4 ms·

I hate when people use the word "algorithms" like that. Which algorithms? Dijkstra? I know this is the way non-technical people refer to the spooky AI action a

by msm_ 4y ago

I hate when people use the word "algorithms" like that. Which algorithms? Dijkstra? I know this is the way non-technical people refer to the spooky AI action at a distance happening in social media to maximise engagement, but in my opinion it just makes everything confusing when actually trying to understand the argument.

WediBlino 4y ago
algorithm /ˈalɡərɪð(ə)m/ noun plural noun: algorithms a process or set of rules to be followed in calculations or other problem-solving operations, especially by a computer. "a basic algorithm for division"
whitemary 4y ago
> like that. Like what? Correctly?
lcnPylGDnU4H9OF 4y ago
I didn't go through academia for my computer science knowledge so I may be wrong about the definition of "algorithm". I understand it as a series of steps to accomplish a task (e.g. a food recipe is a list of ingredients and an algorithm for rendering something delicious). If that's not mistaken, then the algorithm is something that's being used to make decisions. A different term to use could be "machine-made decisions", which I personally think describes the issue better. Still, (and unless my understanding is mistaken) the word algorithm is technically accurate to this kind of decision-making.
- groffee 4y ago
  [dead]
- guestbest 4y ago
  I agree with this. If there is any responsibility it is borne by people and management’s business process. Blaming algorithms is absolving people of responsibility
  - jkestner 4y ago
    And that’s exactly what the people responsible want. Another layer to launder any culpability, just like consultants, contractors and corporations. The next step is to grant algorithms personhood so we can prosecute them while their creators continue to collect money. The executives and shareholders who profit the most from the decisions must be the ones held responsible.
groffee 4y ago
[dead]
cheeze 4y ago
It's a simplification for folks who don't understand the nuances past "an algorithm" The intended audience here isn't computer scientists... it's... the public. It's pretty obvious what they mean by "algorithms", no? I feel like this is just being "akshyually" guy
- netsharc 4y ago
  Ah, so just like people wanting to avoid "chemicals" in what they eat?
- mrWiz 4y ago
  I'm skeptical when this kind of language is used, particularly in the media. It feels like they're trying to drum up outrage by intentionally phrasing such that the public understanding differs from the technical meaning. I'm not saying that's what's happening in this case, just my typical initial reaction to seeing this pattern.
kelseyfrog 4y ago
It could be anything from a regression to DL, to who knows what else. Even in the simple case, say a linear regression with all the assumptions satisfied, predictions are interpretable as means, which is to say that (and again, assumptions holding, errors are normally distributed) half of observations fall below the mean and half above it. Using the mean in a normative capacity implies that half of patients will get kicked out before they would otherwise end their treatment. There are ways around this, like quantile regression, or using prediction intervals, but again, we don't know the specifics of the method which leaves us in the dark as to the specific way the problem manifests.
jerf 4y ago
There is no size limitation on the word "algorithm". It is perfectly acceptable to define a complete decision tree as being an "algorithm", in addition to "decision trees" in general being an "algorithm". There is no restriction on it being the sort of cute little thing you can stick in a textbook on a single page and do intensive mathematical analysis on it. You can refer to the complete rendering pipeline for a frame in a modern AAA game as an "algorithm", or the entire code for instantiating an EC2 instance, or, yes, even the algorithm that takes an input a person and their medical cases and returns what the organization should do about it including the possibility of just cutting them loose. It's a perfectly sensible use of the term as is.
- mort96 4y ago
  That's the problem though. If there are clear eligibility criteria, and a computer program lets the operator enter information and answer eligible or ineligible, that's an algorithm used to cut off care. If a computer program decides that someone is ineligible because they're not a US citizen, that's an algorithm deciding. If the computer figures out that you no longer live in the service area for your plan, that's the algorithm deciding to cut off your care. I would expect any insurance provider to use computer programs (i.e algorithms) to decide whether someone is eligible or not. The problem is what kind of algorithm.
  - nerdponx 4y ago
    > The problem is what kind of algorithm. ...and how its outputs are interpreted. Ultimately it's less about what kind of algorithm and more about how it's used, or in this case abused. As another user mentioned, that 16.6 days should have been the mean or mode of a posterior distribution, from which you could obtain 75% and 95% credible intervals, or at least a conditional expectation accompanied by an (estimated) sampling error and traditional 75/85/95/whatever% confidence interval. Crossing the 17-day line could have flagged this person's case for manual review. Even if the 95% credible interval was 15-18 days, anyone who understands how statistical models work would understand that manual review is a more sensible reaction than automatic cutoff, unless you simply don't care about the 5% chance of falsely denying people health coverage. The fact that the model is statistical is not more problematic than a chain of if/else decisions. The problem is the fact that good statistical practice was disregarded in its usage. And as others have said, there was no intention of pursing good statistical practice here, and no reason to ascribe any benefit of doubt to any decision-makers involved in deploying such a model/algorithm. It's just a fig leaf for wanting to deny people health coverage.
stronglikedan 4y ago
It's correctly used in this context, where it's not referring to a particular well known algorithm in the context of computer science, but rather the " process or set of rules to be followed in calculations or other problem-solving operations" that the computer takes to determine eligibility. Considering the context and the intended audience, it's used correctly here.
jart 4y ago
Now that algorithm has become a snarl word, what word can serve as a suitable replacement? So that computer scientists can talk about algorithms, without people getting triggered that they're talking about the algorithm.
InCityDreams 4y ago
I just asked someone what they thought an algorithm is: A set of hidden calculations that can be tweaked mainly for the benefit of the owner.
- htag 4y ago
  Jargon often gets hijacked to mean something different to the general population than it does to experts. This happens in a lot of different ways, but often the word is picked up in FUD campaigns and it picks up a negative connotation.
jxidjhdhdhdhfhf 4y ago
It's like people worried about "chemicals" in food. Ok, which chemicals are you actually worried about?
wahern 4y ago
Join the club. Some of us are still trying to get over the ever expanding scope of the term, AI.

by msm_ 4y ago

WediBlino 4y ago

algorithm /ˈalɡərɪð(ə)m/ noun plural noun: algorithms a process or set of rules to be followed in calculations or other problem-solving operations, especially by a computer. "a basic algorithm for division"

whitemary 4y ago

> like that. Like what? Correctly?

lcnPylGDnU4H9OF 4y ago

I didn't go through academia for my computer science knowledge so I may be wrong about the definition of "algorithm". I understand it as a series of steps to accomplish a task (e.g. a food recipe is a list of ingredients and an algorithm for rendering something delicious). If that's not mistaken, then the algorithm is something that's being used to make decisions. A different term to use could be "machine-made decisions", which I personally think describes the issue better. Still, (and unless my understanding is mistaken) the word algorithm is technically accurate to this kind of decision-making.

groffee 4y ago
[dead]
guestbest 4y ago
I agree with this. If there is any responsibility it is borne by people and management’s business process. Blaming algorithms is absolving people of responsibility
- jkestner 4y ago
  And that’s exactly what the people responsible want. Another layer to launder any culpability, just like consultants, contractors and corporations. The next step is to grant algorithms personhood so we can prosecute them while their creators continue to collect money. The executives and shareholders who profit the most from the decisions must be the ones held responsible.

groffee 4y ago

[dead]

cheeze 4y ago

It's a simplification for folks who don't understand the nuances past "an algorithm" The intended audience here isn't computer scientists... it's... the public. It's pretty obvious what they mean by "algorithms", no? I feel like this is just being "akshyually" guy

netsharc 4y ago
Ah, so just like people wanting to avoid "chemicals" in what they eat?
mrWiz 4y ago
I'm skeptical when this kind of language is used, particularly in the media. It feels like they're trying to drum up outrage by intentionally phrasing such that the public understanding differs from the technical meaning. I'm not saying that's what's happening in this case, just my typical initial reaction to seeing this pattern.

kelseyfrog 4y ago

It could be anything from a regression to DL, to who knows what else. Even in the simple case, say a linear regression with all the assumptions satisfied, predictions are interpretable as means, which is to say that (and again, assumptions holding, errors are normally distributed) half of observations fall below the mean and half above it. Using the mean in a normative capacity implies that half of patients will get kicked out before they would otherwise end their treatment. There are ways around this, like quantile regression, or using prediction intervals, but again, we don't know the specifics of the method which leaves us in the dark as to the specific way the problem manifests.

jerf 4y ago

There is no size limitation on the word "algorithm". It is perfectly acceptable to define a complete decision tree as being an "algorithm", in addition to "decision trees" in general being an "algorithm". There is no restriction on it being the sort of cute little thing you can stick in a textbook on a single page and do intensive mathematical analysis on it. You can refer to the complete rendering pipeline for a frame in a modern AAA game as an "algorithm", or the entire code for instantiating an EC2 instance, or, yes, even the algorithm that takes an input a person and their medical cases and returns what the organization should do about it including the possibility of just cutting them loose. It's a perfectly sensible use of the term as is.

mort96 4y ago
That's the problem though. If there are clear eligibility criteria, and a computer program lets the operator enter information and answer eligible or ineligible, that's an algorithm used to cut off care. If a computer program decides that someone is ineligible because they're not a US citizen, that's an algorithm deciding. If the computer figures out that you no longer live in the service area for your plan, that's the algorithm deciding to cut off your care. I would expect any insurance provider to use computer programs (i.e algorithms) to decide whether someone is eligible or not. The problem is what kind of algorithm.
- nerdponx 4y ago
  > The problem is what kind of algorithm. ...and how its outputs are interpreted. Ultimately it's less about what kind of algorithm and more about how it's used, or in this case abused. As another user mentioned, that 16.6 days should have been the mean or mode of a posterior distribution, from which you could obtain 75% and 95% credible intervals, or at least a conditional expectation accompanied by an (estimated) sampling error and traditional 75/85/95/whatever% confidence interval. Crossing the 17-day line could have flagged this person's case for manual review. Even if the 95% credible interval was 15-18 days, anyone who understands how statistical models work would understand that manual review is a more sensible reaction than automatic cutoff, unless you simply don't care about the 5% chance of falsely denying people health coverage. The fact that the model is statistical is not more problematic than a chain of if/else decisions. The problem is the fact that good statistical practice was disregarded in its usage. And as others have said, there was no intention of pursing good statistical practice here, and no reason to ascribe any benefit of doubt to any decision-makers involved in deploying such a model/algorithm. It's just a fig leaf for wanting to deny people health coverage.

stronglikedan 4y ago

It's correctly used in this context, where it's not referring to a particular well known algorithm in the context of computer science, but rather the " process or set of rules to be followed in calculations or other problem-solving operations" that the computer takes to determine eligibility. Considering the context and the intended audience, it's used correctly here.

jart 4y ago

Now that algorithm has become a snarl word, what word can serve as a suitable replacement? So that computer scientists can talk about algorithms, without people getting triggered that they're talking about the algorithm.

InCityDreams 4y ago

I just asked someone what they thought an algorithm is: A set of hidden calculations that can be tweaked mainly for the benefit of the owner.

htag 4y ago
Jargon often gets hijacked to mean something different to the general population than it does to experts. This happens in a lot of different ways, but often the word is picked up in FUD campaigns and it picks up a negative connotation.

jxidjhdhdhdhfhf 4y ago

It's like people worried about "chemicals" in food. Ok, which chemicals are you actually worried about?

wahern 4y ago

Join the club. Some of us are still trying to get over the ever expanding scope of the term, AI.