Jeongsik Park’s Post

View profile for Jeongsik Park, graphic

MSCS(AI)@USC | NLP, LLMs, Multimodal | Research, MLE

I recently had the pleasure of attending #YRRSDS (Sep 16-17) and #SIGdial (Sep 18-20) in the beautiful city of Kyoto, Japan. It was a fulfilling experience, particularly meeting new people who share the same research interests. At YRRSDS, I had the opportunity to exchange interesting ideas with fellow attendees, and throughout the conferences, I explored remarkable work from others in the field. First and foremost, I want to extend my gratitude to my co-authors, who worked tirelessly on our research project, the organizers for putting together such a successful event, and all the researchers who presented their outstanding work. SIGdial, which stands for the Special Interest Group on Discourse and Dialogue, significantly expanded my understanding of Discourse and Dialogue. Although my prior knowledge was limited, I left the conference with a wealth of insights and experience. Thanks to advancements in LLMs, it's clear that dialogue systems have made tremendous progress and will continue to evolve. I noticed a particular focus on areas such as persona and personalization. Despite the ongoing challenge of achieving real-time performance, there is growing emphasis on optimization and using smaller models to enhance speed and reduce costs while maintaining effectiveness. I left feeling optimistic that, thanks to LLMs, we are approaching practical solutions to these challenges. Another key takeaway was the realization that dialogue systems, through chatbots and robots, are already being implemented in the education sector. I'm looking forward to the next conference and witnessing the continued progress and success of researchers in this field. FYI, this is our work that was presented at SIGdial 2024: MemeIntent: Benchmarking Intent Description Generation for Memes (https://lnkd.in/dMWyE2gC)

  • No alternative text description for this image

To view or add a comment, sign in

Explore topics