EMAN：通过多任务学习中的路径涌现实现优化驱动的容量增长
EMAN: Optimization-Driven Capacity Growth through Path Emergence in Multi-Task Learning
浏览论文内容
中文总结 AI 辅助
该研究针对现有多任务学习方法容量变化受限的问题，提出优化驱动的EMAN框架，通过潜在相对相位实现路径涌现，在多数据集实验中以竞争力的计算成本提升了性能。
中文摘要 AI 辅助
现有多任务学习方法依赖硬共享、多路径或专家、自适应共享及动态扩展，但它们的容量变化通常受预定义结构限制，或由任务边界与冲突信号触发。这引出一个根本问题：网络能否从精确的单路径计算开始，仅在出现持续优化证据时才生成新的独立路径？我们提出Emergent Modular Atomic Network（EMAN），这是一个优化驱动框架，通过潜在相对相位暴露反对称增长方向，无需实例化第二条路径，并在训练期间监控多个决策信号，将局部优化证据转化为结构决策。EMAN仅在验证后才实现两条容量相等的独立路径，自适应分配共享与任务特定的表示容量以适应不同任务需求。在受控秩设置、PASCAL-Context和NYUv2上的大量实验验证了其有效性，在具有竞争力的计算成本下实现了性能提升。
英文摘要
Existing multi-task learning methods rely on hard sharing, multiple paths or experts, adaptive sharing, and dynamic expansion. However, their capacity changes are usually constrained by predefined structures or triggered by task boundaries and conflict signals. This raises a fundamental question: can a network start from exact single-path computation and grow a new independent path only when persistent optimization evidence appears? We propose the Emergent Modular Atomic Network (EMAN), an optimization-driven framework for exposing an antisymmetric growth direction through latent relative phases without instantiating a second path, and for monitoring multiple decision signals during training to transform local optimization evidence into a structural decision. EMAN materializes two equal-capacity independent paths only after certification. EMAN adaptively allocates shared and task-specific representation capacity to accommodate varying task requirements. Extensive experiments on controlled rank settings, PASCAL-Context, and NYUv2 validate its effectiveness, achieving improved performance at a competitive computational cost.
