Learning an explainable classifier often results in low accuracy model or ends up with a huge rule set, while learning a deep model is usually more capable of handling noisy data at scale, but with the cost of hard to explain the result and weak at generalization. To mitigate this gap, we propose an end-to-end deep explainable learning approach that combines the advantage of deep model in noise handling and expert rule-based interpretability. Specifically, we propose to learn a deep data assessing model which models the data as a graph to represent the correlations among different observations, whose output will be used to extract key data features. The key features are then fed into a rule network constructed following predefined noisy expert rules with trainable parameters. As these models are correlated, we propose an end-to-end training framework, utilizing the rule classification loss to optimize the rule learning model and data assessing model at the same time. As the rule-based computation is none-differentiable, we propose a gradient linking search module to carry the gradient information from the rule learning model to the data assessing model. The proposed method is tested in an industry production system, showing comparable prediction accuracy, much higher generalization stability and better interpretability when compared with a decent deep ensemble baseline, and shows much better fitting power than pure rule-based approach.
This article presents a hardware platform including stimulating implants wirelessly powered and controlled by a shared transmitter (TX) for coordinated leadless multisite stimulation. The adopted novel single-TX, multiple-implant structure can flexibly deploy stimuli, improve system efficiency, easily scale stimulating channel quantity, and relieve efforts in device synchronization. In the proposed system, a wireless link leveraging magnetoelectric (ME) effect is co-designed with a robust and efficient system-on-chip (SoC) to enable reliable operation and individual programming of every implant. Each implant integrates a 0.8-mm2 chip, a 6-mm2 ME film, and an energy storage capacitor within a 6.2-mm3 size. ME power transfer is capable of safely transmitting milliwatt power to devices placed several centimeters away from the TX coil, maintaining good efficiency with size constraints, and tolerating 60 degree, 1.5-cm misalignment in angular and lateral movement. The SoC robustly operates with 2-V source amplitude variations that spans a 40-mm TX-implant distance change, realizes individual addressability through physical unclonable function (PUF) IDs, and achieves 90% efficiency for 1.5-3.5-V stimulation with fully programmable stimulation parameters.
This paper presents the first wireless and programmable neural stimulator leveraging magnetoelectric (ME) effects for power and data transfer. Thanks to low tissue absorption, low misalignment sensitivity and high power transfer efficiency, the ME effect enables safe delivery of high power levels (a few milliwatts) at low resonant frequencies (~250 kHz) to mm-sized implants deep inside the body (30-mm depth). The presented MagNI (Magnetoelectric Neural Implant) consists of a 1.5-mm$^2$ 180-nm CMOS chip, an in-house built 4x2 mm ME film, an energy storage capacitor, and on-board electrodes on a flexible polyimide substrate with a total volume of 8.2 mm$^3$ . The chip with a power consumption of 23.7 $\mu$W includes robust system control and data recovery mechanisms under source amplitude variations (1-V variation tolerance). The system delivers fully-programmable bi-phasic current-controlled stimulation with patterns covering 0.05-to-1.5-mA amplitude, 64-to-512-$\mu$s pulse width, and 0-to-200Hz repetition frequency for neurostimulation.
A compact, accurate, and bitwidth-programmable in-memory computing (IMC) static random-access memory (SRAM) macro, named CAP-RAM, is presented for energy-efficient convolutional neural network (CNN) inference. It leverages a novel charge-domain multiply-and-accumulate (MAC) mechanism and circuitry to achieve superior linearity under process variations compared to conventional IMC designs. The adopted semi-parallel architecture efficiently stores filters from multiple CNN layers by sharing eight standard 6T SRAM cells with one charge-domain MAC circuit. Moreover, up to six levels of bit-width of weights with two encoding schemes and eight levels of input activations are supported. A 7-bit charge-injection SAR (ciSAR) analog-to-digital converter (ADC) getting rid of sample and hold (S&H) and input/reference buffers further improves the overall energy efficiency and throughput. A 65-nm prototype validates the excellent linearity and computing accuracy of CAP-RAM. A single 512x128 macro stores a complete pruned and quantized CNN model to achieve 98.8% inference accuracy on the MNIST data set and 89.0% on the CIFAR-10 data set, with a 573.4-giga operations per second (GOPS) peak throughput and a 49.4-tera operations per second (TOPS)/W energy efficiency.
Object detection has been widely used in the field of Internet, and deep learning plays a very important role in object detection. However, the existing object detection methods need to be trained in the static setting, which requires obtaining all the data at one time, and it does not support training in the way of class-incremental. In this paper, an object detection framework named class-incremental object detection (CIOD) is proposed. CIOD divides object detection into two stages. Firstly, the traditional OpenCV cascade classifier is improved in the object candidate box generation stage to meet the needs of class increment. Secondly, we use the concept of prototype vector on the basis of deep learning to train a classifier based on class-incremental to identify the generated object candidate box, so as to extract the real object box. A large number of experiments on CIOD have been carried out to verify that CIOD can detect the object in the way of class-incremental and can control the training time and memory capacity.
Incorporating deep neural networks in image compressive sensing (CS) receives intensive attentions recently. As deep network approaches learn the inverse mapping directly from the CS measurements, a number of models have to be trained, each of which corresponds to a sampling rate. This may potentially degrade the performance of image CS, especially when multiple sampling rates are assigned to different blocks within an image. In this paper, we develop a multi-channel deep network for block-based image CS with performance significantly exceeding the current state-of-the-art methods. The significant performance improvement of the model is attributed to block-based sampling rates allocation and model-level removal of blocking artifacts. Specifically, the image blocks with a variety of sampling rates can be reconstructed in a single model by exploiting inter-block correlation. At the same time, the initially reconstructed blocks are reassembled into a full image to remove blocking artifacts within the network by unrolling a hand-designed block-based CS algorithm. Experimental results demonstrate that the proposed method outperforms the state-of-the-art CS methods by a large margin in terms of objective metrics, PSNR, SSIM, and subjective visual quality.
This paper introduces a model based upon games on an evolving network, and develops three clustering algorithms according to it. In the clustering algorithms, data points for clustering are regarded as players who can make decisions in games. On the network describing relationships among data points, an edge-removing-and-rewiring (ERR) function is employed to explore in a neighborhood of a data point, which removes edges connecting to neighbors with small payoffs, and creates new edges to neighbors with larger payoffs. As such, the connections among data points vary over time. During the evolution of network, some strategies are spread in the network. As a consequence, clusters are formed automatically, in which data points with the same evolutionarily stable strategy are collected as a cluster, so the number of evolutionarily stable strategies indicates the number of clusters. Moreover, the experimental results have demonstrated that data points in datasets are clustered reasonably and efficiently, and the comparison with other algorithms also provides an indication of the effectiveness of the proposed algorithms.
Enormous successes have been made by quantum algorithms during the last decade. In this paper, we combine the quantum game with the problem of data clustering, and then develop a quantum-game-based clustering algorithm, in which data points in a dataset are considered as players who can make decisions and implement quantum strategies in quantum games. After each round of a quantum game, each player's expected payoff is calculated. Later, he uses a link-removing-and-rewiring (LRR) function to change his neighbors and adjust the strength of links connecting to them in order to maximize his payoff. Further, algorithms are discussed and analyzed in two cases of strategies, two payoff matrixes and two LRR functions. Consequently, the simulation results have demonstrated that data points in datasets are clustered reasonably and efficiently, and the clustering algorithms have fast rates of convergence. Moreover, the comparison with other algorithms also provides an indication of the effectiveness of the proposed approach.
We have proposed a model based upon flocking on a complex network, and then developed two clustering algorithms on the basis of it. In the algorithms, firstly a \textit{k}-nearest neighbor (knn) graph as a weighted and directed graph is produced among all data points in a dataset each of which is regarded as an agent who can move in space, and then a time-varying complex network is created by adding long-range links for each data point. Furthermore, each data point is not only acted by its \textit{k} nearest neighbors but also \textit{r} long-range neighbors through fields established in space by them together, so it will take a step along the direction of the vector sum of all fields. It is more important that these long-range links provides some hidden information for each data point when it moves and at the same time accelerate its speed converging to a center. As they move in space according to the proposed model, data points that belong to the same class are located at a same position gradually, whereas those that belong to different classes are away from one another. Consequently, the experimental results have demonstrated that data points in datasets are clustered reasonably and efficiently, and the rates of convergence of clustering algorithms are fast enough. Moreover, the comparison with other algorithms also provides an indication of the effectiveness of the proposed approach.
The enormous successes have been made by quantum algorithms during the last decade. In this paper, we combine the quantum random walk (QRW) with the problem of data clustering, and develop two clustering algorithms based on the one dimensional QRW. Then, the probability distributions on the positions induced by QRW in these algorithms are investigated, which also indicates the possibility of obtaining better results. Consequently, the experimental results have demonstrated that data points in datasets are clustered reasonably and efficiently, and the clustering algorithms are of fast rates of convergence. Moreover, the comparison with other algorithms also provides an indication of the effectiveness of the proposed approach.