The International Conference on Learning Representations (ICLR), one of the top-tier conferences in deep learning, saw outstanding participation from Renmin University's Gaoling School of Artificial Intelligence at its 2025 edition. With36 accepted papers, a delegation of over 40 faculty and students—led by Prof. Wei Zhewei, Assoc. Prof. Yan Rui, Asst. Prof. Li Chongxuan, and postdoctoral researcher Zheng Yanping—attended the conference in Singapore from April 24–28, engaging with global scholars and showcasing cutting-edge research.

Highlights from the Conference

Qu Changle(PhD), first author of"From Exploration to Mastery: Enabling LLMs to Master Tools via Self-Driven Interactions", delivered an oral presentation and poster. He noted:"This first experience at a major conference deepened my understanding of diverse research approaches and highlighted the importance of communication skills."

Ma Shengjie(PhD) presented*"Think-on-Graph 2.0"*, describing interactions with leading researchers as"immensely inspiring."

Sun Zhongxiang(PhD) shared work ondetecting and mitigating hallucinations in RAG systemsat the Trustworthy AI session, emphasizing how exchanges with experts broadened his perspective.

Tan Wenhui(PhD), presenting"Think Then React: Motion Generation for Embodied AI", captured the conference’s energy:"From the airport to the venue, everyone carried posters and ideas—discussions sparked new insights at every turn."

Pei Qizhi(PhD) showcased*"3D-MolT5"*for molecule-text modeling, gaining clarity on future directions through peer feedback.

Key Takeaways

Students highlighted the value of:

Active engagement: Poster sessions offered deeper discussions than oral presentations.

Strategic attendance: Prioritizing keynotes, workshops, and industry booths (for career opportunities).

Visual appeal: Well-designed posters and slides attracted more meaningful exchanges.

Looking Ahead

Participants expressed hope for expanded opportunities to represent the school globally. Their experiences—from technical growth to networking—underscore Gaoling’s rising profile in AI research. As one PhD candidate noted,"ICLR didn’t just expand my ideas—it showed me where I need to improve, like my English fluency, to compete on this stage."With these insights, Renmin’s researchers are poised to contribute even more prominently to AI’s future.