Revision as of 22:52, 21 June 2017 by Jason
All papers, abstracts, and related materials such as slides are fully browsable here. Here's a quick and reasonably complete list though:
- Eisner and Filardo (2011) is the place to start. It is a long book chapter that motivates and explains the language design, and includes example Dyna code to solve a variety of problems.
- Filardo and Eisner (2012) details the first steps towards the new Dyna inference engine.
- Vieira et al. (2017) lays out our plans for using reinforcement learning to auto-tune the execution strategies and data structures as a Dyna program executes.
- Eisner (2009) is not a paper, but a video of a talk, with slides available.
- Eisner (2008) is an extended abstract giving the motivation and goals of the project. There are extensive accompanying slides.
- Eisner, Goldlust & Smith (2005) clearly describes an early restricted version of the language, from an NLP point of view. (A shorter early paper was Eisner, Goldlust & Smith (2004).)
- Blatz & Eisner (2007) is a long paper. It gives a denotational semantics for a somewhat more general version of the language, and then lays out several program transformations on Dyna programs, to turn one algorithm automatically into a more efficient algorithm.
- Cohen et al. (2007) describe another program transformation on Dyna programs; this won the Best Student Paper Award at ICLP'08. (longer tech report).
- www.dyna.org may also be worth a look, but is currently very out of date.
- Eisner et al. (2006) is a short paper about the Dynasty hypergraph browser to examine computation graphs.
See also our list of publications that are not primarily about Dyna, but which use Dyna.