Welcome to TsvetShop! Yulia Tsvetkov's research group with members at the Paul G. Allen School of Computer Science & Engineering at the University of Washington and the Language Technologies Institute of Carnegie Mellon University. Our work focuses on natural language processing, particularly cross-lingual approaches, low-resource settings, and social good.

News

Older posts…


People

Faculty

Yulia Tsvetkov
Assistant Professor

Graduate Students

Anjalie Field
PhD Student, CMU
Sachin Kumar
PhD Student, CMU
Vidhisha Balachandran
PhD Student, CMU
Chan Young Park
PhD Student, CMU
Zirui Wang
PhD Student, CMU
Artidoro Pagnoni
PhD Student, UW
Xiaochuang Han
PhD Student, UW
Rishabh Joshi
MLT Student, CMU
Alissa Ostapenko
MLT Student, CMU
Antonio Theophilo
Visiting Scholar, CMU
Nupoor Gandhi
PhD Student, CMU
Co-advisor: Alex Chouldechova
Alumni

Research

Strategic Non-collaborative Dialog Agents

Strategic Non-collaborative Dialog Agents

We study a non-collaborative dialog setting, where a seller negotiates with a buyer over a given product. Specifically, we 1) design a dynamic coach that recommends tactics in real time to the seller to get a better deal, 2) propose to model tactic history with an FST for better dialog planning and generation, and 3) propose a novel framework to model negotiations strategies and their dependencies as graph structures, via GNNs.

Factuality Evaluation Metrics

Factuality Evaluation Metrics

We develop a typology of factual errors occurring in abstractive summarization and built a benchmark for factuality metrics in abstractive summarizaiton.

Detection of Discrimination, Bias, and Microaggressions in Text

Detection of Discrimination, Bias, and Microaggressions in Text

We aim to detect subtle forms of bias and veiled hostility in text, including microaggressions, condescending language, and dehumanization

Politics, Propaganda, and Polarization in Online Media

Politics, Propaganda, and Polarization in Online Media

We are investigating media bias through a highly multidisciplinary approach that combines elements from NLP, political science, and economics. We are also working to identify false information spread on social networks through phylogeny analysis.

Demoting Spurious Confounds in Text Classification

Demoting Spurious Confounds in Text Classification

We develop methods of discovering and demoting latent confounds in text classifcation which correspond to superficial patterns specific to the training set but don’t generalize well.

Contextual Affective Analysis

Contextual Affective Analysis

How are people portrayed? Our generalizable methodology combines NLP with social psychology theory in order to address this question in various domains, including news coverage of the #MeToo movement and Wikipedia biographies of LGBT people.

Summarization: Structure and Mutlilinguality

Summarization: Structure and Mutlilinguality

We incorporate document structure into neural document representation models to make summaries more abstractive and improve interpretability. We explore extending summarization to low and no resource languages.

Model Interpretability

Model Interpretability

We investigate two interpretation methods in NLP that either highlight salient input words or identify influential training examples. We show when and how the latter one might provide more valuable insights.

Conditional Language Generation with Continuous Outputs

Conditional Language Generation with Continuous Outputs

We develop a method of conditional language generation using seq2seq models which produce word embeddings instead of a softmax based distribution over the vocabulary at each step enabling much faster training while maintaining generation quality.

Understanding Code-Mixing in Controlled Dialogues

Understanding Code-Mixing in Controlled Dialogues

We design a Spanish-English code-mixing dialogue system which uses diverse, linguistically-informed code-mixing strategies when talking to Spanish-English bilinguals.


Publications

2021
A Survey of Race, Racism, and Anti-Racism in NLP
Anjalie Field, Su Lin Blodgett, Zeerak Waseem, and Yulia Tsvetkov. Proc. ACL.
Machine Translation into Low-resource Language Varieties
Sachin Kumar, Antonios Anastasopoulos, Shuly Wintner, and Yulia Tsvetkov. Proc. ACL.
Synthesizing Adversarial Negative Responses for Robust Response Ranking and Evaluation
Prakhar Gupta, Yulia Tsvetkov, and Jeffrey P. Bigham. Proc. Findings of ACL.
Understanding Factuality in Abstractive Summarization with FRANK: A Benchmark for Factuality Metrics
Artidoro Pagnoni, Vidhisha Balachandran, and Yulia Tsvetkov. Proc. NAACL-HLT.
Controlling Dialogue Generation with Semantic Exemplars
Prakhar Gupta, Jeffrey P. Bigham, Yulia Tsvetkov, and Amy Pavel. Proc. NAACL-HLT.
DialoGraph: Incorporating Interpretable Strategy-Graph Networks into Negotiation Dialogues
Rishabh Joshi, Vidhisha Balachandran, Shikhar Vashishth, Alan Black, and Yulia Tsvetkov. Proc. ICLR.
Gradient Vaccine: Investigating and Improving Multi-task Optimization in Massively Multilingual Models
Zirui Wang, Yulia Tsvetkov, Orhan Firat, and Yuan Cao. Proc. ICLR.
StructSum: Incorporating Latent and Explicit Sentence Dependencies for Single Document Summarization
Vidhisha Balachandran, Artidoro Pagnoni, Jay Yoon Lee, Dheeraj Rajagopal, Jaime Carbonell, and Yulia Tsvetkov. Proc. EACL.
Cross-Cultural Similarity Features for Cross-Lingual Transfer Learning of Pragmatically Motivated Tasks
Jimin Sun, Hwijeen Ahn, Chan Young Park, Yulia Tsvetkov, and David R. Mortensen. Proc. EACL.
Multilingual Contextual Affective Analysis of LGBT People Portrayals in Wikipedia
Chan Young Park, Xinru Yan, Anjalie Field, and Yulia Tsvetkov. Proc. ICWSM.
An Exploration of Data Augmentation Techniques for Improving English to Tigrinya Translation
Lidia Kidane, Sachin Kumar, and Yulia Tsvetkov. Proc. AfricaNLP.
2020
End-to-End Differentiable GANs for Text Generation
Sachin Kumar and Yulia Tsvetkov. Proc. ICBINB.
Understanding Linguistic Accommodation in Code-Switched Human-Machine Dialogues
Tanmay Parekh, Emily Ahn, Yulia Tsvetkov, and Alan W. Black. Proc. CoNLL.
Automatic Extraction of Rules Governing Morphological Agreement
Aditi Chaudhary, Antonios Anastasopoulos, Adithya Pratapa, David R. Mortensen, Zaid Sheikh, Yulia Tsvetkov, and Graham Neubig. Proc. EMNLP.
Unsupervised Discovery of Implicit Gender Bias
Anjalie Field and Yulia Tsvetkov. Proc. EMNLP.
On Negative Interference in Multilingual Models: Findings and A Meta-Learning Treatment
Zirui Wang, Zachary C. Lipton, and Yulia Tsvetkov. Proc. EMNLP.
Fortifying Toxic Speech Detectors Against Veiled Toxicity
Xianchuang Han and Yulia Tsvetkov. Proc. EMNLP.
A Computational Analysis of Polarization onIndian and Pakistani Social Media
Aman Tyagi, Anjalie Field, Priyank Lathwal, Yulia Tsvetkov, and Kathleen M. Carley. Proc. SocInfo.
LTIatCMU at SemEval-2020 Task 11: Incorporating Multi-Level Features for Multi-Granular Propaganda Span Identification
Sopan Khosla, Rishabh Joshi, Ritam Dutt, Alan W. Black, and Yulia Tsvetkov. Proc. SemEval.
A framework for the computational linguistic analysis of dehumanization
Julia Mendelsohn, Yulia Tsvetkov, and Dan Jurafsky. Frontiers in Artificial Intelligence.
Demoting Racial Bias in Hate Speech Detection
Mengzhou Xia, Anjalie Field, and Yulia Tsvetkov. Proc. SocialNLP.
A Generative Approach to Titling and Clustering Wikipedia Sections
Anjalie Field, Sascha Rothe, Simon Baumgartner, Cong Yu, and Abe Ittycheriah. Proc. WNGT.
A Deep Reinforced Model for Cross-Lingual Summarization with Bilingual Semantic Similarity Reward
Zi-Yi Dou, Sachin Kumar, and Yulia Tsvetkov. Proc. WNGT.
Balancing Training for Multilingual Neural Machine Translation
Xinyi Wang, Yulia Tsvetkov, and Graham Neubig. Proc. ACL.
Explaining Black Box Predictions and Unveiling Data Artifacts through Influence Functions
Xiaochuang Han, Byron C. Wallace, and Yulia Tsvetkov. Proc. ACL.
Stress and Burnout in Open Source: Toward Finding, Understanding, and Mitigating Unhealthy Interactions
Naveen Raman, Minxuan Cao, Yulia Tsvetkov, Christian Kästner, and Bogdan Vasilescu. International Conference on Software Engineering -- New Ideas Track (ICSE-NIER).
Augmenting Non-Collaborative Dialog Systems with Explicit Semantic and Strategic Dialog History
Yiheng Zhou, Yulia Tsvetkov, Alan W Black, and Zhou Yu. Proc. ICLR.
What Code-Switching Strategies are Effective in Dialog Systems?
Emily Ahn, Cecilia Jimenez, Yulia Tsvetkov, and Alan W Black. Proc. SCiL.
Where New Words Are Born: Distributional Semantic Analysis of Neologisms and Their Semantic Neighborhoods
Maria Ryskina, Ella Rabinovich, Taylor Berg-Kirkpatrick, David Mortensen, and Yulia Tsvetkov. Proc. SCiL.
2019
Topics to Avoid: Demoting Latent Confounds in Text Classification
Sachin Kumar, Shuly Wintner, Noah A. Smith, and Yulia Tsvetkov. Proc. EMNLP.
Finding Microaggressions in the Wild: A Case for Locating Elusive Phenomena in Social Media Posts
Luke M. Breitfeller, Emily Ahn, David Jurgens, and Yulia Tsvetkov. Proc. EMNLP.
Learning to Generate Word- and Phrase-Embeddings for Efficient Phrase-Based Neural Machine Translation
Chan Young Park and Yulia Tsvetkov. Proc. WNGT.
A Margin-based Loss with Synthetic Negative Samples for Continuous-output Machine Translation
Gayatri Bhat, Sachin Kumar, and Yulia Tsvetkov. Proc. WNGT.
A Dynamic Strategy Coach for Effective Negotiation
Yiheng Zhou, He He, Alan W Black, and Yulia Tsvetkov. Proc. SIGdial.
Entity-Centric Contextual Affective Analysis
Anjalie Field and Yulia Tsvetkov. Proc. ACL.
CMU-01 at the SIGMORPHON 2019 Shared Task on Crosslinguality and Context in Morphology
Aditi Chaudhary, Elizabeth Salesky, Gayatri Bhat, David R. Mortensen, Jaime G. Carbonell, and Yulia Tsvetkov. Proc. SIGMORPHON.
Quantifying Social Biases in Contextual Word Representations
Keita Kurita, Nidhi Vyas, Ayush Pareek, Alan W Black, and Yulia Tsvetkov. Proc. of Workshop on Gender Bias for NLP.
Contextual Affective Analysis: A Case Study of People Portrayals in Online #MeToo Stories
Anjalie Field, Gayatri Bhat, and Yulia Tsvetkov. Proc. ICWSM.
Black is to Criminal as Caucasian is to Police: Detecting and Removing Multiclass Bias in Word Embeddings
Thomas Manzini, Yao Chong, Yulia Tsvetkov, and Alan W Black. Proc. NAACL.
Von Mises-Fisher Loss for Training Sequence to Sequence Models with Continuous Outputs
Sachin Kumar and Yulia Tsvetkov. Proc. ICLR.
2018
Framing and Agenda-setting in Russian News: a Computational Analysis of Intricate Political Strategies
Anjalie Field, Doron Kliger, Shuly Wintner, Jennifer Pan, Dan Jurafsky, and Yulia Tsvetkov. Proc. EMNLP.
RtGender: A corpus for studying differential responses to gender
Rob Voigt, David Jurgens, Vinodkumar Prabhakaran, Dan Jurafsky, and Yulia Tsvetkov. Proc. LREC'18.
Native Language Cognate Effects on Second Language Lexical Choice
Ella Rabinovich, Yulia Tsvetkov, and Shuly Wintner. TACL.
Style Transfer Through Back-Translation
Shrimai Prabhumoye, Yulia Tsvetkov, Ruslan Salakhutdinov, and Alan W Black. Proc. ACL.

Download BibTeX.


Funding

Our work has been supported by the following organizations/companies:

DARPA Logo DARPA Logo NSF Logo AWS Logo Adobe Logo NVF Logo