Neural Information Processing Systems http://nips.cc/

This paper introduces stochastic recurrent neural networks which glue a deterministic recurrent neural network and a state space model together to form a stochastic and sequential neural generative model. The clear separation of deterministic and stochastic layers allows a structured variational inference network to track the factorization of the model's posterior distribution. By retaining both the nonlinear recursive structure of a recurrent neural network and averaging over the uncertainty in a latent path, like a state space model, we improve the state of the art results on the Blizzard and TIMIT speech modeling data sets by a large margin, while achieving comparable performances to competing methods on polyphonic music modeling.

Neural Information Processing Systems http://nips.cc/

Neural Information Processing Systems http://nips.cc/

As privacy protection gains increasing importance, more models are being trained on edge devices and subsequently merged into the central server through Federated Learning (FL). However, current research overlooks the impact of network topology, physical distance, and data heterogeneity on edge devices, leading to issues such as increased latency and degraded model performance. To address these issues, we propose a new federated learning scheme on edge devices that called Federated Learning with Encrypted Data Sharing(FedEDS). FedEDS uses the client model and the model's stochastic layer to train the data encryptor. The data encryptor generates encrypted data and shares it with other clients. The client uses the corresponding client's stochastic layer and encrypted data to train and adjust the local model. FedEDS uses the client's local private data and encrypted shared data from other clients to train the model. This approach accelerates the convergence speed of federated learning training and mitigates the negative impact of data heterogeneity, making it suitable for application services deployed on edge devices requiring rapid convergence. Experiments results show the efficacy of FedEDS in promoting model performance.

This paper presents a model and method which are likely to be of interest to many in the community. The formulation allows stochastic layers of a convolution-deconvolution structured autoencoder with stochastic layers to be parameterized and manipulated in such a way that inference can be performed in a much more computationally efficient way compared to Gibbs sampling and MCEM techniques. Results on CIFAR are fairly far from state of the art, but illustrate the key contributions related to efficient inference well. Results for a few other methods would be useful to include in Table 1 to give the reader some context as to what regime this approach and the examined model are operating within. The Flickr8k, 30k and MS COCO results seem quite strong, especially given that this work does not focus on the application, but rather the proposed VAE method.