Morgan,
I was good chating with you...
Here is the .tex (and again .pdf) version of the paper.
I also give you the .sty file, so you should be able to recompile it easily if you
want to. 
Feel free to edit anything since on our side we feel we are done.
Also, tell me frankly if you want ot be a co-author on it. I certainly have no problem
with that.
--Herve


---------- Original Message -----------
From: Nelson Morgan <morgan@icsi.berkeley.edu>
To: Herve Bourlard <bourlard@idiap.ch>
Sent: Wed, 27 Oct 2004 11:47:26 -0700
Subject: Re: last version of the paper

> Herve, I read through it carefully and have a moderate number of small 
> corrections (none of them to the formulae, all to the text).
> It would be much easier for me to just fix them rather than list them - 
> can you send me the latex? If you can do this tonight then I can fix 
> them this afternoon (my time) and it will be ready for you in your 
> morning. I will distinguish between the things that are just English and 
> the things that have any substance at all so that you can ignore the 
> former and briefly think about the latter.
> 
> The only "big" thing you might want to think about is this: since you 
> want to fork off a different paper for a journal, are you sure you want 
> to include the hierarchical stuff here, before you have any results?
> 
> Also, do we have our guys trying out the trained transition probability 
> case?
> 
> Herve Bourlard wrote:
> 
> >Here is the last/final version of the DARPA-EARS paper.
> >I intend to send it off tomorow, so you still have a last chance
> >to catch bugs...
> >Thanks,
> >--Herve
> >  
> >
------- End of Original Message -------