Contrastive Alignment for Preference-Driven KG-to-Text Generation

Room A008

• place: B013
• time: 14h00

The topic of my presentation is “Contrastive Alignment for Preference-Driven KG-to-Text Generation”. I will introduce multilingual cross-modal alignment between Knowledge Graphs and Text, and discuss how recent reinforcement learning techniques (DPO) can be applied to the KG-to-Text Generation task. The presentation will last approximately 30 minutes, followed by a Q&A session.