Actuarial Outpost
 
Go Back   Actuarial Outpost > Actuarial Discussion Forum > Careers - Employment
FlashChat Actuarial Discussion Preliminary Exams CAS/SOA Exams Cyberchat Around the World Suggestions



Reply
 
Thread Tools Search this Thread Display Modes
  #11  
Old 01-27-2018, 03:21 PM
AbedNadir's Avatar
AbedNadir AbedNadir is offline
Member
CAS SOA
 
Join Date: Mar 2014
Studying for enlightenment
Posts: 2,124
Default

Quote:
Originally Posted by JohnLocke View Post
The fact that needlessly complex solutions to simpler problems can be generated does not imply that there aren't more complex problems in need of more complex models.
i like your way of saying what I said in my post
__________________
Reply With Quote
  #12  
Old 01-27-2018, 03:39 PM
MathStatFin MathStatFin is offline
Member
Non-Actuary
 
Join Date: Jul 2014
Posts: 4,703
Default

I don't think that's "AI". Predicting recidivism isn't a core aspect of human intelligence like being able to differentiate between a cat and a dog.

I don't think people understand the subtle differences between AI and just regular mathematical/statistical modeling.

Quote:
Originally Posted by JMO View Post
Apparently, complex algorithms are not necessarily better than simpler ones. Or even untrained people.
The study is here: https://www.documentcloud.org/docume...Algorithm.html

News articles about it are ubiquitous: https://www.google.com/search?hl=en&...pe=&as_rights=

Based on what I have read, there has been ongoing criticism of a certain proprietary model to predict recidivism, used by judges for setting bail and deciding on sentences. The latest study compared that algorithm with human judgement by means of "mechanical turk" (i.e., online people recruited to do various tasks). The also used a much simpler algorithm involving only age and prior convictions.

When the researchers looked at predictions from two years ago and the results since then, the proprietary algorithm was right about 2/3 of the time. So was the mechanical turk. And so was the far simpler algorithm.
Reply With Quote
  #13  
Old 01-27-2018, 03:40 PM
MathStatFin MathStatFin is offline
Member
Non-Actuary
 
Join Date: Jul 2014
Posts: 4,703
Default

Quote:
Originally Posted by Woodrow View Post
Maybe it depends on the topic, but for a lot of AI, being "as good a untrained people" would be an amazing success.

Think about AI tasks like "recognize a picture of a cat"
good point.
Reply With Quote
  #14  
Old 01-29-2018, 12:02 AM
whoanonstop's Avatar
whoanonstop whoanonstop is offline
Member
Non-Actuary
 
Join Date: Aug 2013
Location: Los Angeles, CA
Studying for Spark / Scala
College: College of William and Mary
Favorite beer: Orange Juice
Posts: 5,845
Blog Entries: 1
Default

If the definition of AI for the majority is different from my definition of AI, does that mean I don't know what AI is?

I don't even know what passes as a complex model anymore.

Is it a regression with an interaction term in it?

Or a deep network mutated with some genetic algorithm scheme?

Not even sure I understand what the term "model" means by itself.

Is it a few numbers entered into Excel? Are spreadsheets alone models?

Or does it actually require something more than that?

I just don't know anymore.

-Riley
__________________
Reply With Quote
  #15  
Old 01-29-2018, 12:04 AM
nonlnear nonlnear is offline
Member
Non-Actuary
 
Join Date: May 2010
Favorite beer: Civil Society Fresh IPA
Posts: 29,851
Default

Quote:
Originally Posted by whoanonstop View Post
If the definition of AI for the majority is different from my definition of AI, does that mean I don't know what AI is?

I don't even know what passes as a complex model anymore.

Is it a regression with an interaction term in it?

Or a deep network mutated with some genetic algorithm scheme?

Not even sure I understand what the term "model" means by itself.

Is it a few numbers entered into Excel? Are spreadsheets alone models?

Or does it actually require something more than that?

I just don't know anymore.

-Riley
A model is complicated when the eyes of the person in control of the project's budget glaze over.
Reply With Quote
  #16  
Old 01-29-2018, 12:11 AM
Colonel Smoothie's Avatar
Colonel Smoothie Colonel Smoothie is online now
Member
CAS
 
Join Date: Sep 2010
College: Jamba Juice University
Favorite beer: AO Amber Ale
Posts: 47,815
Default

Quote:
Originally Posted by whoanonstop View Post
If the definition of AI for the majority is different from my definition of AI, does that mean I don't know what AI is?

I don't even know what passes as a complex model anymore.

Is it a regression with an interaction term in it?

Or a deep network mutated with some genetic algorithm scheme?

Not even sure I understand what the term "model" means by itself.

Is it a few numbers entered into Excel? Are spreadsheets alone models?

Or does it actually require something more than that?

I just don't know anymore.

-Riley
one-line-per-sentence-guy on LinkedIn imo
__________________
Recommended Readings for the EL Actuary || Recommended Readings for the EB Actuary

Quote:
Originally Posted by Wigmeister General View Post
Don't you even think about sending me your resume. I'll turn it into an origami boulder and return it to you.
Reply With Quote
  #17  
Old 02-08-2018, 04:27 PM
rhoucag rhoucag is offline
Member
 
Join Date: Oct 2001
Posts: 274
Default

Bad news for AI and data scientists? It certainly is if they didn't see it coming!
Reply With Quote
  #18  
Old 02-08-2018, 04:32 PM
DoctorNo's Avatar
DoctorNo DoctorNo is online now
Member
SOA AAA
 
Join Date: Dec 2004
Location: Denver, CO
College: Western Washington, Colorado-Boulder
Posts: 13,051
Default

__________________
The opinions of Doctor No do not necessarily represent the opinions of mathematicians or consulting actuaries. Facts cited by Doctor No are not necessarily facts. Find me on Twitter: @NorrisDoug. If you send me a LinkedIn invitation, please let me know who you are (unless it's obvious how I know you).
Reply With Quote
  #19  
Old 02-08-2018, 04:50 PM
Fracktuary Fracktuary is offline
Member
CAS
 
Join Date: Aug 2012
Posts: 1,798
Default

Quote:
Originally Posted by DoctorNo View Post
Jaron Lanier wrote an interesting book on this.

He basically states that these models require absurd amounts of basic human input just to not be catastrophically bad.

All the gains from AI accrue to the model and its owner, and those who provide the input get nothing except the satisfaction of training their replacements.
Reply With Quote
  #20  
Old 02-08-2018, 05:01 PM
Colonel Smoothie's Avatar
Colonel Smoothie Colonel Smoothie is online now
Member
CAS
 
Join Date: Sep 2010
College: Jamba Juice University
Favorite beer: AO Amber Ale
Posts: 47,815
Default

Okay so I've got a question. The last time I tried clicking on random squares just to see if I could proceed, the webpage stopped me and asked me to please click on the correct squares. If these quizzes are being used to tune AI, what are my answers being compared against to determine that I am wrong?
__________________
Recommended Readings for the EL Actuary || Recommended Readings for the EB Actuary

Quote:
Originally Posted by Wigmeister General View Post
Don't you even think about sending me your resume. I'll turn it into an origami boulder and return it to you.
Reply With Quote
Reply

Tags
data science, predictive analytics

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off


All times are GMT -4. The time now is 01:50 PM.


Powered by vBulletin®
Copyright ©2000 - 2018, Jelsoft Enterprises Ltd.
*PLEASE NOTE: Posts are not checked for accuracy, and do not
represent the views of the Actuarial Outpost or its sponsors.
Page generated in 0.20483 seconds with 9 queries