Meituan's 10 Papers at CVPR 2025 and ICLR 2025
Meituan presented 10 papers at CVPR 2025 and ICLR 2025, featuring breakthroughs in LLMs, multimodal generation, and computer vision, such as ternary preference alignment, temporal-decay DPO, joint-embedding architectures, 4-bit quantization, and spatial-temporal models, emphasizing model efficiency, cross-modal alignment, and high-resolution synthesis, and will host an ICLR 2025 booth in Singapore for collaboration and knowledge exchange.
This article introduces 10 papers presented by Meituan's technical team at CVPR 2025 and ICLR 2025, covering advancements in large language models (LLMs), multimodal generation, and computer vision. Topics include LLM alignment with ternary preferences, temporal-decay DPO for preference optimization, joint-embedding predictive architectures for multimodal generation, 4-bit quantization methods for LLMs, and spatial-temporal understanding models. The papers highlight technical innovations in areas like model efficiency, cross-modal alignment, and high-resolution image synthesis. Additionally, Meituan will host an ICLR 2025 booth (J07) in Singapore from April 24-28, 2025, inviting collaboration and knowledge exchange.
Meituan Technology Team
Over 10,000 engineers powering China’s leading lifestyle services e‑commerce platform. Supporting hundreds of millions of consumers, millions of merchants across 2,000+ industries. This is the public channel for the tech teams behind Meituan, Dianping, Meituan Waimai, Meituan Select, and related services.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.