Flawed Algorithms Are Grading Millions of Students’ Essays • Hacker News Board

Re: Flawed Algorithms Are Grading Millions of Students’ Essays

#251

Post by lr4444lr » Fri, Aug 30, 2019, 1:35 PM UTC

piokoch wrote: post #181

dahart wrote: post #133

> Utah has been using AI as the primary scorer on its standardized tests for several years. “It was a major cost to our state to hand score, in addition to very time consuming,” said Cydnee Carter, the state’s assessment development coordinator. The automated process also allowed the state to give immediate feedback to students and teachers, she said. Yes, education takes time and costs money. Yes, not educating is b…

Frankly, I can't believe what I am reading. The idea that some "AI" grades essays automatically is idiotic and has nothing to do with education. Where is the place for discussions? Where is the place for ideas confrontation? Where is the place for writing style development? How this AI is supposed to grade things like repetitions (that can be either good rhetorical tool or a mistake, depending on context), etc? Who t…

I agree with you wholeheartedly but I think there's a stronger argument to be made here: the algorithms being used "work" only on a correlation based on an ignorance of the scoring metric. If the students under test knew even sketchily how the system worked, e.g., points deducted if your average sentence word length > 7, points added if your word length stddev is greater than 2, and the students could meaningfully push their scores up by focusing on these proxies that don't _actually_ measure what a human would say is quality work - or even they can even get gibberish[0] rated highly - then the whole thing is a fraud. No one will stand for a grading system that only works by virtue of obscurity.

[0] https://www.nytimes.com/2012/04/23/education/robo-readers-us...

Re: Flawed Algorithms Are Grading Millions of Students’ Essays

#252

Post by JustSomeNobody » Fri, Aug 30, 2019, 1:35 PM UTC

crankylinuxuser wrote: post #220

thombat wrote: post #185

Earlier quoted context omitted.

The final paragraph from that example is steaming gibberish that nobody could mistake for English: "Calling has not, and undoubtedly never will be aggravating in the way we encounter mortification but delineate the reprimand that should be inclination. Nonetheless, armed with the knowledge that the analysis augurs stealth with propagandists, almost all of the utterances on my authorization journey. Since sanctions ar…

The let me posit another idea... There's 2 ways of finding out these artifacts of AI essay grading: pure luck, and being able to afford extensive test-prep (rich). The luck one can't be accounted for. So I im lead to believe that the purpose of these essays and their AI grading is to find and escalate rich people.

> So I im lead to believe that the purpose of these essays and their AI grading is to find and escalate rich people.

Well, of course. How many poor people are allowed to decide what is good for children's education?

Re: Flawed Algorithms Are Grading Millions of Students’ Essays

#253

Post by jkdufair » Fri, Aug 30, 2019, 1:35 PM UTC

ssivark wrote: post #141

jkdufair wrote: post #122

Earlier quoted context omitted.

I have a grant where were are doing just that. Implementing more or less SOTA research using fairly vanilla LSTM networks from 2-3 years ago (primarily Taghipour & Ng) to provide low stakes feedback to students on their essays in one of our teaching tools at Purdue. It’s based on research using the Kaggle ASAP database and we have found it to be pretty accurate across a variety of domains in early testing. Though som…

Very interesting. Are there any meta-analyses / reviews that summarize progress in this area? Would it be possible to share your grant proposal -- I'd be curious to get an idea of what is being attempted.

It's an internal grant and I'm not sure I'd be allowed to share it. We are adding AES to our peer-review app. Currently as an additional "grader" to the peer reviews since that's what the PI requested. Since the tool allows unlimited submissions until the review date, I hope to add it as a "pre-flight" estimate to give students a chance to get a rough prediction of the score they will receive and a metric they can use as they revise until the due date.

I'm not aware of any meta-analyses myself. I have been keeping up with the ASAP competition and various attempts to improve on the initial systems for a number of years. The two papers I believe are having the most success are [1] and [2]. [3] seems promising for balancing the opposing forces of high accuracy for true positives and the risk of false positives via adversarially crafted inputs.

I'm also vaguely aware of research happening around extracting features from neural nets. I'd love to be able to help students understand why the system is predicting a particular score.

[1] https://www.aclweb.org/anthology/D16-1193 [2] https://arxiv.org/pdf/1606.04289.pdf [3] https://arxiv.org/pdf/1804.06898.pdf

Re: Flawed Algorithms Are Grading Millions of Students’ Essays

#254

Post by jcranberry » Fri, Aug 30, 2019, 1:37 PM UTC

Aromasin wrote: post #234

clay_the_ripper wrote: post #152

Earlier quoted context omitted.

Wow this is pretty shocking. I can understand using automated systems for something like math problems, it makes sense. There’s (usually) one right answer. But essays? This should be banned.

Having been forced to use an online math software for all my homework while at school, I vehemently disagree. It was so poor that it became a meme within my year group. It would mark you as incorrect for using too many decimal places, even though it wouldn't tell you how many significant figures was required. I often remember it marking my answer as incorrect, even though it was identical to the answer they gave. Som…

That's just poor software. As long as the software and teachers or professors allow for going over answers and checking for correctness it should be alright.

Re: Flawed Algorithms Are Grading Millions of Students’ Essays

#255

Post by JustSomeNobody » Fri, Aug 30, 2019, 1:40 PM UTC

stochastic_monk wrote: post #129

dmoy wrote: post #82

Earlier quoted context omitted.

For the not eating, it's pretty easy to get data. It's like 1 in 5 children live in food-insecure households in the US and maybe 1 in 20 of those very insecure, so not eating before school provided lunch is common enough that if you're grading tons of papers you'll run into kids like that.

It could also be a student suffering from anorexia nervosa, which the confessional aspects of the essay would fit well with.

I'm confident that your example would of a less percentage than those mentioned in dmoy's comment.

Re: Flawed Algorithms Are Grading Millions of Students’ Essays

#256

Post by imtringued » Fri, Aug 30, 2019, 1:40 PM UTC

dahart wrote: post #133

> Utah has been using AI as the primary scorer on its standardized tests for several years. “It was a major cost to our state to hand score, in addition to very time consuming,” said Cydnee Carter, the state’s assessment development coordinator. The automated process also allowed the state to give immediate feedback to students and teachers, she said. Yes, education takes time and costs money. Yes, not educating is b…

I can't even comprehend how someone can use automation for a task like this... It completely goes against human nature. In a world where all jobs have been automated teachers would be the last ones to go before humanity is completely obsolete.

Re: Flawed Algorithms Are Grading Millions of Students’ Essays

#257

Post by jcranberry » Fri, Aug 30, 2019, 1:41 PM UTC

dmoy wrote: post #82

harry8 wrote: post #34

Earlier quoted context omitted.

Some of these will be 100% true as well. But don't make the mistake that there are no kids who go for shock value or are wantonly manipulative when they know it can't come back to them. So how many are true and how many false? I have no clue. Literally none. And no it doesn't make me feel any better about the screams of existential agony even if that were a low percentage. Could be high too.

For the not eating, it's pretty easy to get data. It's like 1 in 5 children live in food-insecure households in the US and maybe 1 in 20 of those very insecure, so not eating before school provided lunch is common enough that if you're grading tons of papers you'll run into kids like that.

https://www.ers.usda.gov/topics/food-nutrition-assistance/fo...

Food Security Status of U.S. Households with Children in 2017 Among U.S. households with children under age 18:

84.3 percent were food secure in 2017. In 8.0 percent of households with children, only adults were food insecure. Both children and adults were food insecure in 7.7 percent of households with children (2.9 million households). Although children are usually protected from substantial reductions in food intake even in households with very low food security, nevertheless, in about 0.7 percent of households with children (250,000 households), one or more child also experienced reduced food intake and disrupted eating patterns at some time during the year.

Re: Flawed Algorithms Are Grading Millions of Students’ Essays

#258

Post by dahart » Fri, Aug 30, 2019, 1:53 PM UTC

_0ffh wrote: post #208

dahart wrote: post #133

> Utah has been using AI as the primary scorer on its standardized tests for several years. “It was a major cost to our state to hand score, in addition to very time consuming,” said Cydnee Carter, the state’s assessment development coordinator. The automated process also allowed the state to give immediate feedback to students and teachers, she said. Yes, education takes time and costs money. Yes, not educating is b…

>Turns out the instructions said the essay would be scored on verbal efficiency; getting the point across clearly with the fewest words. I started playing around and realized that the more words I added, the higher the score Apart from the fact that your story is straight up frightening, isn't this part completely backwards, too? I mean, clearly using more words to convey the same message is /less/ efficient, not mor…

Yes. Exactly. Before I figured out how to game the program, my son and wife were editing shorter. That’s what the instructions said to do. And, that’s also a major strategy for decent writing: brainstorm a lot, then edit down to the good parts. What this means is the software’s scoring is an anti-incentive to good writing. Used as a teaching aid, it’s actually doing pure damage, not good. Not only can it not score reliably, nor provide meaningful feedback, it’s actually actively teaching a very wrong way to write. But it is cheaper than humans, and it does give immediate feedback, so there’s that.

Re: Flawed Algorithms Are Grading Millions of Students’ Essays

#259

Post by JoeAltmaier » Fri, Aug 30, 2019, 2:21 PM UTC

mannykannot wrote: post #244

JoeAltmaier wrote: post #240

Earlier quoted context omitted.

Clearly, that referred to 'doing well on essays' correlating to good analytical thinkers. Which is does.

That assumption lacks justification when the scoring does not actually measure analytical thinking. Any statistical evidence for it is suspect as a predictor of future outcomes when a high score can more easily be gamed than 'honestly' achieved.

Scoring is not the point here; the analytical thinker is gaming the test to pump the score, thus proving they are an analytical thinker. Not a statistical argument; a suggestion that the screen works, when it is abused. Because it is abused.

Anyway, yeah, not really a correlation.

Re: Flawed Algorithms Are Grading Millions of Students’ Essays

#260

Post by DATACOMMANDER » Fri, Aug 30, 2019, 3:16 PM UTC

bendbro wrote: post #127

DATACOMMANDER wrote: post #124

Earlier quoted context omitted.

That’s what you’re doing wrong.

Thus aren't you as well?

I meant that your self-consciousness and constant worry is probably harming your social interactions more than anything. Or not. It was for me in the past.