AI will now be able to detect depression, anxiety

A recent report suggested that work is underway to create an AI that would detect early signs of anxiety and depression – a mental health disorder that is quite prevalent in contemporary times.

The findings, published in the journal Language Resources and Evaluation, mention that the AI ​​will also collaborate with micro-blogging platform Twitter to do the same.

Researchers from the University of São Paulo (USP) in Brazil said preliminary findings from the model suggested the possibility of detecting a person’s likelihood of developing depression based only on their social media friends and followers.

The first phase of this study involved the creation of a database, called SetembroBR, of information relating to 47 million publicly posted Portuguese texts and networks of connections among 3,900 Twitter users. These users were reportedly diagnosed or treated for mental health problems prior to the survey. The tweets were collected during the COVID-19 pandemic.

Because people with mental health problems follow certain accounts such as discussion forums, influencers and celebrities who publicly acknowledge DepressionThe study also collected tweets from friends and followers.

The second phase, still in progress, has provided some preliminary findings, such as the likelihood of a person developing depression based solely on their social media friends and followers, not their own posts. without taking into account

After pre-processing the corpus to retain the original text by removing non-standard characters, the researchers Deployed Deep Learning (AI)To build four text classifiers and word embeddings (context-dependent mathematical representations of relationships between words) using models based on bidirectional encoder representations from Transformers (BERT), a machine learning algorithm employed for NLP.

These models correspond to a neural network that learns contexts and meanings by monitoring sequential data relationships, such as the words in a sentence. The training input consisted of a sample of 200 tweets randomly selected from each user.

The researchers found that among the models, BERT performed best in terms of predicting depression and anxiety. They said that because the model analyzed sequences of words and complete sentences, it was possible to observe that people with depression, for example, tend to write about topics related to themselves, using verbs and phrases in the first person. , as well as such topics as death, crisis and psychology.

(With inputs from PTI)