|
This page includes abstracts and handouts for talks presented at conferences and workshops, which have not been written up as papers, organized by topic.
Computational Linguistics
Identification of subculture groups and its reflection on topic propagation in the Blogosphere (with James Davidson)
- Presented at Advances in Human Factors and Ergonomics 2012
,San Francisco, CA, July 2012.
Our investigation of language use and link structures in the Persian blogosphere shows that homophily in this network can develop along gender and ethnic lines. But more importantly, it was noted that although value homophily is crucial in the formation of the progressive/reformist social group, gender homophily governs the formation of the conservative/fundamentalist subculture. These results suggest that the connections in the online network groups in the Persian blogosphere directly reflect the ideological separations that exist in the physical society along gender and ethnic lines, which will in turn affect the
propagation of information and topics in the network.
- [abstract][long abstract (pdf)]
-
Computing linguistic creativity in the Persian blogosphere
- Invited talk; presented at the Department of Computer and Information Science
, at Linköpings Universitet, Sweden; June 2011.
Linguistic creativity - such as coinage of new words, creation of novel metaphors, play with words, and new usages of non-standard spelling - is an important property of human language. The automatic identification of instances of linguistic creativity and the computation of their meaning represent an important challenge for a variety of Natural Language Processing tasks. Social media such as blogs provide new insights into the use of linguistic creativity as bloggers' writings more directly represent modern language than formal texts found in traditional websites and news sources. In this talk, I will present a study of neologisms (newly created words) and metaphors used in Persian blog language and will discuss methods for their automatic recognition. When relevant, comparisons to Blogspeak in other languages (e.g., English and Chinese) will also be discussed.
- [abstract]
Automatic topic detection in Persian blogs (with Ali Hadjarian)
- Presented at Human Social Cultural Behavioral Modeling (HSCB) Focus 2011
,Chantilly, Virginia; February 2011.
- [abstract][slides]
MITRE Infrastructure and Evaluation for Identity Resolution Technology (with Keith Miller, Elizabeth Schroeder, Sarah McLeod, Azar Ulrich, James Finley, Gail Hamilton, Andre Milota, Ken Samuel, Sherri Condon and Mark Arehart)
- 77th MORS Symposium. Fort Leavenworth, Kansas. June 2009.
Across the government, analysts must search multiple data sources in order to create a composite view of information about a person. This process is time consuming, difficult, and it is impossible for a human to carry out thoroughly when dealing with large amounts of data and shared data sources. Various systems claim to offer such identity matching and resolution capability as an automatic function. Before incorporating such systems into an analyst workflow or an information sharing environment it is important to determine how well those systems perform. This research focuses on developing an appropriate set of metrics to use in evaluating identity matching technologies, and creating an infrastructure and test data that can be brought to bear on performing such evaluations. The project also investigates maintaining privacy when sharing information and extending system evaluations to incorporate privacy concerns. The evaluation infrastructure is extensible to multiple use cases, such as sharing data across agencies or resolving identities in a single data source, and enables an objective comparison of different identity resolution systems to determine which is most appropriate for a given mission.
- [abstract]
Method for comparing Identity Resolution Technology (with Keith Miller, Elizabeth Schroeder, Sarah McLeod, Azar Ulrich, James Finley, Gail Hamilton, Andre Milota, Ken Samuel, Sherri Condon and Mark Arehart)
- 77th MORS Symposium. Fort Leavenworth, Kansas. June 2009.
With vast amounts of information currently available digitally, it is difficult to sift through the data and discover important details. One particularly difficult task is identifying "who is who" in different data sources and being able to recognize when the same person appears multiple times throughout. Various systems claim to offer this as an automatic capability, known as identity matching or resolution, yet it can be difficult to determine exactly how well those systems perform the reported functions. Building on previous work in evaluating and comparing name and record matching tools, this study addresses the problems involved with the evaluation of identity matching and resolution technologies. To develop an objective means of comparing alternative identity resolution technologies, this research focuses on determining appropriate evaluation metrics. The research also involves creating an infrastructure and developing test data that can be brought to bear on performing such evaluations. The evaluation infrastructure is extensible to multiple use cases, such as sharing data across agencies or resolving identities in a single data source. As a result, a user is able to define their mission, run each identity resolution system using applicable test data, and then calculate the evaluation metrics for each system to determine which produced the best score.
- [abstract]
Machine Translation of Persian Complex Predicates (with Jan W. Amtrup)
- Talk presented at the Second International Conference on Iranian Linguistics in Hamburg, Germany; August 2007.
- [abstract][slides]
Heritage Persian and Language Teaching
Computational tools for learners of Persian
- Invited talk; presented at the STARTALK Persian Teachers Development Workshop
, at the University of Pennsylvania, Philadelphia; July 2010.
This talk presented information about existing computer tools that can be used by Persian language teachers in the classroom or by Persian language learners who are doing research, preparing reports, or doing exercises. Most of these tools are freely available but we have also listed a few tools that are in preparation and are not yet available to the public. The latter have been included in this list to show the potential for Persian language tools in the coming future. See "website" link for the presentation.
- [abstract]
[website]
Heritage Persian Characteristics and Needs
- Invited talk; presented at the STARTALK Persian Teacher's Workshop for Professional Curriculum and Materials Development
, at the University of Pennsylvania, Philadelphia; June 2009.
The talk presents an overview of the linguistic and functional competence of Heritage and L2 speakers or Persian, based on recent research in the field, and shows that they have different characteristics. The talk covers issues in lexicon, morphology, syntax and literacy. I argue that instruction should build upon the strengths of heritage learners and address their specific needs.
- [abstract]
[slides]
Difficulties for Learning Persian as a Second Language
- Keynote speech; presented at the Workshop for Teachers of Persian, at the University of Maryland, College Park; January 25-26, 2008.
In this talk, I present an overview of the various factors that raise difficulties for a speaker of English learning Persian. Persian and English are both Indo-European languages and thus share a number of linguistic features, yet a number of issues make Iranian Persian and Dari Persian more difficult than more commonly taught languages. These include the Arabic-based writing system, lack of cognate vocabulary, morphological patterns borrowed from Arabic, verb-final word order with relatively free movement, and light verb constructions. The effects of interference from English and diglossia are also discussed.
I argue, however, that many of the difficulties of the language can be simplified for students if we use linguistic insight to help the learners detect the existing patterns. This is demonstrated using the properties of plural formation and the patterns found in the choice of the light verb. Hence, what may seem difficult often follows distinct patterns that can be generalized, thus helping students avoid the need for rote learning methods. In addition, the talk presents an introduction to computational linguistics with applications for Persian and shows how a number of computational systems could be used in the classroom to facilitate both teaching and learning by helping to discover language patterns in authentic media or by generating automated feedback to the student.
- [abstract]
[slides]
[pdf (with comments) ]
Testing the Linguistic Competence of Persian Heritage Speakers (with Ilhan Cagri and Scott Jackson)
- Talk presented at the Second Language Research Forum: Colloquium on Understanding the interlanguage systems of heritage language learners, in Urbana-Champaign, Illinois; October 2007.
- [abstract]
[slides]
Complex Predicates
Preverbal nominals and telicity in Persian complex predicates
- Talk presented at the International Conference on Complex Predicates in Iranian Languages in Paris, France; July 2008.
-
[abstract]
[handout]
On Incorporation in Persian Verbal Predicates
- Talk presented at the Second International Conference on Iranian Linguistics in Hamburg, Germany; August 2007.
-
[abstract][handout]
A Typology of Persian Complex Predicates
- Talk presented at the International Conference on Aspects of Iranian Linguistics in Leipzig, Germany; June 2005.
-
[abstract][handout]
Reflexive Verbs and Voice Morphology in Eastern Armenian
- Talk presented at the Workshop on Armenian Syntax, Pithiviers, France, May 2005.
- [abstract]
Aspect in Complex Predicates
- Talk presented at the Workshop on Complex Predicates, Particles and Subevents, University of Konstanz, Germany. Sept. 30-Oct. 2, 2002.
- [abstract][handout]
Primitive Elements of Verbal Predicates: Evidence from Persian
- Talk presented at the Workshop on Syntax of Predication, ZAS, Berlin, Germany. Nov. 2-3, 2001.
- [handout]
Theories of Verb Formation
- Invited lecture, presented at New Trends in Linguistics Workshop, Allameh Tabatabai University, Tehran, Iran. May 27-29, 2001.
- [handout]
Clitics, Case and vP Structure
Focus and the Auxiliary in Eastern Armenian (based on work done with Arsalan Kahnemuyipour)
- Presented at the 37th Annual Meetin of the Berkeley Linguistics Society (BLS), special session on Languages of the Caucasus. Bekeley, CA. February 2011.
The auxiliary verb in Eastern Armenian is a clitic that carries tense and agreement and is used to form all the indicative forms of the verb with the exception of the aorist tense. Although the auxiliary typically follows the participle, it is also able to be fronted and attached to various, seemingly unrelated elements such as the direct object, the manner adverb, the first component of a compound verb, negation, and questioned elements or wh-phrases in the clause. The close relationship between the auxiliary and contrastive focus has been noted by several researchers (Comrie 1984, Tamrazian 1994, Tragut 2009). A purely focus-based analysis of the auxiliary, however, fails to explain the obligatory fronting of the auxiliary in neutral sentences involving nonspecific objects, measure adverbs, and preverbs.
The goal of this paper is to account for the puzzling positional distribution of the auxiliary clitic. A closer investigation reveals that the elements that host the auxiliary in neutral sentences occupy the leftmost position within the vP domain, while the marked elements are in a focus position and are structurally higher in the sentence (cf. Kahnemuyipour and Megerdoomian 2010). The paper then contrasts the Armenian data to the mobile auxiliary clitics in Udi (Harris 2002) and Talyshi (Stilo 2008) where similar phenomena have been noted, pointing to the possibility of focus-marking auxiliaries as an areal feature in the southern Caucasus.
- [abstract]
[handout]
Some bare nouns are more equal than others
- Invited talk presented at the USC Student Workshop 2008, Los Angeles; May 7, 2008.
- [abstract][handout]
Parallel Domains and the Projection of Arguments
- USC-GSIL Student Workshop. University of Southern California, Los Angeles. April 29, 2002.
- [handout]
Wh-phrases
D-Linked Wh-Phrases and Focus-Fronting in Persian (with Shadi Ganjavi)
- Talk presented at the Third International Conference on Iranian Linguistics in Paris, France; September 2009.
- [abstract][handout]
Morphology-Phonology Interface
The Derivation/Inflection Distinction and Post-Syntactic Merge
(with Arsalan Kahnemuyipour)
- Talk presented at the Canadian Linguistic Association, University of Toronto, Canada. May 25-28, 2002.
- [abstract][handout]
|