PAL to Lend a Helping Hand: Towards Building an Emotion Adaptive Polite and Empathetic Counseling Conversational Agent

DOI: 10.18653/v1/2023.acl-long.685 Publication Date: 2023-08-05T00:57:42Z
ABSTRACT
The World Health Organization (WHO) has significantly emphasized the need for mental health care. social stigma associated with illness prevents individuals from addressing their issues and getting assistance. In such a scenario, relevance of online counseling increased dramatically. feelings attitudes that client counselor express towards each other result in higher or lower experience. A should be friendly gain clients' trust to make them share problems comfortably. Thus, it is essential adequately comprehend client's emotions ensure welfare, i.e. s/he adapt deal clients politely empathetically provide pleasant, cordial personalized Motivated by this, this work, we attempt build novel Polite empAthetic counseLing conversational agent PAL lay down support substance addict crime victims. To have emotion-based polite empathetic responses, two datasets laying addicts victims are annotated. These annotated used reinforcement learning framework. reward function formulated correct politeness empathy preferences as per naturalness non-repetitiveness responses. Thorough automatic human evaluation showcase usefulness strength designed function. Our proposed system scalable can easily modified different modules preference models need.
SUPPLEMENTAL MATERIAL
Coming soon ....
REFERENCES (0)
CITATIONS (2)