Arrow Research search

Author name cluster

Roxana Girju

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

4 papers
1 author row

Possible papers

4

KR Conference 2014 Short Paper

Tracking Beliefs and Intentions in the Werewolf Game

  • Codruta Girlea
  • Eyal Amir
  • Roxana Girju

We present a new approach to Belief Revision in Situation Calculus. We overcome the need to represent perlocutions by assuming an ’observation model’, describing what beliefs, intentions, and unknown properties the utterances expose. The agents’ belief states are then filtered (Shirazi and Amir 2011) with the observed utterances, resulting in an updated Kripke structure. Our new approach allows us to describe dialogues combined with actions. From a Kriple Structure perspective, we add an observation model of utterances, and a revision model, accounting for how the utterances are used to change beliefs. From a Situation Calculus perspective, we model beliefs over beliefs, belief change, and perlocutionary effects. From a Belief Revision perspective, we represent how utterances affect dialogue participants on an individual level, depeding on each agents’ internal state. Our model does not use any linguistic input. We assume the utterances are already parsed to logical formulas that encode both their propositional content and the speech act they function as. Utterances expressed in natural language may provide further insight into the agents’ beliefs, encoded in presuppositions and propositional attitudes. We plan to embed those phenomena in our model as future work. We propose a model of belief and intention change over the course of a dialogue, in the case where the decisions taken during the dialogue affect the possibly conflicting goals of the agents involved. We use Situation Calculus to model the evolution of the world and an observation model to analyze the evolution of intentions and beliefs. In our formalization, utterances, that only change the beliefs and intentions, are observations. We illustrate our formalization with the game of Werewolf.

AAAI Conference 2000 Short Paper

Domain-Specific Knowledge Acquisition Using WordNet

  • Roxana Girju

In many knowledge intensive applications, it is necessary to have extensive domain-specific knowledge in addition to general-purpose knowledge. This paper presents a methodology for discovering domain-specific concepts and relationships in an attempt to extend WordNet. The method was tested on five seed concepts selected from the financial domain: interest rate, stock market, inflation, economic growth, and employment. Queries were formed with each of these concepts and a corpus of 1000 sentences/seed was extracted automatically from the Internet and the TREC-8 corpora. The system discovered a total of 362 new concepts and 62 new relationships while working in an interactive mode.

v2026.09.13