High-capacity Architectures Research Articles

CERN: Compact facial expression recognition net

Since the renaissance of deep learning (DL), facial expression recognition (FER) has received a lot of interest, with continual improvement in the performance. Hand-in-hand with performance, new challenges have come up. Modern FER systems deal with face images captured under uncontrolled conditions (also called in-the-wild scenario) including occlusions and pose variations. They successfully handle such conditions using deep networks that come with various components like transfer learning, attention mechanism and local-global context extractor. However, these deep networks are highly complex with large number of parameters, making them unfit to be deployed in real scenarios. Is it possible to build a light-weight network that can still show significantly good performance on FER under in-the-wild scenario? In this work, we methodically build such a network and call it as Compact Expression Recognition Net (CERN). We leverage on the aforementioned components of deep networks for FER, and analyse, and appropriately fit them to arrive at CERN. Our CERN is a low-calorie net with only 1.45M parameters, which is almost 50x less than that of a state-of-the-art (SOTA) architecture. It requires only 17MB of storage. Further, during inference, it can process at the real time rate of 40 frames per second (fps) in an intel-i7 cpu. Though it is low-calorie, it is still power-packed in its performance, overpowering other light-weight architectures, and even few high capacity architectures. Specifically, CERN reports 87.09%, 88.17% and 62.06% accuracies on in-the-wild datasets RAFDB, FERPlus and AffectNet respectively. It also exhibits superior robustness under occlusions and pose variations in comparison to other light-weight architectures from the literature. Codes are publicly available at https://github.com/1980x/CFERNet.

A Novel Method for Function Smoothness in Neural Networks

Existing methods for function smoothness in neural networks have limitations. These methods can make training sensitive to their hyperparameters, or their smoothness constraints can limit model capacity. These methods can impose too much smoothness, even in areas without data, or they can impose non-meaningful smoothness constraints. The way these methods measure smoothness can also be computationally hard. One of the main methods for function smoothness, Lipschitz continuity, does not even imply differentiability theoretically, let alone continuous differentiability, that is smoothness. In this paper, we propose a method based on the theoretical definition of the derivative to ensure that the derivative of the parametrized function should tend toward its theoretical value for the given neural network parameters in the vicinity of training samples. The method changes the classifier and its training minimally and has no added hyperparameters. The proposed method is shown to achieve a smoother function in the vicinity of both training and testing samples for all tested datasets, as measured with decreased values of the Frobenius norm of the Jacobian with respect to inputs. Due to the correlation between function smoothness and generalization, the method makes classifiers generalize better and achieve higher accuracy than default classifiers for Restricted ImageNet, CIFAR10 and MNIST. Due to the correlation between function smoothness and adversarial robustness, the proposed method makes classifiers with high-capacity architecture more robust to adversarial samples generated with the PGD attack compared to default classifiers for the Restricted ImageNet, CIFAR10, Fashion-MNIST and MNIST datasets.

Performance Analysis of Analog IF Over Fiber Fronthaul Link With 4G and 5G Coexistence

Fifth generation (5G) mobile communications will require a dense deployment of small cell antenna sites and higher channel bandwidth, in conjunction with a cloud radio access network (C-RAN) architecture. This necessitates low latency and high-capacity architecture in addition to energy- and cost-efficient fronthaul links. An efficient way of achieving such connectivity is to make use of an opticalfiber- based infrastructure where multiple wireless services may be distributed over the same fiber to remote radio head (RRH) sites. In this work, we demonstrate the spectral containment of fourth generation (4G) Long-Term Evolution (LTE) signals and 5G candidate waveforms—generalized frequency division multiplexing and universally filtered orthogonal frequency division multiplexing (UF-OFDM) through a directly modulated link. Seventy-five bands of LTEand 10 bands of 5Gwaveforms are successfully transmitted over a 25 km analog intermediate frequency signal over fiber (AIFoF) link through our setup, limited only by the bandwidth of the laser. For the first time, to the best of our knowledge, we demonstrate the fronthaul network for providing simultaneous 4G and 5G services by propagating LTE signals in coexistence with UF-OFDM.