texai.org

Note that the system now can only understand and generate a single utterance. Progress towards a broader coverage of English awaits the completion of the Grammar Acquisition Skill, and the Vocabulary Acquisition Skill.

I have completed coding the English Generation Skill and I tested it on my use case utterance “the book is on the table“. As expected, I wrote the supporting Generation Rule Application Library with less effort than its counterpart on the parsing side - Parsing Rule Application Library. The same rule set is used by both libraries, and the method of rule application is similar. In my opinion, this is strong validation of the Fluid Construction Grammar engine that I adopted from Luc Steel’s work here. Continue Reading »

Note that the system now can only understand a single utterance. Progress towards a broader coverage of English awaits the completion of the Grammar Acquisition Skill, and the Vocabulary Acquisition Skill.

I’ve completed writing the English Comprehension Skill, described with graphs here, and tested it on my use case utterance “the book is on the table“. This skill contains, as a component, the parsing rule application library described in my previous system status post. In February and March I completed two additional libraries: one to perform Kintsch spreading activation, and the other to perform discourse elaboration. Both use the AI technique of spreading activation. Accordingly, I factored out a generally useful Java library for spreading activation and released it on the SourceForge project site as a separate package. Continue Reading »

Previously, I posted the bootstrap dialog system design. See the project roadmap to see how this component supports the creation of artificial intelligence. I’m currently writing the parsing rule application library. At this pace, I hope that within a couple of months the system will echo back the single use case sentence that I used in the Incremental Fluid Construction Grammar library. Continue Reading »

The Bootstrap Dialog System is the key deliverable for the Texai project this year. Its use case is described this previous comment. Although this application could be written without fitting into an Albus Hierarchical Control System, I chose to do so because eventually that is the Texai architecture, and spending the extra effort now will avoid a subsequent rewrite and greatly facilitate system elaboration. Continue Reading »

On the SourceForge project site, I just released the Java library for Incremental Fluid Construction Grammar.

Continue Reading »

The Synthetic Teammate is a cognitive model of a pilot flying a UAV in a team setting. The focus of the project is communicating the right information at the right time, through natural language. The project contributors are Jerry Ball and Andrea Heiberg from the U.S. Air Force Research Laboratory. Dr. Ball is the author of the Double R Theory whose grammar the Texai project is adopting. I am now committed to sharing lexical and ontological resources with the Synthetic Teammate project.

As a first step in the collaboration, I created documentation: The Texai Lexicon RDF File Explained, so that collaborators can more easily extract entries for their own projects from the Texai Lexicon.

As the next step I checked out the Synthetic Teammate sources and installed the free Allegro Common Lisp Express Edition for Linux. After fixing a small Lisp bug when launching a shell command on Linux, I was able to parse my use case sentence “the book is on the table”.  The process created the following labeled syntax tree diagram, whose parsing behavior I will implement using my Fluid Construction Grammar Java engine. Click the image for a full size version.

Double R Grammar and the use case sentence are described in Dr. Ball’s paper Double R Grammar.

Fluid Construction Grammar is a natural language parsing and generation system developed by researchers at www.emergent-languages.org. The system features a production rule mechanism for both parsing and generation using a reversible grammar. I just released a Java reimplementation of the original developers’ Lisp implementation. It is available at the SourceForge project site.

There are enough grammar rules developed to parse the sentence “Jill slides blocks to Jack”, creating logical propositions, equivalent to RDF statements as a result. These same propositions can be input to the system to render the original sentence.

Because the original parsing design operates on whole sentences at once, this release will be feature-frozen and development will proceed on a fork called IncrementalFCG. In order for a parser to function in a cognitively plausible manner, it must process an utterance one unit (e.g. word) at a time. I am developing an incremental FCG parser that will process Controlled English utterances left-to-right and word-by-word. The production rules will implement a portion of Jerry Ball’s Double-R grammar for English.

I’ve discovered Fluid Construction Grammar which has both a parser and generator implemented in lisp. I am now writing a java implemenation of the FCG core to replace my current construction grammar and parser. FCG accomodates construction composition using feature structures and unification. One of the FCG authors, Joachim De Beule has offered encouragement.

Progress will be tracked on the project’s SourceForge feature tracker.

At this point in the development of the texai dialog system, it is neccessary to create the bootstrap grammar constructions. These will enable the system to acquire the remainder of the grammar via dialog interactions. I investigated a couple of alternatives before settling on a user-editable construction grammar serialization. The first was to hard-code the construction object creation in the ConstructionInitializer class. This alternative was both tedious and opaque. The second alternative was to represent the construction entities in RDF directly in TriG format, which includes context. This approach is less opaque, but still tedious because RDF entity persistence requires bookkeeping statements. Continue Reading »