Intern-S2-Preview：科学智能体基础模型
Intern-S2-Preview: Scientific Agentic Foundation Model
浏览论文内容
中文总结 AI 辅助
本研究提出Intern-S2-Preview系列科学智能体基础模型，通过多阶段训练与架构优化，在多类基准上取得领先结果，相关模块可提升科学任务表现且无需修改主干模型。
中文摘要 AI 辅助
科学发现越来越需要能够对异构模态的科学证据进行推理、与科学工具和环境交互、并在长期任务中持续推进的AI系统。我们提出Intern-S2-Preview，这是一系列旨在支持多模态科学理解、推理、生成及长期任务的科学智能体基础模型。其训练流程始于对渲染科学文档、交错图文数据及各类科学语料的科学多模态预训练；从预训练检查点出发，我们采用统一的后训练流程，包含监督微调、可扩展多任务强化学习（RL）、黑盒与白盒智能体RL，以及在线蒸馏。该流程辅以提升推理与训练稳定性及效率的实用技术，包括带离策略校正的部分推理、自适应长度正则化、在线投机解码、鲁棒多任务优化，以及面向智能体任务的轨迹感知经验组装。架构层面，Intern-S2-Preview-397B将时间序列建模从高效长序列理解扩展至数值预测，同时研究了Memory Decoder作为独立的记忆增强路径，以实现快速科学专业化，且无需修改冻结的397B主干。在科学、多模态、智能体及通用基准上的评估显示，Intern-S2-Preview-397B在多种场景下取得了具竞争力或领先的结果：时间序列模块提升了SciTS上的科学信号理解与预测，而独立的Intern-MemDec-4B扩展无需修改冻结的397B主干，便将Biology-Instructions的平均分数从56.92提升至60.32。
英文摘要
Scientific discovery increasingly requires AI systems that can reason over scientific evidence of heterogeneous modalities, interact with scientific tools and environments, and sustain progress across long task horizons. We present Intern-S2-Preview, a series of scientific agentic foundation models designed to support multimodal scientific understanding, reasoning, generation, and long-horizon tasks. The training pipeline begins with scientific multimodal pre-training over rendered scientific documents, interleaved image-text data, and diverse scientific corpora. Starting from the pretrained checkpoint, we apply a unified post-training pipeline consisting of supervised fine-tuning, scalable multi-task reinforcement learning (RL), black- and white-box agentic RL, and on-policy distillation. This pipeline is supported by practical techniques that improve rollout and training stability and efficiency, including partial rollout with off-policy correction, adaptive length regularization, online speculative decoding, robust multi-task optimization, and trace-aware experience assembly for agentic tasks. At the architecture level, Intern-S2-Preview-397B extends time series modelling from efficient long-sequence understanding to numerical forecasting, while Memory Decoder is studied as a separate memory-augmented path for rapid scientific specialization without modifying the frozen 397B backbone. Evaluations across scientific, multimodal, agentic, and general-purpose benchmarks show that Intern-S2-Preview-397B achieves competitive or leading results in multiple settings. The time series modules improve scientific signal understanding and forecasting on SciTS, while the separate Intern-MemDec-4B extension improves the Biology-Instructions average score from 56.92 to 60.32 without modifying the frozen 397B backbone.
发表机构
- Shanghai AI Laboratory（上海人工智能实验室）
机构由 AI 辅助整理，请以论文原文为准。
