Ani Nenkova
Associate professor
University of Pennsylvania
Computer and Information Science
3330 Walnut Street
Philadelphia, PA 19104

Office:  Levine 505
Phone: (215) 898-8745
Fax:      (215) 898-0587

The Penn Engineering Magazine ran an article on my work in Spring 2017.

Ani Nenkova is an associate professor of computer and information science at the University of Pennsylvania. Her main areas of research are computational linguistics and artificial intelligence, with emphasis on developing computational methods for analysis of text quality and style, discourse, affect recognition and summarization. She obtained her PhD degree in computer science from Columbia University.

Ani and her collaborators are recipients of the best student paper award at SIGDial in 2010 and best paper award at EMNLP-CoNLL in 2012. The Penn team co-led by Ani won the audio-visual emotion recognition challenge (AVEC) for word-level prediction in 2012.

Ani is a co-editor-in-chief of the Transactions of the Association for Computational Linguistics (TACL). She was a member of the editorial board of Computational Linguistics (2009--2011) and an associate editor for the IEEE/ACM Transactions on Audio, Speech and Language Processing (2015--2018). She regularly serves as an area chair/senior program committee member for ACL, NAACL and AAAI. Ani was a program co-chair for SIGDial 2014 and NAACL-HLT in 2016.

Owen Rambow and I wrote a report on our work as program co-chairs for NAACL-HLT 2016. We made important changes to the structure of the conference areas and the process for recruiting reviewers. We hope some of these successful changes will be preserved in future editions of the conference.

[ Research | Downloads | Google Scholar | DBLP | Older Publications ]

Recent highlights

Word Embeddings (Also) Encode Human Personality Stereotypes, Agarwal et al, *SEM@NAACL-HLT 2019.

How to Compare Summarizers Without Target Length? Pitfalls, Solutions and Re-Examination of the Neural Summarization Literature, Simeng Sun, Ori Shapira, Ido Dagan, Ani Nenkova, To appear at the Workshop on Methods for Optimizing and Evaluating Neural Language Generation at NAACL 2019.

Predicting Annotation Difficulty to Improve Task Routing and Model Performance for Biomedical Information Extraction, Yang et al, NAACL-HLT 2019.

A Corpus with Multi-Level Annotations of Patients, Interventions and Outcomes to Support Language Processing for Medical Literature, Nye et al, ACL 2018.

Combining Lexical and Syntactic Features for Detecting Content-Dense Texts in News, Yang and Nenkova, JAIR. 60: 179-219 (2017).

Fast and Accurate Prediction of Sentence Specificity, Li and Nenkova, AAAI 2015.

Prosodic cues for emotion: analysis with discrete characterization of intonation, Cao et al, Speech prosody, 2014.


How can NLP help improve access to the medical literature?

The efficacy of medical interventions is most meaningfully assessed in Randomized Control Trials (RCTs), in which groups of people are randomly assigned to undergo the intervention or a comparison treatment. Clinical outcomes are measured and compared between groups, to decide which treatment leads to better clinical outcomes. Access to the growing number of medical papers describing the results from RCTs however are difficult to access using free text search. In many cases, patients and medical practitioners alike in fact do not have a completely formulated search query, but instead may want to browse all studied treatments for a given condition, or all RCTs that have measured a given outcome of interest. To organize the medical literature according to condition, intervention and outcome would require a certain level of automation.

My collaborators and I oversaw the creation of EBM-NLP, an annotated crowdsourced corpus of about 5,000 abstracts of medical papers reporting the findings of RCTs. We created simple annotation guidelines that could be carried out by people with no specific medical training, to identify the spans of text describing the patients enrolled in an RCT, the medical interventions compared in the trial and the outcome measures used as basis for the comparison [ACL'18]. This corpus allows researchers to develop meaningful supervised models for span extraction from medical documents, which is a necessary step towards organizing the medical literature in a browsable interface.

>Creating annotated data in a specialized domain leads to questions that do not arise in typical language technology research. We were particularly interested in developing ways to predict if crowdsourced annotations would be of acceptable quality and how we can decide if a given abstract should be annotated by an expert or via crowdsourcing. To probe these questions, we developed a model to predict the difficulty of an abstract [NAACL'19]. Discarding crowdsourced annotations for difficult abstracts increases extraction performance and cost and resources can be optimized if difficulty is used to route annotations to experts. For extracting interventions, using a smaller expert-annotated corpus containing a mix of difficult and typical abstracts leads to better performance that more than double the size corpus of crowdsourced annotations.

What is a good news summary?

A good summary conveys the main points of a text much more succinctly. So if language technologies worked perfectly, automatic evaluation of summaries will be trivial, reduced to computing which summary is most similar to the original text in meaning. Even at the time when text meaning was approximated by counting word overlap, this intuition was meaningful for automatically estimating summary quality [CL'13]. A key insight of this work is that we do not need human reference summary in order to estimate the quality of a machine summaries. Moreover in our work we have shown that using several machine summaries as reference is even more useful in automatically approximating people's judgements of summary informativeness. Our most recent result show that neural representations of meaning hold similar potential for evaluation and that using human references and word overlap is only due to the inertia of habit in research [EMNLP'19, short].

When summaries are of different length, longer ones are likely to be judged as more informative. These judgements however do not directly reflect the balance between informativeness and time spent to get the information. We have proposed a length normalization method that provides a more meaningful comparison between systems of different length [NeuralGen'19].

What aspects of text create the impression of smooth flow from sentence to sentence in a well-written text?

Dominant theories in computational linguistics fail to provide an account for a large portion of sentence-to-sentence transitions even in highly polished and edited newspaper text [NAACL'10]. My students and I have proposed novel accounts of local text coherence, modeling writer's intent [EMNLP'12] and sentence specificity [IJCNLP'11, AAAI'15].

Our model of writer intent draws from the intuition that syntactic structure would give strong indication of the discourse purpose of a sentence. The model uses no lexical information at all. It is able to differentiate text quality in the science journalism domain [TACL'13], i.e. it detects statistically significant differences in the organization of regular science articles from the New York Times and science articles from recognition-winning science writers [LN12]. The model is also able to detect differences in the organization of abstracts of academic papers (from the computational linguistics domain) published in top venues and those published at workshops [EMNLP'12]. The strong connection between syntactic form and quality and type of text [EACL'09, NetAl'10, AAAI'14, EMNLP'15] is currently not well-understood and my ongoing work aims to clarify the mechanisms via which they interact.

Our other line of investigation into the missing components of local coherence and general impressions of writing quality revolves around the balance between specific and general information in a text. We have built tools to quantify sentence specificity [IJCNLP 2011, AAAI 2015]. We have also shown that changes in sentence and overall text specificity are strongly associated with perception of text quality: great science writing is overall more general than regular NYT science writing and great articles contain fewer stretches of specific content [LN'12]. Automatic summaries, which are often judged to be incoherent, are significantly more specific than human-written summaries for the same events [LN'11]. Sentence specificity is also more robust than sentence length as indicator of which sentences may pose comprehension problems and need to be simplified for given audiences [AAAI'15].

Where is emotion encoded in speech?

One of the conundrums in emotion recognition from voice is from what regions of speech one should extract features for the task: full utterance, individual words or even frames. The accompanying question is whether to represent the regions with high-level descriptions of prosodic events or low-level descriptors of the acoustic signal.

My collaborators and I have extensively studied these questions. We adopted an approach that relies on regions of interest related to properties of phoneme [SC'10] or word [SCLa'15] classes. We compute a separate representation for all frames in an utterance that fall in each region. This approach gives a robust improvement over other approaches for representing speech for emotion recognition, in both acted and spontaneous speech [SC'10, SCLa'15].

Another issue in emotion recognition is accounting for the fact that some speakers are simply more expressive than others. We have developed a speaker-independent approach in which no subject is seen in both training and testing, but which is nevertheless speaker-sensitive because each speaker is treated as a query in a ranking problem. The task is to identify, in a collection of samples from the same speaker, the utterances that are most likely to express a given emotion. Then the ranking of utterances is used as features for predicting the most likely emotion conveyed by the utterance. The method is not designed to work in real time but to analyze recorded speech or interactions, in which a sample of speech from the same speaker is available. The method leads to significant improvement over standard classification approaches on acted emotion datasets, and is much more accurate in finding utterances conveying pure emotion in spontaneous speech where neutral utterances form the dominant class. Furthermore, our approach combines well with standard approaches to improve overall recognition accuracy [SLPb'15].

We have also examined the utility in emotion recognition of identifying discrete prosodic events, such as pauses and pitch accent types. This question was inspired by work in emotion recognition from visual features, where discrete facial configurations such as lip, brow or cheek configurations are often used instead of general representations of the face. We found strong evidence that detection of discrete events in speech would complement existing feature representations [SP'14].


Speciteller: A tool for predicting sentence specificity.

CATS: The corpus of science journalism articles used for our TACL 13 paper.

SumRepo04: A Repository of State of the Art and Competitive Baseline Summaries for Generic News Summarization on DUC 2004. Introduced in our LREC 2014 paper.

Newsworthy and personal perspective verbs from our NAACL'15 short paper.

General specific annotations: Sentence-level annotations of specificity used for our work reported in IJCNLP '11 and ACL workshop '11 papers.

SIMetrix (Summary Input similarity Metrics): Tool to perform the automatic summary evaluation introduced in our EMNLP'09 and CL'14 papers. The tool implements the various input-summary similarity measures investigated in that work.

Accent ratio dictionaries from our work on word prominence is spoken utterances (NAACL'07). This one is derived from conversational speech and this one from read news.


Junyi Jessy Li, Ani Nenkova, Detecting Content-Heavy Sentences: A Cross-Language Case Study, EMNLP 2015: 1271-1281

Junyi Jessy Li and Ani Nenkova, Fast and Accurate Prediction of Sentence Specificity, The 29th AAAI Conference on Artificial Intelligence (AAAI), pp. 2281-2287, 2015.

Benjamin Nye and Ani Nenkova, Identification and Characterization of Newsworthy Verbs in World News, Conference of the North American Chapter of the Association for Computational Linguistics - Human Language Technologies (NAACL HLT), short paper, 2015.

Houwei Cao, Ragini Verma and Ani Nenkova, Speaker-sensitive Emotion Recognition via Ranking: Studies on Acted and Spontaneous Speech, Computer Speech and Language, Special Issue: Next Generation Paralinguistics, 29 (1), 186-202, 2015.

Houwei Cao, Arman Savran, Ragini Verma and Ani Nenkova, Acoustic and lexical representations for affect prediction in spontaneous conversations, Computer speech and language 29 (1), 203-217, 2015.

Kai Hong, Ani Nenkova, Mary March, Amber Parker, Ragini Verma and Christian Kohler, Lexical use in emotional autobiographical narratives of persons with schizophrenia and healthy controls, Psychiatry research 225 (1-2), 40-49, 2015.

Arman Savran, Houwei Cao, Ani Nenkova and Ragini Verma, Temporal Bayesian Fusion for Affect Sensing: Combining Video, Audio, and Lexical Modalities, IEEE Transactions on Cybernetics, 2015 Sep;45(9):1927-41.

Ellie Pavlick and Ani Nenkova, Inducing Style for Paraphrase and Genre Differentiation, Conference of the North American Chapter of the Association for Computational Linguistics - Human Language Technologies (NAACL HLT), short paper, 2015.

Kai Hong, Mitchell Marcus, Ani Nenkova, System Combination for Multi-document Summarization, EMNLP 2015: 107-117.

Yinfei Yang and Ani Nenkova, Detecting Information-Dense Texts in Multiple News Domains, The Twenty-Eighth Conference on Artificial Intelligence (AAAI), pp. 1650- 1656, 2014.

Junyi Jessy Li and Ani Nenkova, Reducing Sparsity Improves the Recognition of Implicit Discourse Relations, Special Interest Group on Discourse and Dialogue (SIGDIAL), pp 199-207, 2014.

Junyi Jessy Li and Ani Nenkova, Addressing Class Imbalance for Improved Recognition of Implicit Discourse Relations, Special Interest Group on Discourse and Dialogue (SIGDIAL), pp. 142-150, 2014.

Junyi Jessy Li, Marine Carpuat and Ani Nenkova, Assessing the Discourse Factors that Influence the Quality of Machine Translation, Annual Meeting of the Association for Computational Linguistics (ACL), short paper, pp. 283-288, 2014.

Junyi Jessy Li, Marine Carpuat and Ani Nenkova, Cross-lingual Discourse Relation Analysis: A corpus study and a semi-supervised classification system, International Conference on Computational Linguistics (COLING), pp. 577-587, 2014.

Houwei Cao, Stefan Benus, Ruben Gur, Ragini Verma, and Ani Nenkova, Prosodic cues for emotion: analysis with discrete characterization of intonation, Speech Prosody, pp. 130-134, 2014.

Kai Hong, John Conroy, Benoit Favre, Alex Kulesza, Hui Lin and Ani Nenkova, A Repository of State of the Art and Competitive Baseline Summaries for Generic News Summarization, Ninth International Conference on Language Resources and Evaluation (LREC), pp. 1608-1616.

Houwei Cao, David Cooper, Michael Keutmann, Ruben Gur, Ani Nenkova and Ragini Verma, CREMA-D: Crowd-Sourced Emotional Multimodal Actors Dataset, IEEE Transactions on Affective Computing, 5 (4), 377 - 390, 2014.

Annie Louis and Ani Nenkova, Verbose, Laconic or Just Right: A Simple Computational Model of Content Appropriateness under Length Constraints, Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics EACL, pp. 636-644, 2014.

Kai Hong and Ani Nenkova, Improving the Estimation of Word Importance for News Multi-Document Summarization, Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics (EACL), pp. 712-721, 2014.

Annie Louis and Ani Nenkova, A Text quality corpus for science journalism, Dialogue and Discourse, Special Issue on Annotating Pragmatic and Discourse Phenomena, 4(2): 8 7-117, 2013.

Annie Louis and Ani Nenkova, What Makes Writing Great? First Experiments on Article Quality Prediction in the Science Journalism Domain, Transactions of the Association for Computational Lingustics (TACL), 1(July):341-352, 2013.

Annie Louis and Ani Nenkova Automatically assessing machine summary content without a gold-standard Computational Linguistics, 39 (2): 267-300, 2013.

Peter A. Rankel, John M. Conroy, Hoa Trang Dang and Ani Nenkova, A Decade of Automatic Content Evaluation of News Summaries: Reassessing the State of the Art ACL 2013, short paper.

Miraj Shah, David Cooper, Houwei Cao, Ruben Gur, Ani Nenkova and Ragini Verma, Action Unit Models of Facial Expression of Emotion in the Presence of Speech, Fifth biannual Humaine Association Conference on Affective Computing and Intel ligent Interaction, ACII 2013.

Benoit Favre, Kyla Cheung, Siavash Kazemian, Adam Lee, Yang Liu, Cosmin Munteanu, Ani Nenkova, Dennis Ochei, Gerald Penn, Stephen Tratz, Clare Voss and Frauke Zeller, Automatic Human Utility Evaluation of ASR Systems: Does WER Really Predict Performance?, INTERSPEECH 2013.

Arman Savran, Houwei Cao, Miraj Shah, Ani Nenkova, Ragini Verma, Combining Audio, Video and Lexical Indicators of Affect in Spontaneous Conversation via Particle Filtering, Proceedings of the 14th ACM international conference on Multimodal interaction (ICMI), 2012. [Winner of the AVEC'12 word-level emotion prediction challenge].

Annie Louis, Ani Nenkova, A coherence model based on syntactic patterns, EMNLP-CoNLL 2012 [Best paper award].

Kai Hong, Christian G. Kohler, Mary E. March, Amber A. Parker, Ani Nenkova, Lexical Differences in Autobiographical Narratives from Schizophrenic Patients and Healthy Controls, EMNLP-CoNLL 2012.

Houwei Cao, Ragini Verma, Ani Nenkova, Combining Ranking and Classification to Improve Emotion Recognition in Spontaneous Speech, INTERSPEECH 2012.

Rivka Levitan, Agustin Gravano, Laura Willson, Stefan Benus, Julia Hirschberg, Ani Nenkova, Acoustic-Prosodic Entrainment and Social Behavior, NAACL-HLT 2012.

Karolina Owczarzak, John M. Conroy, Hoa Trang Dang, Ani Nenkova, An Assessment of the Accuracy of Automatic Evaluation in Summarization, Proceedings of the NAACL-HLT 2012 Workshop on Evaluation Metrics and System Comparison for Automatic Summarization.

Annie Louis, Ani Nenkova, A corpus of general and specific sentences from news , LREC 2012.

Ani Nenkova, Kathleen McKeown, A Survey of Text Summarization Techniques, Chapter in Mining Text Data, pp 43-76, 2012.

Libo Sun, Alexander Shoulson, Pengfei Huang, Nicole Nelson, Wenhu Qin, Ani Nenkova and Norman Badler, Animating synthetic dyadic conversations with variations based on context and agent attributes , Journal of Visualization and Computer Animation 23(1): 17-32, 2012.

Advaith Siddharthan, Ani Nenkova and Kathleen McKeown, Information Status Distinctions and Referring Expressions: An Empirical Study of References to People in News Summaries , Computational Linguistics 37(4): 811-842, 2011.

Ani Nenkova and Kathleen McKeown, Automatic Summarization [pdf]
Foundations and Trends in Information Retrieval, Vol 5, No 2-3, pp. 103-233.
[A printed and bound version of this article is available at a special discount price of US$35 \ from Now Publishers. This can be obtained by entering the promotional code INR015015 on the order form at now publishers.]

Agustin Gravano, Rivka Levitan, Laura Willson, Stefan Benus, Julia Hirschberg, Ani Nenkova, Acoustic and Prosodic Correlates of Social Behavior, Interspeech 2011.

Annie Louis and Ani Nenkova, Automatic identification of general and specific sentences by leveraging discourse annotations, Proceedings of 5th International Joint Conference on Natural Language Processing (IJCNLP), 2011.

Annie Louis and Ani Nenkova, Text specificity and impact on quality of news summaries, Proceedings of ACL-HLT Workshop on Monolingual Text to Text Generation, 2011.

Annie Louis, Ani Nenkova, Creating Local Coherence: An Empirical Assessment HLT-NAACL 2010: 313-316, 2010.

Annie Louis, Aravind Joshi and Ani Nenkova, Discourse indicators for content selection in summarization, Proceedings of SIGDIAL 2010. [Best student paper]

Annie Louis, Aravind Joshi, Rashmi Prasad and Ani Nenkova, Using entity features to classify implicit discourse relations, Proceedings of SIGDIAL 2010.

Emily Pitler, Annie Louis and Ani Nenkova, Automatic Evaluation of Linguistic Quality in Multi-Document Summarization, ACL 2010.

Dmitri Bitouk, Ragini Verma and Ani Nenkova,Class-level spectral features for emotion recognition, Speech Communication, Volume 52, Issues 7-8, July-August 2010, Pages 613-625.

Ani Nenkova, Jieun Chae, Annie Louis and Emily Pitler, Structural Features for Predicting the Linguistic Quality of Text: Applications to Machine Translation, Automatic Summarization and Human-Authored Text, In Emiel Krahmer and Mariet Theune, editors, Empirical Methods in Natural Language Generation: Data-oriented Methods and Empirical Evaluation, 2010.

Annie Louis and Ani Nenkova, Creating Local Coherence: An Empirical Assessment, NAACL-HLT 2010, short paper.

Annie Louis and Ani Nenkova, Automatically Evaluating Content Selection in Summarization without Human Models, EMNLP'09.

Emily Pitler, Annie Louis and Ani Nenkova, Automatic sense prediction for implicit discourse relations in text, ACL-IJCNLP 2009.

Emily Pitler and Ani Nenkova, Using Syntax to Disambiguate Explicit Discourse Connectives in Text, ACL-IJCNLP 2009, short paper.

Dmitri Bitouk, Ani Nenkova Ragini Verma, Improving Emotion Recognition using Class-Level Spectral Features, INTERSPEECH 2009.

Annie Louis and Ani Nenkova, Performance Confidence Estimation for Automatic Summarization, EACL 2009.

Jieun Chae and Ani Nenkova, Predicting the Fluency of Text with Shallow Structural Features: Case Studies of Machine Translation and Human-Written Text, EACL 2009.

Emily Pitler and Ani Nenkova, Revisiting Readability: A Unified Framework for Predicting Text Quality, EMNLP, 2008.

Emily Pitler, Mridhula Raghupathy, Hena Mehta, Ani Nenkova, Alan Lee, Aravind Joshi, Easily Identifiable Discourse Relations, COLING 2008, short paper.

Ani Nenkova, Agustin Gravano, Julia Hirschberg, High frequency word entrainment in spoken dialogue , ACL:HLT 2008, short paper.

Ani Nenkova and Annie Louis, Can you summarize this? Identifying correlates of input difficulty for generic multi-document summarization, ACL:HLT 2008.

Vivek Kumar Rangarajan Sridhar, Ani Nenkova, Shrikanth Narayanan, Dan Jurafsky, Detecting prominence in conversational speech: pitch accent, givenness and focus, 4th Conference on Speech Prosody 2008.

Ani Nenkova, Entity-driven Rewrite for Multi-document Summarization, The Third International Joint Conference on Natural Language Processing (IJCNLP08).

Ani Nenkova and Dan Jurafsky, Automatic Detection of Contrastive Elements in Spontaneous Speech, IEEE workshop on Automatic Speech Recognition and Understanding (ASRU) 2007.

Volker Strom, Ani Nenkova, Robert Clark, Yolanda Vazquez-Alvarez, Jason Brenier, Simon King, Dan Jurafsky, Modelling Prominence and Emphasis Improves Unit-Selection Synthesis, Intrespeech 2007.

Surabhi Gupta, Ani Nenkova and Dan Jurafsky, Measuring Importance and Query Relevance in Topic-focused Multi-document Summarization, ACL 2007, short paper.

Ani Nenkova, Jason Brenier, Anubha Kothari, Sasha Calhoun, Laura Whitton, David Beaver, Dan Jurafsky, To Memorize or to Predict: Prominence Labeling in Conversational Speech , NAACL-HLT 2007.

Ani Nenkova, Rebecca Passonneau, Kathleen McKeown, The pyramid method: incorporating human content selection variation in summarization evaluation , ACM Transactions on Speech and Language Processing, volume 4, issue 2, 2007.

Lucy Vanderwende, Hisami Suzuki, Chris Brockett, Ani Nenkova, Beyond SumBasic: Task-Focused Summarization with Sentence Simplification and Lexical Expansion , Information Processing and Management, Special issue on summarization volume 43, issue 6, 2007.

Jason Brenier, Ani Nenkova, Anubha Kothari, Laura Whitton, David Beaver, Dan Jurafsky, The (Non)Utility of Linguistic Features for Predicting Prominence in Spontaneous Sp eech , IEEE/ACL 2006 Workshop on Spoken Language Technology, 2006.

Ani Nenkova, Summarization Evaluation for Text and Speech: Issues and Approaches , INTERSPEECH 2006.

Ani Nenkova, Lucy Vanderwende, Kathleen McKeown, A compositional context sensitive multi-document summarizer: exploring the factors that influence summarization , SIGIR 2006.

Ani Nenkova, Understanding the process of multi-document summarization: content selection, rew\ rite and evaluation, PhD Thesis, Columbia University, January 2006.

Ani Nenkova, Advaith Siddharthan and Kathleen McKeown, Automatically learning cognitive status for multi-document summarization of newswire, HLT/EMNLP 2005.

Ani Nenkova, Automatic Text Summarization of Newswire: Lessons Learned from the Document Understanding Conference, 20th National Conference on Artificial Intelligence (AAAI 2005).

Ani Nenkova, Discourse Factors in Multi-Document Summarization, 10th Annual AAAI/SIGART Doctoral Consortium at AAAI'05.

Kathleen McKeown, Rebecca Passonneau, David Elson, Ani Nenkova, Julia Hirschberg, Do Summaries Help? A Task-Based Evaluation of Multi-Document Summarization, 28th Annual International ACM SIGIR Conference on Research and Development in Information Retreival (SIGIR 2005).

Aaron Harnly, Ani Nenkova, Rebecca Passonneau, Owen Rambow, Automation of summary evaluation by the pyramid method, RANLP-2005.

Advaith Siddharthan, Ani Nenkova and Kathleen McKeown, Syntactic Simplification for Improving Content Selection in Multi-Document Summarization, 20th International Conference on Computational Linguistics (COLING 2004).

Julia Hirschberg, Agustin Gravano, Ani Nenkova, Elisa Sneed, Gregory Ward, Intonational overload: uses of the H* !H* L- L% contour in read and spontaneous speech, Labphon-9.

Ani Nenkova and Rebecca Passonneau, Evaluating Content Selection in Summarization: the Pyramid Method, NAACL-HLT 2004.

Rebecca J. Passonneau and Ani Nenkova, Evaluating Content Selection in Human- or Machine-Generated Summaries: The Pyramid Scoring Method, Columbia University, CS Department Technical Report, CUCS-025-03.

Ani Nenkova and Kathleen McKeown, References to Named Entities: a Corpus Study, NAACL-HLT, short paper, 2003.

Ani Nenkova and Kathleen McKeown, Improving the Coherence of Multi-document Summaries: a Corpus Study for Modeling the Syntactic Realization of Entities, Columbia University, CS Department Technical Report, CUCS-001-03.

Ani Nenkova and Amit Bagga, Facilitating Email Thread Access by Extractive Summary Generation, RANLP 2003.

Ani Nenkova and Amit Bagga, Email Classification for Contact Centers, 18th ACM Symposium on Applied Computing SAC 2003.

Barry Schiffman, Ani Nenkova and Kathleen McKeown, Experiments in Multidocument Summarization, Proceedings of HLT 2002 Human Language Technology Conference, 2002.

Kathleen R. McKeown, Regina Barzilay, David Evans, Vasileios Hatzivassiloglou, Judith L. Klavans, Ani Nenkova, Carl Sable, Barry Schiffman and Sergey Sigelman, Tracking and Summarizing News on a Daily Basis with Columbia's Newsblaster, Proceedings of HLT 2002 Human Language Technology Conference, 2002.

Ani Nenkova, A tableau method for graded intersection of modalities: a case for concept languages, Journal of Logic, Language and Information, Vol.11, Issue 1, 2002, pp. 67-77.

Ani Nenkova, Tableau Methods for Concept Languages, MS thesis (in Bulgarian), advisor Tinko Tinchev, June 2000.

Svetla Boytcheva, Ognian Kalaydjiev, Ani Nenkova and Galia Angelova, Integrating Resources and Components in a Knowledge-Based Environment for Terminology Learning, Proc. of the 9th Intern. Conference "Artificial Intelligence: Methodology, Systems, Applications" (AIMSA-00), September 2000, Lecture Notes in AI 1904, Springer, pp. 210-220, 2000.

Angelova, G., A. Nenkova, S. Boytcheva and T. Nikolov, Conceptual graphs as a knowledge representation core in a complex language learning environment, Proc. ICCS-2000, August 2000, Shaker Verlag, pp. 45-58, 2000.

Ani Nenkova and Galia Angelova, User Modeling as an Application of Actors, Proc. of 7th International Conference on Conceptual Structures (ICCS'99), 1999, LNAI, pp.83-89.