[ADMB Users] recursive code

[John Sibert]

Anybody have any experience with recursive code in ADMB?

-------- Original Message --------
Subject: 	[New comment] Automatic Differentiation: The most criminally 
underused tool in the potential machine learning toolbox?
Date: 	Tue, 25 Jan 2011 10:45:37 +0000
From: 	Justin Domkes Weblog <no-reply at wordpress.com>
To: 	sibert at hawaii.edu



*Anders S* added a new comment to the post Automatic Differentiation: 
The most criminally underused tool in the potential machine learning 
toolbox? 
<http://justindomke.wordpress.com/2009/02/17/automatic-differentiation-the-most-criminally-underused-tool-in-the-potential-machine-learning-toolbox>.

	


    Anders S
    <http://justindomke.wordpress.com/2009/02/17/automatic-differentiation-the-most-criminally-underused-tool-in-the-potential-machine-learning-toolbox#comment-393>
    said on Automatic Differentiation: The most criminally underused
    tool in the potential machine learning toolbox?
    <http://justindomke.wordpress.com/2009/02/17/automatic-differentiation-the-most-criminally-underused-tool-in-the-potential-machine-learning-toolbox>

January 25, 2011 at 10:45 am

In response to /justindomke/ on February 17, 2009 at 2:08 am:

    I recently got back reviews of a paper in which I used automatic
    differentiation.  Therein, a reviewer clearly thought I was using
    finite difference, or “numerical” differentiation. This has led me
    to wondering: Why don’t machine learning people use automatic
    differentiation more?  Why don’t they use it…constantly? Before
    recklessly speculating on the answer, let me [...]

Awesome blog entry and great links. I think I finally understand 
automatic differentiation and its scope.

One question: How well does it cope (in theory and in practice) with 
recursive functions? I parse function expressions from input files and 
store them in binary tree structures. To evaluate the function I call 
the evaluate() function recursively and then propagate the values from 
the branches to the root.

Could automatic differentiation software create a version of the 
recursive evaluate() to give derivatives? (I guess I could implement the 
ideas manually, I did something similar to create symbolic derivatives.)

See all comments on this post here 
<http://justindomke.wordpress.com/2009/02/17/automatic-differentiation-the-most-criminally-underused-tool-in-the-potential-machine-learning-toolbox#comments>.


WordPress 	

WordPress.com <http://wordpress.com> | Thanks for flying with WordPress!
Manage Subscriptions 
<http://subscribe.wordpress.com/?key=e797a46efd81a101e7118d6e0c02573a&email=sibert%40hawaii.edu> 
| Unsubscribe 
<http://subscribe.wordpress.com/?key=e797a46efd81a101e7118d6e0c02573a&email=sibert%40hawaii.edu&c=uf%2F3JS8PlTk8DGS5IBdj%2FIo%5D2BiUfD20AuaiQJrY2%259> 
| Publish text, photos, music, and videos by email using our Post by 
Email <http://support.wordpress.com/post-by-email/> feature.

*Trouble clicking? Copy and paste this URL into your browser:* 
http://subscribe.wordpress.com