As I understand it, Siri is derived from SRI's CALO project (Cognitive Assistant that Learns and Organizes) which is itself a part of SRI's PAL project (Personalized Assistant that Learns). The work from CALO was used as the basis for the Siri startup company that was ultimately acquired by Apple to give them Siri.
SRI's CALO project page no longer exists so most of the links in Wikipedias entry on CALO (http://en.wikipedia.org/wiki/CALO) no longer work but I did find SRI's PAL page that is still live and, despite them saying that one needs to register, the links to papers presented at conferences can be accessed without registration as can some basic information on the project.
For those who are interested, the PAL project page is here: https://pal.sri.com/Plone/framework .
It's pretty detailed, and in some cases turgid stuff, but might be of interest to some people who would like some glimpse as to just how complicated this whole problem area is. This really isn't just another implementation of voice control and, since the Siri developers are presumeably still with Apple and now far better funded than they were before, lots of the issues being tackled for the PAL project will also be being worked on by the Siri team.
If anyone else has any interesting links to publications that are further up the food chain (e.g. ones that Siri researchers presented at conferences before Apple took them over, or stuff directly from the CALO project) then please do post them here because I for one would be interested in reading them, especially higher level architecture overviews rather than highly specific stuff such as "Cascaded Lexicalised Classifiers for Second-Person Reference Resolution" (although stuff like that is interesting too).
- Julian
SRI's CALO project page no longer exists so most of the links in Wikipedias entry on CALO (http://en.wikipedia.org/wiki/CALO) no longer work but I did find SRI's PAL page that is still live and, despite them saying that one needs to register, the links to papers presented at conferences can be accessed without registration as can some basic information on the project.
For those who are interested, the PAL project page is here: https://pal.sri.com/Plone/framework .
It's pretty detailed, and in some cases turgid stuff, but might be of interest to some people who would like some glimpse as to just how complicated this whole problem area is. This really isn't just another implementation of voice control and, since the Siri developers are presumeably still with Apple and now far better funded than they were before, lots of the issues being tackled for the PAL project will also be being worked on by the Siri team.
If anyone else has any interesting links to publications that are further up the food chain (e.g. ones that Siri researchers presented at conferences before Apple took them over, or stuff directly from the CALO project) then please do post them here because I for one would be interested in reading them, especially higher level architecture overviews rather than highly specific stuff such as "Cascaded Lexicalised Classifiers for Second-Person Reference Resolution" (although stuff like that is interesting too).
- Julian
Last edited: