A Baseline for Self-state Identification and Classification in Mental Health Data: CLPsych 2025 Task
We present a baseline for the CLPsych 2025 A.1 task: classifying self-states in mental health data taken from Reddit. We use few-shot learning with a 4-bit quantized Gemma 2 9B model and a data preprocessing step which first identifies relevant sentences indicating self-state evidence, and then performs a binary classification to determine whether the sentence is evidence of an adaptive or maladaptive self-state. This system outperforms our other method which relies on an LLM to highlight spans of variable length independently. We attribute the performance of our model to the benefits of this sentence chunking step for two reasons: partitioning posts into sentences 1) broadly matches the granularity at which self-states were human-annotated and 2) simplifies the task for our language model to a binary classification problem. Our system places third out of fourteen systems submitted for Task A.1, achieving a test-time recall of 0.579.
Code (0)
등록된 구현이 없습니다.
Tasks
AttributeBinary ClassificationChunkingFew-Shot LearningLanguage ModelingLanguage ModellingSentenceSimilar Papers 제목 키워드 기반
Wildlife Target Re-Identification Using Self-supervised Learning in Non-Urban Settings
Wildlife re-identification aims to match individuals of the same species across different observations. Current state-of-the-art (SOTA) models rely on class labels to train supervised models for individual classification…
Self-Supervised LearningTransfer LearningTask-adaptive Pre-training and Self-training are Complementary for Natural Language Understanding
Task-adaptive pre-training (TAPT) and Self-training (ST) have emerged as the major semi-supervised approaches to improve natural language understanding (NLU) tasks with massive amount of unlabeled data. However, it's unc…
named-entity-recognitionNamed Entity RecognitionNamed Entity Recognition (NER)Natural Language Inference+4ProtoX-AD: Self-Explainable Time Series Anomaly Detection and Characterization
Recent advances in time series anomaly detection (TSAD) have highlighted the effectiveness of self-supervised classification-based approaches. These methods apply transformations to normal training samples, training a cl…
Time Series Anomaly DetectionRSDD-Time: Temporal Annotation of Self-Reported Mental Health Diagnoses
Self-reported diagnosis statements have been widely employed in studying language related to mental health in social media. However, existing research has largely ignored the temporality of mental health diagnoses. In th…
General ClassificationEvolveReason: Self-Evolving Reasoning Paradigm for Explainable Deepfake Facial Image Identification
With the rapid advancement of AIGC technology, developing identification methods to address the security challenges posed by deepfakes has become urgent. Face forgery identification techniques can be categorized into two…
Reinforcement Learning