Papers by Natalie Parde

23 papers
From Heart to Words: Generating Empathetic Responses via Integrated Figurative Language and Semantic Context Signals (2025.findings-acl)

Copied to clipboard

Challenge: Existing research on empathy generation focuses on understanding the emotions of the speaker rather than on how the responder conveys empathy.
Approach: They propose to use figurative language and causal semantic context to facilitate targeted empathy generation in a mental health support domain.
Outcome: The proposed approach achieves 7.6% improvement in BLEU, 36.7% reduction in Perplexity, and 7.6% increase in lexical diversity.
Telling a Lie: Analyzing the Language of Information and Misinformation during Global Health Events (2022.lrec-1)

Copied to clipboard

Challenge: a new dataset is available to stimulate research on health misinformation . linguistic characteristics of health misinfonia are unique to COVID-19 and other events .
Approach: They propose a new dataset that analyzes health misinformation at scale . it includes 2.8 million news articles and social media posts covering diseases . authors propose an annotation framework that allows for strong agreement between annotators .
Outcome: The proposed dataset is based on 2.8 million news articles and social media posts spanning 1900s to present . it shows that the proposed model is robust and can be used to detect misinformation .
Modeling Dialogue in Conversational Cognitive Health Screening Interviews (2020.lrec-1)

Copied to clipboard

Challenge: Dementia is one of the most pressing healthcare concerns as median age rises . a conversational agent capable of conducting cognitive health screening interviews could be an inexpensive, flexible, low-stress alternative .
Approach: They propose an annotation schema for assigning dialogue act labels to utterances in patient-interviewer conversations collected as part of a clinically-validated cognitive health screening task.
Outcome: The proposed system is characterized by high inter-annotator agreement and is able to perform clinically-validated cognitive health screening tasks.
Towards Domain-Agnostic and Domain-Adaptive Dementia Detection from Spoken Language (2023.acl-long)

Copied to clipboard

Challenge: Domain adaptation (DA) techniques have been used to improve performance of NLP systems for healthcare tasks due to numerous complexities of data.
Approach: They propose to use domain adaptation techniques to improve generalizability across diverse datasets for dementia detection.
Outcome: The proposed model achieves a 22% increase in accuracy adapting from a conversational to task-oriented dataset compared to a jointly trained baseline.
TweetTaglish: A Dataset for Investigating Tagalog-English Code-Switching (2022.lrec-1)

Copied to clipboard

Challenge: a large dataset is available to study Tagalog-English code-switching in low-resource settings.
Approach: They propose to use a large dataset to investigate Tagalog-English code-switching . they use linguistic data from Tagalogue and Tagalit-English to investigate their results .
Outcome: The proposed dataset achieves a strong performance benchmark for Tagalog-English code-switching.
The AI Doctor Is In: A Survey of Task-Oriented Dialogue Systems for Healthcare Applications (2022.acl-long)

Copied to clipboard

Challenge: Task-oriented dialogue systems have been surveyed in the medical community from a non-technical perspective, but a systematic review from . a rigorous computational perspective has to date remained noticeably absent.
Approach: They analyze 4070 papers on task-oriented dialogue systems for healthcare applications and identify gaps in their analysis.
Outcome: The proposed system-level implementation details remain limited or underspecified, slowing the pace of innovation in this area.
Pouring Your Heart Out: Investigating the Role of Figurative Language in Online Expressions of Empathy (2024.acl-long)

Copied to clipboard

Challenge: Empathy is a social mechanism used to support and strengthen emotional connection with others, including in online communities.
Approach: They propose to use a subcategory of linguistic phenomena, figurative language, to detect online expressions of empathy.
Outcome: The proposed model improves on the use of metaphor, idiom, and hyperbole in online expressions of empathy, with 0.942 and 0.809 F1 scores.
SLaCAD: A Spoken Language Corpus for Early Alzheimer’s Disease Detection (2024.lrec-main)

Copied to clipboard

Challenge: Recent studies show that cerebrospinal fluid (CSF) levels serve as useful early biomarkers for identifying early AD, but CSF biomarker collection is challenging.
Approach: They propose to use speech data to identify early Alzheimer's disease (AD) trajectory to identify cognitive deficits in early disease stages.
Outcome: The proposed dataset relates speech and speech characteristics with CSF and plasma biomarkers to clinical diagnoses, CSF levels, and biomarker scores.
CareCorpus: A Corpus of Real-World Solution-Focused Caregiver Strategies for Personalized Pediatric Rehabilitation Service Design (2024.lrec-main)

Copied to clipboard

Challenge: Pediatric rehabilitation services focus on functional skills and participation, defined as attendance and involvement in home, school, and community activities.
Approach: They propose to use a dataset of 780 real-world strategies written by caregivers to sort caregiver strategies for use in designing pediatric rehabilitation care plans.
Outcome: The proposed model can be used to sort caregiver strategies for use in designing pediatric rehabilitation care plans.
Identifying Medical Self-Disclosure in Online Communities (2021.naacl-main)

Copied to clipboard

Challenge: a new dataset of health-related posts from online social platforms is available for analysis . medical self-disclosure may be useful for early detection and treatment of medical issues .
Approach: They propose to analyze medical self-disclosure in online health conversations . they release a dataset of health-related posts from online social platforms with high inter-annotator agreement .
Outcome: The proposed model achieves an accuracy of 81.02% and sets a strong performance benchmark.
CORI: CJKV Benchmark with Romanization Integration - a Step towards Cross-lingual Transfer beyond Textual Scripts (2024.lrec-main)

Copied to clipboard

Challenge: Naively assuming English as a source language may hinder cross-lingual transfer . despite recent advances in cross-linguistic research, most studies have restricted themselves to two major assumptions .
Approach: They propose to integrate Romanized transcription beyond textual scripts to capture contact between these languages . they propose to use a benchmark dataset to further encourage in-depth studies of language contact .
Outcome: The proposed method allows for enhanced cross-lingual representations and effective zero-shot cross-linguistic transfer.
CareCorpus+: Expanding and Augmenting Caregiver Strategy Data to Support Pediatric Rehabilitation (2024.emnlp-main)

Copied to clipboard

Challenge: Existing studies on caregiver strategy classification in pediatric rehabilitation contexts are under-resourced and under-studied.
Approach: They propose to manually categorized 4,037 caregiver strategies in a pediatric rehabilitation setting, and manually supplement target task data with publicly relevant child health forums.
Outcome: The proposed method improves the quality of the dataset and shows that it performs well.
Exploring Variation of Results from Different Experimental Conditions (2023.findings-acl)

Copied to clipboard

Challenge: Recent research has shown that reproducibility of NLP experiments is not guaranteed by arbitrary factors like random seed and different data splits.
Approach: They propose to run the training and testing pipeline for three neural text simplification models under varying experimental conditions using the same data and train/dev/test set splits.
Outcome: The proposed pipeline yields a large number of results for three neural text simplification models under varying experimental conditions.
Reproducibility in Computational Linguistics: Is Source Code Enough? (2022.emnlp-main)

Copied to clipboard

Challenge: a recent study finds that source code releases leave much to be desired . authors can include small-scale experiments and explicit scripts to generate each result .
Approach: They propose that conferences require self-contained artifacts to evaluate their reproducibility . authors suggest authors include small-scale experiments and explicit scripts to generate results .
Outcome: The authors find that source code releases leave much to be desired . authors can include small-scale experiments and explicit scripts to generate each result .
Humanistic Buddhism Corpus: A Challenging Domain-Specific Dataset of English Translations for Classical and Modern Chinese (2024.lrec-main)

Copied to clipboard

Challenge: Compared to existing machine translation datasets, HBC presents unique challenges . classical and modern Chinese texts are often translated in distant languages .
Approach: They propose a dataset containing 80,000 Chinese-English parallel phrases extracted and translated from publications in the domain of Buddhism.
Outcome: The Humanistic Buddhism Corpus (HBC) contains 80,000 parallel Chinese-English phrases extracted and translated from publications in the domain of Buddhism.
Towards Intelligent Clinically-Informed Language Analyses of People with Bipolar Disorder and Schizophrenia (2022.findings-emnlp)

Copied to clipboard

Challenge: Existing studies on social media data have limited the extent to which they can produce meaningful or generalizable conclusions.
Approach: They propose to use transcribed conversations with people with bipolar disorder and schizophrenia to create a large dataset of transcriptions.
Outcome: The proposed dataset extracts 100+ temporal, sentiment, psycholinguistic, emotion, and lexical features and establishes classification validity.
AcnEmpathize: A Dataset for Understanding Empathy in Dermatology Conversations (2024.lrec-main)

Copied to clipboard

Challenge: Existing studies on empathy and mental health-related corpora focus on broader contexts and lack domain specificity.
Approach: They propose a dataset that captures empathy expressed in acne-related discussions from forum posts focused on its emotional and psychological effects.
Outcome: The AcnEmpathize dataset shows that it performs well at empathy classification.
Domain Adaptation via Prompt Learning for Alzheimer’s Detection (2024.findings-emnlp)

Copied to clipboard

Challenge: Existing attempts to fine-tune pre-trained language models for AD detection are limited by the small and disparate corpora of spoken language.
Approach: They propose to use domain-adaptive prompt fine-tuning to optimize for AD detection by using AD classification loss as the training objective and spoken language corpora from a variety of language tasks.
Outcome: The proposed approach achieves the highest detection rate with the highest-performing prompting paradigm and the domain-adaptive prompt fine-tuning paradigm.
Towards Comprehensive Language Analysis for Clinically Enriched Spontaneous Dialogue (2024.lrec-main)

Copied to clipboard

Challenge: Contemporary NLP has progressed from feature-based classification to fine-tuning and prompt-based techniques . many of these techniques remain understudied in the context of real-world, clinically enriched spontaneous dialogue.
Approach: They investigate the efficacy and overall performance of a range of NLP techniques on transcribed speech from patients with schizophrenia and other disorders.
Outcome: The proposed methods are effective in analyzing transcribed speech from patients with schizophrenia and healthy controls taking a clinically-validated language test.
Enriching Neural Models with Targeted Features for Dementia Detection (P19-2)

Copied to clipboard

Challenge: In the United States, adults over 65 are expected to comprise one-fifth of the population by 2030, and a larger proportion of the . population than those under 18 by 2035.
Approach: They propose a neural model that takes into account both long language samples and hand-crafted linguistic features to distinguish between dementia affected and healthy patients.
Outcome: The proposed model achieves an F1 score of 0.929 on the DementiaBank dataset and the state-of-the-art on the dataset.
A Corpus of Metaphor Novelty Scores for Syntactically-Related Word Pairs (L18-1)

Copied to clipboard

Challenge: Existing data on metaphor novelty are limited, making it difficult to perform research on this topic.
Approach: They propose to release a corpus of metaphor novelty scores for syntactically related word pairs . they establish a performance benchmark to which future researchers can compare .
Outcome: The proposed corpus of metaphor novelty scores is compared to other datasets . it performs better than chance or nave strategies, the authors show .
Demystifying Neural Fake News via Linguistic Feature-Based Interpretation (2022.coling-1)

Copied to clipboard

Challenge: Recent advances to neural fake news generators have made it difficult to understand how misinformation generated by these models may best be confronted.
Approach: They conduct feature-based analysis to gain an interpretative understanding of the linguistic attributes that neural fake news generators may most effectively exploit.
Outcome: The proposed models are compared with models trained on subsets of features and confronted with increasingly advanced neural fake news.
What Clued the AI Doctor In? On the Influence of Data Source and Quality for Transformer-Based Medical Self-Disclosure Detection (2023.eacl-main)

Copied to clipboard

Challenge: Recognizing medical self-disclosure is important in many healthcare contexts, but it has been under-explored by the NLP community.
Approach: They analyze a social media-based task to expand existing medical self-disclosure corpus and compare Transformer-based models to determine their merits.
Outcome: The proposed dataset outperforms the state-of-the-art dataset for this task by 16.73%.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations