学习调制随机权重：受神经调制启发的神经网络用于高效持续学习
Learning to Modulate Random Weights: Neuromodulation-inspired Neural Networks For Efficient Continual Learning
AI总结：
受生物神经调制启发，提出一种持续学习新架构，仅学习少量情境参数来调制随机权重，以高效解决灾难性遗忘并加速训练。
摘要：
现有的持续学习（CL）方法主要集中于通过利用正则化方法、回放缓冲区和任务特定组件来解决灾难性遗忘问题。然而，现实的持续学习解决方案不仅应受到灾难性遗忘指标的制约，还应考虑计算效率和运行时间。在此，我们引入了一种受生物神经系统中的神经调制启发的新型神经网络架构，以经济高效地解决灾难性遗忘问题，并为解释学习到的表征提供新途径。神经调制是一种在机器学习中受到有限关注的生物机制；它实时动态控制和微调突触动力学，以跟踪不同行为情境的需求。受此启发，我们提出的架构为每个任务情境学习一组相对较小的参数，这些参数对将输入转换的不变的随机权重的活动进行神经调制。我们表明，尽管可学习参数的数量非常少，但这种方法在每个任务上都具有强大的学习性能。此外，由于情境向量非常紧凑，多个网络可以同时存储，互不干扰且占用空间很小，从而完全消除了灾难性遗忘并加速了训练过程。
英文摘要：
Existing Continual Learning (CL) approaches have focused on addressing catastrophic forgetting by leveraging regularization methods, replay buffers, and task-specific components. However, realistic CL solutions must be shaped not only by metrics of catastrophic forgetting but also by computational efficiency and running time. Here, we introduce a novel neural network architecture inspired by neuromodulation in biological nervous systems to economically and efficiently address catastrophic forgetting and provide new avenues for interpreting learned representations. Neuromodulation is a biological mechanism that has received limited attention in machine learning; it dynamically controls and fine tunes synaptic dynamics in real time to track the demands of different behavioral contexts. Inspired by this, our proposed architecture learns a relatively small set of parameters per task context that \emph{neuromodulates} the activity of unchanging, randomized weights that transform the input. We show that this approach has strong learning performance per task despite the very small number of learnable parameters. Furthermore, because context vectors are so compact, multiple networks can be stored concurrently with no interference and little spatial footprint, thus completely eliminating catastrophic forgetting and accelerating the training process.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
