We introduce EV3, a novel meta-optimization framework designed to efficiently train scalable machine learning models through an intuitive explore-assess-adapt protocol. In each iteration of EV3, we explore various model parameter updates, assess them using pertinent evaluation methods, and then adapt the model based on the optimal updates and previous progress history. EV3 offers substantial flexibility without imposing stringent constraints like differentiability on the key objectives relevant to the tasks of interest, allowing for exploratory updates with intentionally-biased gradients and through a diversity of losses and optimizers. Additionally, the assessment phase provides reliable safety controls to ensure robust generalization, and can dynamically prioritize tasks in scenarios with multiple objectives. With inspiration drawn from evolutionary algorithms, meta-learning, and neural architecture search, we investigate an application of EV3 to knowledge distillation. Our experimental results illustrate EV3's capability to safely explore the modeling landscape, while hinting at its potential applicability across numerous domains due to its inherent flexibility and adaptability. Finally, we provide a JAX implementation of EV3, along with source code for experiments, available at: https://github.com/google-research/google-research/tree/master/ev3.

翻译：我们提出EV3，这是一种新颖的元优化框架，通过直观的“探索-评估-适应”协议，旨在高效训练可扩展机器学习模型。在EV3的每次迭代中，我们探索各种模型参数更新，使用相关评估方法对其进行评估，然后基于最优更新和先前进展历史调整模型。EV3具有显著的灵活性，无需对任务关键目标施加可微性等严格约束，允许通过有意偏置的梯度、多种损失函数和优化器进行探索性更新。此外，评估阶段提供可靠的安全控制以确保稳健泛化，并在多目标场景中动态优化任务优先级。受进化算法、元学习和神经架构搜索的启发，我们研究了EV3在知识蒸馏中的应用。实验结果表明，EV3能够安全探索建模空间，同时因其固有灵活性和适应性，展现出在众多领域的潜在适用性。最后，我们提供EV3的JAX实现及实验源代码，访问地址：https://github.com/google-research/google-research/tree/master/ev3。