|
Post by theevan on Jun 22, 2023 10:13:28 GMT -5
|
|
|
Post by Rob Hanesworth on Jun 22, 2023 10:32:31 GMT -5
Members of both parties in the US are most likely to smile when learning about a scandal involving a member of the other party.
|
|
|
Post by Village Idiot on Jun 22, 2023 10:35:23 GMT -5
Members of both parties in the US are most likely to smile when learning about a scandal involving a member of the other party. Shocking.
|
|
|
Post by Russell Letson on Jun 22, 2023 10:42:03 GMT -5
A quick read-through suggests a couple places where this kind of research can go sideways, starting with using the "Face API from Microsoft’s Azure’s Cognitive Services." This from Microsoft's introductory page about that service: Microsoft will be retiring facial recognition capabilities that can be used to try to infer emotional states and identity attributes which, if misused, can subject people to stereotyping, discrimination or unfair denial of services. These include capabilities that predict emotion, gender, age, smile, facial hair, hair and makeup. Existing customers have until June 30, 2023 to discontinue use of these capabilities before they are retired. learn.microsoft.com/en-us/azure/cognitive-services/computer-vision/overview-identityThen there's the question of how exactly facial expressions were characterized and categorized. Our monkey brains do this kind of sorting all the time, but even then our evaluations are hardly infallible. What's a poor algorithm to do? The problem with deriving the human implications of the data is that there has to be a human in the chain somewhere, and the closest thing to a decent validation model I can think of is a large-numbers approach: show the data (pictures of faces) to a very large number of viewers and get them to do the sorting (happy, sad, angry, etc.) and use that statistical data to build a model of facial signaling. And all this before sorting through exactly what "left" and "right" mean in a given political environment.
|
|
|
Post by theevan on Jun 22, 2023 18:03:19 GMT -5
A quick read-through suggests a couple places where this kind of research can go sideways, starting with using the "Face API from Microsoft’s Azure’s Cognitive Services." This from Microsoft's introductory page about that service: Microsoft will be retiring facial recognition capabilities that can be used to try to infer emotional states and identity attributes which, if misused, can subject people to stereotyping, discrimination or unfair denial of services. These include capabilities that predict emotion, gender, age, smile, facial hair, hair and makeup. Existing customers have until June 30, 2023 to discontinue use of these capabilities before they are retired. learn.microsoft.com/en-us/azure/cognitive-services/computer-vision/overview-identityThen there's the question of how exactly facial expressions were characterized and categorized. Our monkey brains do this kind of sorting all the time, but even then our evaluations are hardly infallible. What's a poor algorithm to do? The problem with deriving the human implications of the data is that there has to be a human in the chain somewhere, and the closest thing to a decent validation model I can think of is a large-numbers approach: show the data (pictures of faces) to a very large number of viewers and get them to do the sorting (happy, sad, angry, etc.) and use that statistical data to build a model of facial signaling. And all this before sorting through exactly what "left" and "right" mean in a given political environment. So you're saying you're better looking than the majority?
|
|
|
Post by John B on Jun 22, 2023 18:44:57 GMT -5
|
|
|
Post by TKennedy on Jun 23, 2023 11:41:45 GMT -5
I have long been a student of driving behavior and road rage based on the face of the car.
I contend that being tailgated by a car with an an angry and aggressive face (like a black monster pickup) is much more likely to instill the assumption that the driver is an asshole than would a happy faced vehicle like a VW bug or an Austin Healy Sprite.
|
|
|
Post by Marshall on Jun 23, 2023 11:51:12 GMT -5
In 2015 when I totaled my Corolla, I looked at the new redesigned models. I decided I hated the angry aggressive front ends Toyota was putting out So I bought a Civic instead.
|
|
|
Post by billhammond on Jun 23, 2023 11:52:57 GMT -5
I have long been a student of driving behavior and road rage based on the face of the car. I contend that being tailgated by a car with an an angry and aggressive face (like a black monster pickup) is much more likely to instill the assumption that the driver is an asshole than would a happy faced vehicle like a VW bug or an Austin Healy Sprite.Of course neither of these has enough power to catch up to you and tailgate.
|
|
|
Post by Marshall on Jun 23, 2023 12:00:21 GMT -5
That's all I need to know. Where's my registration link? "STOP THE STEAL ! ! "
|
|
|
Post by TKennedy on Jun 23, 2023 15:22:27 GMT -5
In 2015 when I totaled my Corolla, I looked at the new redesigned models. I decided I hated the angry aggressive front ends Toyota was putting out So I bought a Civic instead. Now THAT'S a happy face!
|
|