A seminar entitled “From Private Chat Annotation to Automatic Identification of Problematic Conversations in Customer Services”, with Gaël Guibon (Synalp team) will take place on Thursday, 8th November 2022 at 10:00 in room A008.
Many companies make use of customer service chats to help the customer and try to solve their problem. However, customer service data is confidential and as such, cannot easily be shared in the research community. This also implies that these data are rarely labeled, making it difficult to take advantage of it with machine learning methods.
In this seminar several works on emotion recognition in conversation and customer service related tasks will be presented. First we will go through data annotation with the annotation campaign from scratch even creating a dedicated annotation tool. Second, we will tackle data and label agnostic models using meta-learning. Then we will present the first work on a few-shot learning approach to emotion recognition in conversation. Finally, we will end the presentation with the first work on problem status and problematic conversation identification.
All this work has been made during a joint post-doctoral position at the Information Processing and Communications Laboratory (LTCI) from Télécom Paris and the SNCF Direction Technologie Innovation Recherche Groupe (DTIPG).
