Model compression has been widely adopted to obtain light-weighted deep neural networks. Most prevalent methods, however, require fine-tuning with sufficient training data to ensure accuracy, which could be challenged by privacy and security issues. As a compromise between privacy and performance, in this paper we investigate few shot network compression: given few samples per class, how can we effectively compress the network with negligible performance drop? The core challenge of few shot network compression lies in high estimation errors from the original network during inference, since the compressed network can easily over-fits on the few training instances. The estimation errors could propagate and accumulate layer-wisely and finally deteriorate the network output. To address the problem, we propose cross distillation, a novel layer-wise knowledge distillation approach. By interweaving hidden layers of teacher and student network, layer-wisely accumulated estimation errors can be effectively reduced. The proposed method offers a general framework compatible with prevalent network compression techniques such as pruning. Extensive experiments on benchmark datasets demonstrate that cross distillation can significantly improve the student network's accuracy when only a few training instances are available.

Generalisation of a deep neural network (DNN) is one major concern when employing the deep learning approach for solving practical problems. In this paper we propose a new technique, named approximated orthonormal normalisation (AON), to improve the generalisation capacity of a DNN model. Considering a weight matrix W from a particular neural layer in the model, our objective is to design a function h(W) such that its row vectors are approximately orthogonal to each other while allowing the DNN model to fit the training data sufficiently accurate. By doing so, it would avoid co-adaptation among neurons of the same layer to be able to improve network-generalisation capacity. Specifically, at each iteration, we first approximate (WW^T)^(-1/2) using its Taylor expansion before multiplying the matrix W. After that, the matrix product is then normalised by applying the spectral normalisation (SN) technique to obtain h(W). Conceptually speaking, AON is designed to turn orthonormal regularisation into orthonormal normalisation to avoid manual balancing the original and penalty functions. Experimental results show that AON yields promising validation performance compared to orthonormal regularisation.

Established recurrent neural networks are well-suited to solve a wide variety of prediction tasks involving discrete sequences. However, they do not perform as well in the task of dynamical system identification, when dealing with observations from continuous variables that are unevenly sampled in time, for example due to missing observations. We show how such neural sequence models can be adapted to deal with variable step sizes in a natural way. In particular, we introduce a time-aware and stationary extension of existing models (including the Gated Recurrent Unit) that allows them to deal with unevenly sampled system observations by adapting to the observation times, while facilitating higher-order temporal behavior. We discuss the properties and demonstrate the validity of the proposed approach, based on samples from two industrial input/output processes.

Joint models are a common and important tool in the intersect ion of machine learning and the physical sciences, particularly in contex ts where real-world measurements are scarce. Recent developments in rainfall-run off modeling, one of the prime challenges in hydrology, show the value of a joint m odel with shared representation in this important context. However, curren t state-of-the-art models depend on detailed and reliable attributes characteriz ing each site to help the model differentiate correctly between the behavior of diff erent sites. This dependency can present a challenge in data-poor regions. In this p aper, we show that we can replace the need for such location-specific attributes w ith a completely data-driven learned embedding, and match previous state-of-the -art results with less information.

ABSTRACT Predicting smartphone users activity using WiFi fingerprints has been a popular approach for indoor positioning in recent years. However, such a high dimensional time-series prediction problem can be very tricky to solve. To address this issue, we propose a novel deep learning model, the convolutional mixture density recurrent neural network (CMDRNN), which combines the strengths of convolutional neural networks, recurrent neural networks and mixture density networks. In our model, the CNN sub-model is employed to detect the feature of the high dimensional input, the RNN sub-model is utilized to capture the time dependency and the MDN sub-model is for predicting the final output. For validation, we conduct the experiments on the real-world dataset and the obtained results illustrate the effectiveness of our method.

Due to the recent advances on Neural Architecture Search (NAS), it gains popularity in designing best networks for specific tasks. Although it shows promising results on many benchmarks and competitions, NAS still suffers from its demanding computation cost for searching high dimensional architectural design space, and this problem becomes even worse when we want to use a large-scale dataset. If we can make a reliable data proxy for NAS, the efficiency of NAS approaches increase accordingly. Our basic observation for making a data proxy is that each example in a specific dataset has a different impact on NAS process and most of examples are redundant from a relative accuracy ranking perspective, which we should preserve when making a data proxy. We propose a systematic approach to measure the importance of each example from this relative accuracy ranking point of view, and make a reliable data proxy based on the statistics of training and testing examples. Our experiment shows that we can preserve the almost same relative accuracy ranking between all possible network configurations even with 10-20$\times$ smaller data proxy.

Uncertainty propagation in nonlinear dynamic systems remains an outstanding problem in scientific computing and control. Numerous approaches have been developed, but are limited in their capability to tackle problems with more than a few uncertain variables or require large amounts of simulation data. In this paper, we propose a data-driven method for approximating joint probability density functions (PDFs) of nonlinear dynamic systems with initial condition and parameter uncertainty. Our approach leverages on the power of deep learning to deal with high-dimensional inputs, but we overcome the need for huge quantities of training data by encoding PDF evolution equations directly into the optimization problem. We demonstrate the potential of the proposed method by applying it to evaluate the robustness of a feedback controller for a six-dimensional rigid body with parameter uncertainty.

Regularization plays a crucial role in machine learning models, especially for deep neural networks. The existing regularization techniques mainly reply on the i.i.d. assumption and only employ the information of the current sample, without the leverage of neighboring information between samples. In this work, we propose a general regularizer called Patch-level Neighborhood Interpolation~(\textbf{Pani}) that fully exploits the relationship between samples. Furthermore, by explicitly constructing a patch-level graph in the different network layers and interpolating the neighborhood features to refine the representation of the current sample, our Patch-level Neighborhood Interpolation can then be applied to enhance two popular regularization strategies, namely Virtual Adversarial Training (VAT) and MixUp, yielding their neighborhood versions. The first derived \textbf{Pani VAT} presents a novel way to construct non-local adversarial smoothness by incorporating patch-level interpolated perturbations. In addition, the \textbf{Pani MixUp} method extends the original MixUp regularization to the patch level and then can be developed to MixMatch, achieving the state-of-the-art performance. Finally, extensive experiments are conducted to verify the effectiveness of the Patch-level Neighborhood Interpolation in both supervised and semi-supervised settings.

To reduce the long training time of large deep neural network (DNN) models, distributed synchronous stochastic gradient descent (S-SGD) is commonly used on a cluster of workers. However, the speedup brought by multiple workers is limited by the communication overhead. Two approaches, namely pipelining and gradient sparsification, have been separately proposed to alleviate the impact of communication overheads. Yet, the gradient sparsification methods can only initiate the communication after the backpropagation, and hence miss the pipelining opportunity. In this paper, we propose a new distributed optimization method named LAGS-SGD, which combines S-SGD with a novel layer-wise adaptive gradient sparsification (LAGS) scheme. In LAGS-SGD, every worker selects a small set of "significant" gradients from each layer independently whose size can be adaptive to the communication-to-computation ratio of that layer. The layer-wise nature of LAGS-SGD opens the opportunity of overlapping communications with computations, while the adaptive nature of LAGS-SGD makes it flexible to control the communication time. We prove that LAGS-SGD has convergence guarantees and it has the same order of convergence rate as vanilla S-SGD under a weak analytical assumption. Extensive experiments are conducted to verify the analytical assumption and the convergence performance of LAGS-SGD. Experimental results show that LAGS-SGD achieves from around 40\% to 95\% of the maximum benefit of pipelining on a 16-node GPU cluster. Combining the benefit of pipelining and sparsification, the speedup of LAGS-SGD over S-SGD ranges from 2.86$\times$ to 8.52$\times$ on our tested CNN and LSTM models, without losing obvious model accuracy.

Deep learning has achieved impressive prediction performance in the field of sequence learning recently. Dissolved oxygen prediction, as a kind of time-series forecasting, is suitable for this technique. Although many researchers have developed hybrid models or variant models based on deep learning techniques, there is no comprehensive and sound comparison among the deep learning models in this field currently. Plus, most previous studies focused on one-step forecasting by using a small data set. As the convenient access to high-frequency data, this paper compares multi-step deep learning forecasting by using walk-forward validation. Specifically, we test Convolutional Neural Network (CNN), Temporal Convolutional Network (TCN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Bidirectional Recurrent Neural Network (BiRNN) based on the real-time data recorded automatically at a fixed observation point in the Yangtze River from 2012 to 2016. By comparing the average accumulated statistical metrics of root mean square error (RMSE), mean absolute error (MAE), and coefficient of determination in each time step, We find for multi-step time series forecasting, the average performance of each time step does not decrease linearly. GRU outperforms other models with significant advantages.