DepressionNet: Learning Multi-modalities with User Post Summarization for Depression Detection on Social Media

Abstract

Twitter is currently a popular online social media platform which allows users to share their user-generated content. This publiclygenerated user data is also crucial to healthcare technologies because the discovered patterns would hugely benefit them in several ways. One of the applications is in automatically discovering mental health problems, e.g., depression. Previous studies to automatically detect a depressed user on online social media have largely relied upon the user behaviour and their linguistic patterns including user's social interactions. The downside is that these models are trained on several irrelevant content which might not be crucial towards detecting a depressed user. Besides, these content have a negative impact on the overall efficiency and effectiveness of the model. To overcome the shortcomings in the existing automatic depression detection methods, we propose a novel computational framework for automatic depression detection that initially selects relevant content through a hybrid extractive and abstractive summarization strategy on the sequence of all user tweets leading to a more fine-grained and relevant content. The content then goes to our novel deep learning framework comprising of a unified learning machinery comprising of Convolutional Neural Network (CNN) coupled with attention-enhanced Gated Recurrent Units (GRU) models leading to better empirical performance than existing strong baselines.

MHA: a multimodal hierarchical attention model for depression detection in social media

“…It can be seen that before the u i point, the number of pictures of normal users is equal to the threshold, and the number of pictures of depressed users is less than the threshold, which will affect the calculation (11)…”

“…In addition, Chinese expressions are more diverse, so these methods often cannot obtain users' potential real emotion in Chinese depression detection. Aiming at this problem, some studies are committed to using multimodal data for depression detection [8][9][10][11]. Figure 1 is an example of pictures posted by normal user and depression user.…”

A Convolutional Attention Network for Unifying General and Sequential Recommenders

“…Inspired by the outperformance of Transformer [27] [28] [29] [9] in NLP tasks, SRs have motivated to use self-attention technique to better capture sequential dependency. BERT4Rec [30] for instance, has used the deep bidirectional selfattention algorithm to model the sequences of users' behaviors.…”

NYCU_TWD@LT-EDI-ACL2022: Ensemble Models with VADER and Contrastive Learning for Detecting Signs of Depression from Social Media

“…For instance, Toto et al ( 2021) introduced a framework with transfer learning to the multi-modality of textual context and audio characteristics of the human voice. Zogan et al (2021) proposed DepressionNet by summarizing history posts as a summary of the user and applying different modalities to infer user behavior, which motivated us to include VAD scores as the additional post feature in this challenge.…”