Tags

• CRFs 9
• relation extraction 2
• coreference 9
• MCMC 4
• canonicalization 1
• Twitter 30
• regression 11
• classification 17
• personality 1
• GMMs 1
• learning from label proportions 6
• sentiment 1
• distant supervision 2
• domain adaptation 5
• marketing 2
• perception 4
• confounding 5
• recommendation systems 1
• politics 1
• Amazon 1
• recalls 1
• PU learning 1
• co-training 1
• Instagram 1
• cyberbullying 1

CRFs

• Tweedr: Mining Twitter to Inform Disaster Response
• Too Neurotic, Not too Friendly: Structured Personality Classification on Textual Data
• Integrating probabilistic extraction models and data mining to discover relations and patterns in text
• Corrective Feedback and Persistent Learning for Information Extraction
• Reducing labeling effort for structured prediction tasks
• Joint deduplication of multiple record types in relational data
• Interactive information extraction with constrained conditional random fields
• Extracting social networks and contact information from email and the Web
• Confidence estimation for information extraction

relation extraction

• Integrating probabilistic extraction models and data mining to discover relations and patterns in text
• Dependency tree kernels for relation extraction

coreference

• An entity-based model for coreference resolution
• Learning and inference in weighted logic with application to natural language processing
• First-Order Probabilistic Models for Coreference Resolution
• Author Disambiguation using Error-driven Machine Learning with a Ranking Loss Function
• Tractable Learning and Inference with High-Order Representations
• Practical Markov logic containing first-order quantifiers with application to identity uncertainty
• Learning field compatibilities to extract database records from unstructured text
• Learning clusterwise similarity with first-order features
• Joint deduplication of multiple record types in relational data

MCMC

• SampleRank: Training factor graphs with atomic gradients
• SampleRank: Learning preferences from atomic gradients
• Sparse Message Passing Algorithms for Weighted Maximum Satisfiability
• Tractable Learning and Inference with High-Order Representations

canonicalization

• Canonicalization of Database Records using Adaptive Similarity Measures

Twitter

• Discovering and Controlling for Latent Confounds in Text Classification Using Adversarial Domain Adaptation
• When do Words Matter? Understanding the Impact of Lexical Choice on Audience Perception using Individual Treatment Effect Estimation
• Robust Text Classification under Confounding Shift
• Learning from noisy label proportions for classifying online social data
• Mining the Demographics of Political Sentiment from Twitter Using Learning from Label Proportions
• Co-training for Demographic Classification Using Deep Learning from Label Proportions
• Are Words Commensurate with Actions? Quantifying Commitment to a Cause from Online Public Messaging
• Controlling for Unobserved Confounds in Classification Using Correlational Constraints
• Using online social networks to measure consumers’ brand perception
• #Polar Scores: Measuring Partisanship Using Social Media Content
• Domain Adaptation for Learning from Label Proportions Using Self-Training
• Robust Text Classification in the Presence of Confounding Bias
• Reducing confounding bias in observational studies that use text classification
• Predicting Twitter User Demographics using Distant Supervision from Website Traffic Data
• Mining brand perceptions from Twitter social networks
• Training a text classifier with a single word using Twitter Lists and domain adaptation
• A demographic and sentiment analysis of e-cigarette messages on Twitter
• Finding truth in cause-related advertising: A lexical analysis of brands' health, environment, and social justice communications on Twitter
• Inferring latent attributes of Twitter users with label regularization
• Using Matched Samples to Estimate the Effects of Exercise on Mental Health from Twitter
• Predicting the Demographics of Twitter Users from Website Traffic Data
• Reducing Sampling Bias in Social Media Data for County Health Inference
• Using county demographics to infer attributes of Twitter users
• Tweedr: Mining Twitter to Inform Disaster Response
• Estimating County Health Statistics with Twitter
• Inferring the Origin Locations of Tweets with Quantitative Confidence
• Lightweight methods to estimate influenza rates and alcohol sales volume from Twitter messages
• A demographic analysis of online sentiment during Hurricane Irene
• Detecting influenza epidemics by analyzing Twitter messages
• Towards detecting influenza epidemics by analyzing Twitter messages

regression

• Predicting Twitter User Demographics using Distant Supervision from Website Traffic Data
• Finding truth in cause-related advertising: A lexical analysis of brands' health, environment, and social justice communications on Twitter
• Inferring latent attributes of Twitter users with label regularization
• Using Matched Samples to Estimate the Effects of Exercise on Mental Health from Twitter
• Predicting the Demographics of Twitter Users from Website Traffic Data
• Reducing Sampling Bias in Social Media Data for County Health Inference
• Using county demographics to infer attributes of Twitter users
• Estimating County Health Statistics with Twitter
• Lightweight methods to estimate influenza rates and alcohol sales volume from Twitter messages
• Detecting influenza epidemics by analyzing Twitter messages
• Towards detecting influenza epidemics by analyzing Twitter messages

classification

• Discovering and Controlling for Latent Confounds in Text Classification Using Adversarial Domain Adaptation
• Robust Text Classification under Confounding Shift
• Learning from noisy label proportions for classifying online social data
• Mining the Demographics of Political Sentiment from Twitter Using Learning from Label Proportions
• Co-training for Demographic Classification Using Deep Learning from Label Proportions
• Controlling for Unobserved Confounds in Classification Using Correlational Constraints
• Identifying leading indicators of product recalls from online reviews using positive unlabeled learning and domain adaptation
• Domain Adaptation for Learning from Label Proportions Using Self-Training
• Robust Text Classification in the Presence of Confounding Bias
• Reducing confounding bias in observational studies that use text classification
• Training a text classifier with a single word using Twitter Lists and domain adaptation
• A demographic and sentiment analysis of e-cigarette messages on Twitter
• Anytime Active Learning
• Tweedr: Mining Twitter to Inform Disaster Response
• Too Neurotic, Not too Friendly: Structured Personality Classification on Textual Data
• Towards Anytime Active Learning: Interrupting Experts to Reduce Annotation Costs
• A demographic analysis of online sentiment during Hurricane Irene

personality

• Too Neurotic, Not too Friendly: Structured Personality Classification on Textual Data

GMMs

• Inferring the Origin Locations of Tweets with Quantitative Confidence

learning from label proportions

• Learning from noisy label proportions for classifying online social data
• Mining the Demographics of Political Sentiment from Twitter Using Learning from Label Proportions
• Co-training for Demographic Classification Using Deep Learning from Label Proportions
• Domain Adaptation for Learning from Label Proportions Using Self-Training
• Inferring latent attributes of Twitter users with label regularization
• Using county demographics to infer attributes of Twitter users

sentiment

• A demographic and sentiment analysis of e-cigarette messages on Twitter

distant supervision

• Predicting Twitter User Demographics using Distant Supervision from Website Traffic Data
• Training a text classifier with a single word using Twitter Lists and domain adaptation

domain adaptation

• Discovering and Controlling for Latent Confounds in Text Classification Using Adversarial Domain Adaptation
• Robust Text Classification under Confounding Shift
• Controlling for Unobserved Confounds in Classification Using Correlational Constraints
• Identifying leading indicators of product recalls from online reviews using positive unlabeled learning and domain adaptation
• Training a text classifier with a single word using Twitter Lists and domain adaptation

marketing

• Using online social networks to measure consumers’ brand perception
• Mining brand perceptions from Twitter social networks

perception

• When do Words Matter? Understanding the Impact of Lexical Choice on Audience Perception using Individual Treatment Effect Estimation
• Are Words Commensurate with Actions? Quantifying Commitment to a Cause from Online Public Messaging
• Using online social networks to measure consumers’ brand perception
• Mining brand perceptions from Twitter social networks

confounding

• Discovering and Controlling for Latent Confounds in Text Classification Using Adversarial Domain Adaptation
• Robust Text Classification under Confounding Shift
• Controlling for Unobserved Confounds in Classification Using Correlational Constraints
• Robust Text Classification in the Presence of Confounding Bias
• Reducing confounding bias in observational studies that use text classification

recommendation systems

• Cold-Start Recommendations for Audio News Stories Using Matrix Factorization

politics

• #Polar Scores: Measuring Partisanship Using Social Media Content

Amazon

• Identifying leading indicators of product recalls from online reviews using positive unlabeled learning and domain adaptation

recalls

• Identifying leading indicators of product recalls from online reviews using positive unlabeled learning and domain adaptation

PU learning

• Identifying leading indicators of product recalls from online reviews using positive unlabeled learning and domain adaptation

co-training

• Co-training for Demographic Classification Using Deep Learning from Label Proportions

Instagram

• Forecasting the presence and intensity of hostility on Instagram using linguistic and social features

cyberbullying

• Forecasting the presence and intensity of hostility on Instagram using linguistic and social features