Technology is the core driver of Ant's future creation.
The ICCV 2025 live session will deep‑dive into two cutting‑edge papers—PixTrace with CopyNCE for precise image copy detection and Skip‑Vision for dramatically faster training and inference of vision‑language models—showcasing their methods, results, and real‑world impact.
Ant Insurance’s tech team leveraged reinforcement learning, focused data selection, and a multi‑dimensional reward system to dramatically reduce hallucinations in LLMs, achieving top‑rank performance on the HHEM leaderboard and robust improvements across instruction‑following and reasoning‑enhanced models.
Since 2016, Alipay's security team has built the Alipay‑SDL 1.0 framework and now integrates AI and large‑model technologies to automate risk identification, enhance security tools, and streamline operations across the entire software development lifecycle, addressing rising business complexity and engineer workload.
At the ACM CCS 2025 live paper showcase, three groundbreaking studies—Gibbon’s fast secure two‑party GBDT training, Panther’s efficient private approximate nearest‑neighbor search on a single server, and PromeFuzz’s knowledge‑driven LLM approach to fuzzing harness generation—are presented, highlighting significant performance and security advances.
Ant Group’s award‑winning ‘Aspect Fusion Intelligence’ system, unveiled at the 2025 Zhejiang Cybersecurity Promotion Week, leverages AI, large models, blockchain and zero‑trust principles to tackle trillion‑scale data, rapid threat response, and financial‑grade reliability, setting new standards for intrusion detection.
The 2025 Inclusion Bund conference revealed a rapidly evolving open‑source large‑model ecosystem, highlighting 114 top projects across 22 fields, a 35% update rate, global developer distribution, community initiatives like Modak and Silicon Flow, and forward‑looking discussions on Vibe Coding, AI Agents, and the future of AI development.
Ring-mini-2.0 is a high‑performance inference MoE model that activates only 1.4 B parameters out of 16 B total, achieving dense‑model quality below 10 B while supporting 128 K context length and ultra‑fast generation speeds of over 300 tokens/s.
The 2025 Inclusion AI Open‑Source Forum unveiled a data‑driven 2.0 report mapping 114 top projects across 22 fields, highlighting rapid growth, regional strategy splits, a surge in AI coding tools, and a detailed timeline of large‑model releases shaping the future of AI development.
Ant Group and Renmin University unveiled LLaDA‑MoE, the industry’s first native MoE‑based diffusion language model trained on 20 TB of data, achieving performance comparable to Qwen2.5 while delivering several‑fold faster inference, and the model will be fully open‑sourced to accelerate global AI research.
At the 2025 Inclusion·Bund Conference forum, leading scholars and industry experts revealed how high‑quality data and AI form a dual‑engine that reshapes model training, improves performance, and drives the next evolution of intelligent systems.
