Search papers, labs, and topics across Lattice.
This paper introduces a novel distributed learning framework that enhances the extraction of diverse and discriminative representations from data samples in large-scale networks. By reformulating the global optimization function with constraints on representation variance and employing a primal-dual approach, the method effectively addresses the limitations of conventional task-specific approaches, particularly in non-i.i.d. scenarios through clustering and block coordinate descent. Extensive simulations on datasets like MNIST and CIFAR demonstrate the framework's capability to maintain optimal solutions with guaranteed convergence while capturing global structural representations.
Achieving diverse and discriminative representations in distributed learning could redefine how we approach complex classification tasks across diverse data distributions.
In large-scale distributed scenarios, increasingly complex tasks demand more intelligent collaboration across networks, requiring the joint extraction of structural representations from data samples. However, conventional task-specific approaches often result in nonstructural embeddings, leading to collapsed variability among data samples within the same class, particularly in classification tasks. To address this issue and fully leverage the intrinsic structure of data for downstream applications, we propose a novel distributed learning framework that ensures both diverse and discriminative representations. For independent and identically distributed (i.i.d.) data, we reformulate and decouple the global optimization function by introducing constraints on representation variance. The update rules are then derived and simplified using a primal-dual approach. For non-i.i.d. data distributions, we tackle the problem by clustering and virtually replicating nodes, allowing model updates within each cluster using block coordinate descent. In both cases, the resulting optimal solutions are theoretically proven to maintain discriminative and diverse properties, with a guaranteed convergence for i.i.d. conditions. Additionally, semantic information from representations is shared among nodes, reducing the need for common neural network architectures. Finally, extensive simulations on MNIST, CIFAR-10 and CIFAR-100 confirm the effectiveness of the proposed algorithms in capturing global structural representations.
