home.
To measure academic achievement they use First Year GPA. It makes sense they're correlated, both require studying known material for a test. But is GPA the best measure, especially first year? I would be interested in other metrics like 3rd-4th year GPA or placements into jobs and such.
I believe there's some research that suggests your college GPA is not correlated strongly with job performance[1]. I suspect if SAT/ACT scores are strongly correlated with GPA, there's reason to suggest it may also not be strongly correlated with job performance (but I can't find anywhere that tests that).
I suppose this shouldn't be surprising. School does not train you to be a good office worker.
The goal of school IS NOT to make you a good office worker. The goal of school is to create the ability to think creatively, rationally and critically and make you a better citizen of society through those processes.
Job performance is subjectively measured by your bosses which has political implications. It is inherently a terrible metric.
> The goal of school IS NOT to make you a good office worker
Who defines this "goal"? Each student has their own individual goals. I bet you if you surveyed students
"The goal of school is to create the ability to think creatively, rationally and critically and make you a better citizen of society through those processes."
Would rank all the way to the bottom and they all are terrible reasons to get into $200k of debt.
Many go to college without a second thought just because they think it's what they are supposed to do. Many go because they don't want to be a manual laborer and they want to be able to make more money. Many go to get drunk and party.
This is such a cynical and tired take. I would imagine that most people attending college at least intend to learn something and level up some ability. Many of them will also enjoy partying in their early 20s. Both can be true, and there’s not a singular purpose of attending school.
Good job, starting off strong with a real substantive criticism.
> I would imagine that most people attending college at least intend to learn something and level up some ability.
Your imagination is lacking and you have not spoken with many 18 year olds, the vast majority are not highly motivated learners. A large number of degrees teach skills that are not even valuable enough to pay back the amount of money that is put in. So they are leveling up in worthless shit. All these indebted graduates are clamoring to have the government pay for their bad decisions. None of them blame the institutions that charge these exorbitant amounts to supposedly become "cultured" and "creative".
> the vast majority are not highly motivated learners
The honest truth is they shouldn't be going to college. Because they went to school and didn't come out cultured or creative.
(I went to a state school. It was easy to structure your schedule with zero liberal arts. I almost did. To the extent I have regrets about my time in college, it's in not taking more liberal arts classes.)
> I also went to a state school, I took the liberal arts classes, they have added nothing of value to my life. If you want to get cultured, read.
I think that means you were very fortunate. For me, 2 years at a community college was essential for getting my career in tech off the ground. Learning to write at the college level, being able to read/search/cite journals, and the practice of being able to synthesize a bunch of information into a cohesive document and/or presentation opened a ton of doors for me.
I get the impression some people figure this stuff out either through more demanding work in high school or from having smart parents who inculcate this knowledge into their kids. For me, I didn't have that.
> I took the liberal arts classes, they have added nothing of value to my life
Sure, many were useless. But the few that were good were golden. Sorry you missed that.
> If you want to get cultured, read
This is necessary but insufficient. The discussion is essential. To be clear, I'm not saying you can't become cultured outside college. (Of course you can.) But that coming out of college uncultured is close to a waste for most graduates, i.e. those graduating with a degree that isn't immediately in high demand at a six-figure wage.
If a student isn't a highly-motivated learner, and their goal is maximising lifetime net earnings, they shouldn't be going to college.
> If a student isn't a highly-motivated learner, and their goal is maximising lifetime net earnings, they shouldn't be going to college
I disagree. I think further education in valuable topics is valuable to people even if they are not motivated. I don't think getting into massive debt for this education is valuable to them. There is nothing inherent about college that requires it to be so expensive and simultaneously useless.
All the defenders of the current system always respond with the tired "college is not about teaching how to do a job, it's about teaching culture, creativity, and critical thinking" bullshit that is not backed by any data. If we go by actual results, college is about getting people into massive debt to fund college administrators.
What college and education should be about is teaching people things that can give them the ability to contribute to themselves and others.
> further education in valuable topics is valuable to people even if they are not motivated
Sure, but that doesn’t need to be college. Many European countries have colleges in name only that actually function as trade schools. An unmotivated learner should go to a cheap 2-year trade school and then start earning.
> college is about getting people into massive debt to fund college administrators
Completely agree. That said, we have standout colleges where the purpose is to educate our next generation of elites. That’s still important.
My girlfriend in high school had no idea what she wanted to major in, no idea what career she wanted, no idea what her future looked like at all. Anytime I'd ask her, she'd get upset, because she didn't want to stress about it before she had to.
We eventually broke up, and she went to college, changed her major 3 times, spent the whole time drunk at parties, and graduated with a degree with a relatively limited scope of jobs. She graduated 4 years ago, during which she's worked as a receptionist, a dog groomer, and a call center rep.
We shortly got back together as she was graduating college, and I encouraged her to apply for jobs for which her degree would come in handy. Every single time, she refused, because she "wasn't qualified enough" even though she had her Bachelors in what they were asking for.
I say all this to say that I think many kids are corralled into college without an actual end goal in mind. They go because it's what they're supposed to do. In my anecdotal experience, there isn't much thought about intending to learn or leveling up some ability.
SAT is well-correlated with 1st year GPA, but not well-correlated with eventual job placement (although it does correlate pretty well with eventual income). However, 1st year GPA is well-correlated with 2nd year GPA, which is well-correlated with 3rd-year GPA, which is well-correlated with 4th year GPA, which is well-correlated with eventual job placement, which is well-correlated with entry-level salary, which is well-correlated with mid-career salary.
A pretty useful model for life is that it's a series of contests, and doing well at the previous contest gives you an advantage for the next couple contests, but only the next couple contests. By the time you get to mid-career, nobody really cares what your high school GPA was. However, because each contest determines which set of subsequent contests you'll face, performance early on can have outsize effects on eventual life outcomes. You typically won't be applying for CEO jobs if you worked retail your whole life, unless you lie your ass off and bullshit convincingly to executive recruiters.
In my own experience, first-year GPA in college was a cakewalk. I had straight As until Junior year, then things got a bit more demanding and I was caught off-guard because college had been pretty easy up to that point. (Large state university).
My experience was the opposite. First two years were a blur, I was in the trenches juggling calculus and organic chemistry and physics along with my major work and the extra busywork courses they throw on top of all that too. By the time I got to my upper division work in the third year it was like grades didn't matter at all. There's be no more assignments or quizzes, maybe two exams to determine your entire grade for the course, that would be pretty easy to get an A on if you attended class and didn't sleep.
It could just be that professors don't want to invest a lot of time designing a bevy of coursework. In the weed out classes they are all lecturing off a textbook, which provides for them slides to use and a general schedule of topics, as well as banks of questions they could pull from and permutate and reuse. Its a lot of time saved for sure. In the upper division, there is no textbook or anyone designing any course, they might literally have a sentence description of what the course is pitched about in the catalog then they take it from there. The professor has to come up with what to talk about from their field for an hour twice a week, and sitting there drafting up those slides is enough work if they don't have any from last year to use, considering they are usually also a full time researcher on top of this.
I agree. Every other comment here seems to be like "well duh" and I'm... skeptical. My experience is that the ACT/SAT seem to be good indicators of getting good grades in well-defined spaces. But things like creativity, curiosity, work ethic are much better predictors of other kinds of success that frankly matter much more in the real world.
I know some really, really unintelligent people who got good grades in college. They just ate books.
Undergraduate GPA predicts lifetime earnings[1], incoming test scores and GPA are highly predictive of both advanced degrees (which increase earnings), and increased earnings within degrees [2], [3].
I suggest these effects are because being a good student aka "eating books" is correlated with conscientiousness. They show up to lectures, prepare, and test well.
And conscientiousness is very highly correlated with lifetime achievement, AND fufillment [4]. So measuring conscientiousness, and signalling high conscientiousness is a really good idea.
IQ is great, but conscientiousness is how you get things done [5]
That is...an extremely narrow study to use to make that broad an assertion. GPA of two classes from 2010 at a single business school in China, their starting salary, and then their salary in 2018?
Why is it narrow? N is a few hundred at a well ranked and good amount of rigor university, and is aimed at the graduating class and then their earnings as they rose through to probably mid career positions past entry level.
I haven't read the entire study mostly because I don't care, but I think you're wrong in your statements.
GPA is well correlated to earning potential, as well as earnings in reality.
Their Figure 1 is why we keep having these discussions in society. It's grossly misleading and not what a scatterplot for a 0.46 correlation looks like. I know what the figure is, it's just done in a way to overstate a case and ignore variability within bin.
If that figure were about anything else, people would be screaming bloody murder about misleading figures and overly generalized interpretation.
I'm in favor of allowing for the use of test scores but they get abused and the language in this report is a good example of how this happens. Scores have these real but modest correlations with real world situations, but then get used as rulers of atomic precision without any context or recognition of their massive limitations.
It makes the authors of this report look either deceiving or ignorant of statistics or both.
The thing is: completely unsuitable students need to fail out in the first year. That's what "weeder" courses are for. They prevent students wasting time and money, only to fail out years later.
Which means: you don't have the same stats for 3rd and 4th year students.
The reason people measure first year GPA is that, in the past, all freshmen at a college took the same set of classes. If you measure 3rd year GPA, you get confounded by the difference between physics students and French students.
Obviously, the worth of the metric goes down over time as first-year curricula differentiate from each other.
> But is GPA the best measure, especially first year?
No. For example, SAT score is a better single measure than GPA is. But you can't use that to check the validity of SAT scores.
I believe they do look at other things (EDIT: or at least the study this paper was based on did):
"Using detailed admissions data from IvyPlus institutions, Chetty, Deming, and Friedman (2023) show that SAT and ACT scores also predict career success, including high levels of earnings and attendance at elite graduate schools, holding family income constant."
Right, but what I think they're trying to do is to confirm that the larger, more comprehensive study also applies to Dartmouth. But the fact that they didn't also run the test against other metrics might suggest they didn't want to publish that data publicly.
Or if you're more cynical, they failed to produce those results.
If I were judged by my first year GPA I'd be homeless. Instead I'm really quite successful in spite of my first try (and miserable failure) at university.
When I went back and was paying for tuition out of my own pocket as opposed to magic sky money falling into my lap from the US Dept of Education, things were a LOT different.
I think some of us (like me) were too immature at 18-22 to make good use of college. I'm not sure the amount of skin in the game would have made much difference, but I wouldn't have been able to pay for it myself, so that's a benefit. When I went back and got high grades, I also had manna-money from Direct loans. I just had a lot more life experience and maturity.
I see your point and relate, but had you asked me at 18 if I wanted to go so badly that I'd work my way through without loans, that would have been a hard pass.
Right, I worded it funny. It would be good if I couldn't have afforded it because I wouldn't have wasted the time and money.
The options in my area after high school were kind of unattractive, though. I could work retail, or maybe go to trucker school, or try to find odd jobs helping people with computers. My current career path (veterinarian) didn't occur to me at that time.
Contrary to what some may think: I would strongly suspect that particularly high scores on SAT/ACT would also correlate fairly strongly with well-paying job placements and long term GPA.
Note I am not saying that people who don't do particularly well on SAT/ACT can't also succeed on college and beyond, I had a pretty average score myself and consider myself fairly successful. But all the people I know who had the highest scores on SAT's were also the people most motivated to study and work extremely hard whether it be based on family pressure or just an inherent drive to be better than everyone else, and even to this day these people I knew who had scores in like the top 95+% are for the most part the most successful people I know primarily just due to an inherent drive to succeed at all costs.
Of course this will also depend on what metric you use for success. Also of course there will be people who do extremely poorly in traditional education but become wildly successful.
There's some nuance to the "work ethic" factor. Work ethic makes a big difference in academic outcomes, but it depends on where that work is targeted.
Anecdotal evidence, part 1:
I grew up in the suburbs and went to public schools that, at the time, were good-not-great within the state. In that setting, I was a high achiever by any sensible metric. I was friends with a lot of the other "smart kids," and I definitely worked harder than some of them, but outside of school, I was also working on different things. My hobbies and extracurriculars weren't strictly academic, but they certainly set me up for academic success better than sports or video games would have.
Anecdotal evidence, part 2:
One of my classmates was "freak of nature" levels of gifted. I shared math classes with him for three years of high school, and to my knowledge he got one math test question wrong in that entire time period. He was also one of the school's best tennis players, and he made it into so many state concert bands/jazz bands/choir groups that their schedules overlapped and he couldn't do all of them. Last I checked, he was finishing up a PhD in neuroscience. But get this: he was our salutatorian, solely because someone else who did no sports and few clubs took more summer classes, and thus had the same GPA with more total credit hours.
I didn't read the entire document in depth, but perhaps it has been shown elsewhere that first year GPA is a good quality predictor of GPA in subsequent years.
https://www.arrantpedantry.com/2020/03/24/umlauts-diaereses-...