Papers by Sungjoon Park
KOLD: Korean Offensive Language Dataset (2022.emnlp-main)
Copied to clipboard
| Challenge: | Recent directions for offensive language detection focus on English and do not transfer well to other languages because of cultural and linguistic differences. |
| Approach: | They present a Korean offensive language dataset annotated with offensive language comments . they use the comments as training data for Korean BERT and RoBERTa models . |
| Outcome: | The proposed model improves offensiveness detection, target classification, and span detection while having room for improvement for target group classification and span prediction. |
Hierarchical Dirichlet Gaussian Marked Hawkes Process for Narrative Reconstruction in Continuous Time Domain (D18-1)
Copied to clipboard
| Challenge: | Existing methods for reconstructing narratives and thread structures of news articles and discussions are lacking . temporal characteristics, triggering event relations, and meta information are used to solve the problem . |
| Approach: | They propose a Hierarchical Dirichlet Gaussian Marked Hawkes process for reconstructing narratives and thread structures of news articles and discussion posts. |
| Outcome: | The proposed model outperforms baseline models on real-world datasets and Wikipedia conversations. |
FedTherapist: Mental Health Monitoring with User-Generated Linguistic Expressions on Smartphones via Federated Learning (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing passive mental health monitoring systems use alternative features such as activity, app usage, and location via smartphones due to data privacy concerns. |
| Approach: | They propose a mobile mental health monitoring system that utilizes continuous speech and keyboard input in a privacy-preserving way via federated learning. |
| Outcome: | The proposed system achieves 0.15 AUROC improvement and 8.21% MAE reduction in self-reported depression, stress, anxiety, and mood from 46 participants. |
Analyzing Norm Violations in Live-Stream Chat (2023.emnlp-main)
Copied to clipboard
Jihyung Moon, Dong-Ho Lee, Hyundong Cho, Woojeong Jin, Chan Park, Minwoo Kim, Jonathan May, Jay Pujara, Sungjoon Park
| Challenge: | Existing methods for detecting toxic language and norm violations are limited to live-streaming platforms . existing methods are less effective when applied to live streaming platforms based on a limited time frame . |
| Approach: | They propose to use contextual information to automatically moderate toxic content on live streaming platforms. |
| Outcome: | The proposed model improves on live-streaming platforms by 35%. |
Suicidal Risk Detection for Military Personnel (2020.emnlp-main)
Copied to clipboard
| Challenge: | a dataset of 2,791 posts with 13,955 expert annotations of suicidal risk levels is available for research . Suicide is one of the major causes of death in the military. |
| Approach: | They analyze posts related to military service in the Republic of Korea and annotate them with military experts and mental health experts. |
| Outcome: | The proposed method predicts the level of suicide risk, reaching .88 F1 for classifying the risks. |
Subword-level Word Vector Representations for Korean (P18-1)
Copied to clipboard
| Challenge: | Existing research on word vectors for English focuses on decomposing words into subword units and using subwords to improve performance. |
| Approach: | They propose to decompose Korean words into the jamo-level, beyond the character-level . they develop Korean test sets for word similarity and analogy and make them publicly available . |
| Outcome: | The proposed method outperforms word2vec and character-level skip-grams on similarity and analogy tasks and contributes positively toward downstream NLP tasks such as sentiment analysis. |
Fast End-to-end Coreference Resolution for Korean (2020.findings-emnlp)
Copied to clipboard
| Challenge: | Recent advances in coreference resolution have come at a cost of computational complexity and have not been addressed. |
| Approach: | They propose a pointer network that leverages the linguistic property of head-final languages to reduce coreference linking search space and achieve 2x speedup in document processing time. |
| Outcome: | The proposed model maintains state-of-the-art performance 66.9% of CoNLL F1 on ETRI test set while achieving 2x speedup (30 doc/sec) in document processing time. |
Dimensional Emotion Detection from Categorical Emotion (2021.emnlp-main)
Copied to clipboard
| Challenge: | Using a model to predict fine-grained emotions along the continuous dimensions of valence, arousal, and dominance (VAD) with a corpus with categorical emotion annotations, we show that our approach reaches comparable performance to that of the state-of-the-art classifiers in categorial emotion classification and shows significant positive correlations with the ground truth VAD scores. |
| Approach: | They propose to train a model to predict fine-grained emotions along the continuous dimensions of valence, arousal, and dominance with a corpus with categorical emotion annotations. |
| Outcome: | The proposed model can predict emotions along the continuous dimensions of valence, arousal, and dominance (VAD) with a corpus with categorical emotion annotations. |
Conversation Model Fine-Tuning for Classifying Client Utterances in Counseling Dialogues (N19-1)
Copied to clipboard
| Challenge: | Recent surge of text-based online counseling applications enables us to collect and analyze interactions between counselors and clients. |
| Approach: | They develop a pre-trained conversation model that learns to classify client utterances into categories that help counselors in diagnosing client status and predicting counseling outcome. |
| Outcome: | The proposed model outperforms state-of-the-art comparison models and shows expected linguistic patterns for each category. |
Towards standardizing Korean Grammatical Error Correction: Datasets and Annotation (2023.acl-long)
Copied to clipboard
| Challenge: | Despite the growing number of Korean learners, little research has been conducted on Korean grammatical error correction (GEC) despite the difficulties of the Korean language, there is no evaluation benchmark for Korean GEC. |
| Approach: | They propose to use Korean grammar error correction datasets to train a machine learning model that can automatically annotate Korean errors from parallel corpora. |
| Outcome: | The proposed model outperforms the currently used statistical Korean GEC system on a wider range of error types. |
Additive Compositionality of Word Vectors (D19-55)
Copied to clipboard
| Challenge: | Existing research on justifying additive compositionality of word embedding models requires a rather strong assumption of uniform word distribution. |
| Approach: | They propose to relax the assumption of uniform word distribution and propose more realistic conditions for proving additive compositionality. |
| Outcome: | The proposed model improves on word similarity and noisy sentence similarity. |