Sourav MISHRA Subhajit CHAUDHURY Hideaki IMAIZUMI Toshihiko YAMASAKI
Our paper attempts to critically assess the robustness of deep learning methods in dermatological evaluation. Although deep learning is being increasingly sought as a means to improve dermatological diagnostics, the performance of models and methods have been rarely investigated beyond studies done under ideal settings. We aim to look beyond results obtained on curated and ideal data corpus, by investigating resilience and performance on user-submitted data. Assessing via few imitated conditions, we have found the overall accuracy to drop and individual predictions change significantly in many cases despite of robust training.
In this paper, we clarify the importance of performance evaluation using a plurality of smartphones in a positioning system based on radio waves. Specifically, in a positioning system using bluetooth low energy, the positioning performance of two types of positioning algorithms is performed using a plurality of smartphones. As a result, we confirmed that the fingerprint algorithm does not always provide sufficient positioning performance. It depends on the model of the smartphone used. On the other hand, the hybrid algorithm that the authors have already proposed is robust in the difference of the received signal characteristics of the smartphone. Consequently, we spotlighted that the use of multiple devices is essential for providing high-quality location-based services in real environments in the performance evaluation of radio wave-based positioning systems using smartphones.
In this short note, we formally show that Keyed-Homomorphic Public Key Encryption (KH-PKE) is secure against key recovery attacks and ciphertext validity attacks that have been introduced as chosen-ciphertext attacks for homomorphic encryption.
A fully homomorphic encryption (FHE) would be the important cryptosystem as the basic scheme for the cloud computing. Since Gentry discovered in 2009 the first fully homomorphic encryption scheme, some fully homomorphic encryption schemes were proposed. In the systems proposed until now the bootstrapping process is the main bottleneck and the large complexity for computing the ciphertext is required. In 2011 Zvika Brakerski et al. proposed a leveled FHE without bootstrapping. But circuit of arbitrary level cannot be evaluated in their scheme while in our scheme circuit of any level can be evaluated. The existence of an efficient fully homomorphic cryptosystem would have great practical implications in the outsourcing of private computations, for instance, in the field of the cloud computing. In this paper, IND-CCA1secure FHE based on the difficulty of prime factorization is proposed which does not need the bootstrapping and it is thought that our scheme is more efficient than the previous schemes. In particular the computational overhead for homomorphic evaluation is O(1).
Expectation propagation (EP) decoding is proposed for sparse superposition coding in orthogonal frequency division multiplexing (OFDM) systems. When a randomized discrete Fourier transform (DFT) dictionary matrix is used, the EP decoding has the same complexity as approximate message-passing (AMP) decoding, which is a low-complexity and powerful decoding algorithm for the additive white Gaussian noise (AWGN) channel. Numerical simulations show that the EP decoding achieves comparable performance to AMP decoding for the AWGN channel. For OFDM systems, on the other hand, the EP decoding is much superior to the AMP decoding while the AMP decoding has an error-floor in high signal-to-noise ratio regime.
Keisuke MAEDA Kazaha HORII Takahiro OGAWA Miki HASEYAMA
A multi-task convolutional neural network leading to high performance and interpretability via attribute estimation is presented in this letter. Our method can provide interpretation of the classification results of CNNs by outputting attributes that explain elements of objects as a judgement reason of CNNs in the middle layer. Furthermore, the proposed network uses the estimated attributes for the following prediction of classes. Consequently, construction of a novel multi-task CNN with improvements in both of the interpretability and classification performance is realized.
In this paper, we propose a design method to design asynchronous circuits with bundled-data implementation on commercial Field Programmable Gate Arrays using placement constraints. The proposed method uses two types of placement constraints to reduce the number of delay adjustments to fix timing violations and to improve the performance of the bundled-data implementation. We also propose a floorplan algorithm to reduce the control-path delays specific to the bundled-data implementation. Using the proposed method, we could design the asynchronous circuits whose performance is close to and energy consumption is small compared to the synchronous counterparts with less delay adjustment.
In this paper, we propose a novel single-template strategy based on a mean template set and locally/globally weighted dynamic time warping (LG-DTW) to improve the performance of online signature verification. Specifically, in the enrollment phase, we implement a time series averaging method, Euclidean barycenter-based DTW barycenter averaging, to obtain a mean template set considering intra-user variability among reference samples. Then, we acquire a local weighting estimate considering a local stability sequence that is obtained analyzing multiple matching points of an optimal match between the mean template and reference sets. Thereafter, we derive a global weighting estimate based on the variable importance estimated by gradient boosting. Finally, in the verification phase, we apply both local and global weighting methods to acquire a discriminative LG-DTW distance between the mean template set and a query sample. Experimental results obtained on the public SVC2004 Task2 and MCYT-100 signature datasets confirm the effectiveness of the proposed method for online signature verification.
Lucas Saad Nogueira NUNES Jacir Luiz BORDIM Yasuaki ITO Koji NAKANO
The volume of digital information is growing at an extremely fast pace which, in turn, exacerbates the need of efficient mechanisms to find the presence of a pattern in an input text or a set of input strings. Combining the processing power of Graphics Processing Unit (GPU) with matching algorithms seems a natural alternative to speedup the string-matching process. This work proposes a Parallel Rabin-Karp implementation (PRK) that encompasses a fast-parallel prefix-sums algorithm to maximize parallelization and accelerate the matching verification. Given an input text T of length n and p patterns of length m, the proposed implementation finds all occurrences of p in T in O(m+q+n/τ+nm/q) time, where q is a sufficiently large prime number and τ is the available number of threads. Sequential and parallel versions of the PRK have been implemented. Experiments have been executed on p≥1 patterns of length m comprising of m=10, 20, 30 characters which are compared against a text string of length n=227. The results show that the parallel implementation of the PRK algorithm on NVIDIA V100 GPU provides speedup surpassing 372 times when compared to the sequential implementation and speedup of 12.59 times against an OpenMP implementation running on a multi-core server with 128 threads. Compared to another prominent GPU implementation, the PRK implementation attained speedup surpassing 37 times.
Myat Hsu AUNG Hiroshi TSUTSUI Yoshikazu MIYANAGA
In this paper, we propose a WiFi-based indoor positioning system using a fingerprint method, whose database is constructed with estimated reference locations. The reference locations and their information, called data sets in this paper, are obtained by moving reference devices at a constant speed while gathering information of available access points (APs). In this approach, the reference locations can be estimated using the velocity without any precise reference location information. Therefore, the cost of database construction can be dramatically reduced. However, each data set includes some errors due to such as the fluctuation of received signal strength indicator (RSSI) values, the device-specific WiFi sensitivities, the AP installations, and removals. In this paper, we propose a method to merge data sets to construct a consistent database suppressing such undesired effects. The proposed approach assumes that the intervals of reference locations in the database are constant and that the fingerprint for each reference location is calculated from multiple data sets. Through experimental results, we reveal that our approach can achieve an accuracy of 80%. We also show a detailed discussion on the results related parameters in the proposed approach.
Ryosuke SUGA Satoshi KURODA Atsushi KEZUKA
Authors had proposed a hybrid electromagnetic field analysis method suitable for an airport surface so far. In this paper, the hybrid method is validated by measurements by using a 1/50 scale-model of an airport considering several layouts of the buildings and sloping ground. The measured power distributions agreed with the analyzed ones within 5 dB errors excepting null points and the null positions of the distribution is also estimated within one wavelength errors.
Liping ZHANG Zongqing LU Qingmin LIAO
This paper proposes a new and effective convolutional neural network model termed OFR-Net for optical flow refinement. The OFR-Net exploits the spatial correlation between images and optical flow fields. It adopts a pyramidal codec structure with residual connections, dense connections and skip connections within and between the encoder and decoder, to comprehensively fuse features of different scales, locally and globally. We also introduce a warp loss to restrict large displacement refinement errors. A series of experiments on the FlyingChairs and MPI Sintel datasets show that the OFR-Net can effectively refine the optical flow predicted by various methods.
Daichi FURUBAYASHI Yuta KASHIWAGI Takanori SATO Tadashi KAWAI Akira ENOKIHARA Naokatsu YAMAMOTO Tetsuya KAWANISHI
A new structure of the electro-optic modulator to compensate the third-order intermodulation distortion (IMD3) is introduced. The modulator includes two Mach-Zehnder modulators (MZMs) operating with frequency chirp and the two modulated outputs are combined with an adequate phase difference. We revealed by theoretical analysis and numerical calculations that the IMD3 components in the receiver output could be selectively suppressed when the two MZMs operate with chirp parameters of opposite signs to each other. Spectral power of the IMD3 components in the proposed modulator was more than 15dB lower than that in a normal Mach-Zehnder modulator at modulation index between 0.15π and 0.25π rad. The IMD3 compensation properties of the proposed modulator was experimentally confirmed by using a dual parallel Mach-Zehnder modulator (DPMZM) structure. We designed and fabricated the modulator with the single-chip structure and the single-input operation by integrating with 180° hybrid coupler on the modulator substrate. Modulation signals were applied to each modulation electrode by the 180° hybrid coupler to set the chirp parameters of two MZMs of the DPMZM. The properties of the fabricated modulator were measured by using 10GHz two-tone signals. The performance of the IMD3 compensation agreed with that in the calculation. It was confirmed that the IMD3 compensation could be realized even by the fabricated modulator structure.
In recent years, deep neural network (DNN) has achieved considerable results on many artificial intelligence tasks, e.g. natural language processing. However, the computation complexity of DNN is extremely high. Furthermore, the performance of traditional von Neumann computing architecture has been slowing down due to the memory wall problem. Processing in memory (PIM), which places computation within memory and reduces the data movement, breaks the memory wall. ReRAM PIM is thought to be a available architecture for DNN accelerators. In this work, a novel design of ReRAM neuromorphic system is proposed to process DNN fully in array efficiently. The binary ReRAM array is composed of 2T2R storage cells and current mirror sense amplifiers. A dummy BL reference scheme is proposed for reference voltage generation. A binary DNN (BDNN) model is then constructed and optimized on MNIST dataset. The model reaches a validation accuracy of 96.33% and is deployed to the ReRAM PIM system. Co-design model optimization method between hardware device and software algorithm is proposed with the idea of utilizing hardware variance information as uncertainness in optimization procedure. This method is analyzed to achieve feasible hardware design and generalizable model. Deployed with such co-design model, ReRAM array processes DNN with high robustness against fabrication fluctuation.
Masahito SHIMAMOTO Yusuke KAMEDA Takayuki HAMAMOTO
We aim at HDR imaging with simple processing while preventing spatial resolution degradation in multiple-exposure-time image sensor where the exposure time is controlled for each pixel. The contributions are the proposal of image interpolation by motion area detection and pixel adaptive weighting method by overexposure and motion blur detection.
Chong WU Le ZHANG Houwang ZHANG Hong YAN
In this letter, we propose a hierarchical segmentation (HS) method for color images, which can not only maintain the segmentation accuracy, but also ensure a good speed. In our method, HS adopts the fuzzy simple linear iterative clustering (Fuzzy SLIC) to obtain an over-segmentation result. Then, HS uses the fast fuzzy C-means clustering (FFCM) to produce the rough segmentation result based on superpixels. Finally, HS takes the non-iterative K-means clustering using priority queue (KPQ) to refine the segmentation result. In the validation experiments, we tested our method and compared it with state-of-the-art image segmentation methods on the Berkeley (BSD500) benchmark under different types of noise. The experiment results show that our method outperforms state-of-the-art techniques in terms of accuracy, speed and robustness.
Advances in deep reinforcement learning have demonstrated its effectiveness in a wide variety of domains. Deep neural networks are capable of approximating value functions and policies in complex environments. However, deep neural networks inherit a number of drawbacks. Lack of interpretability limits their usability in many safety-critical real-world scenarios. Moreover, they rely on huge amounts of data to learn efficiently. This may be suitable in simulated tasks, but restricts their use to many real-world applications. Finally, their generalization capability is low, the ability to determine that a situation is similar to one encountered previously. We present a method to combine external knowledge and interpretable reinforcement learning. We derive a rule-based variant version of the Sarsa(λ) algorithm, which we call Sarsa-rb(λ), that augments data with prior knowledge and exploits similarities among states. We demonstrate that our approach leverages small amounts of prior knowledge to significantly accelerate the learning in multiple domains such as trading or visual navigation. The resulting agent provides substantial gains in training speed and performance over deep q-learning (DQN), deep deterministic policy gradients (DDPG), and improves stability over proximal policy optimization (PPO).
Haruka ARAKAWA Takashi TOMURA Jiro HIROKAWA
The sidelobe level at tilts around 30-40 degrees in both the E and H planes due to a tapered excitation of units of 2×2 radiation slots is suppressed by introducing slit layers over a corporate-feed waveguide slot array antenna. The slit layers act as averaging the excitation of the adjacent radiating slots for sidelobe suppression in both planes. A 16×16-element array in the 70GHz band is fabricated. At the design frequency, the sidelobe levels at tilts around 30-40 degrees are suppressed from -25.4dB to -31.3dB in the E-plane and from -27.1dB to -38.9dB in the H-plane simultaneously as confirmed by measurements. They are suppressed over the desired range of 71.0-76.0GHz frequencies, compared to the conventional antenna.
Shimpei SATO Kano AKAGI Atsushi TAKAHASHI
Routing problems derived from silicon-interposer and etc. are often formulated as a set-pair routing problem where the combination of pin-pairs to be connected is flexible. In this routing problem, a length matching routing pattern is often required due to the requirement of the signal propagation delays be the same. We propose a fast length matching routing method for the set-pair routing problem. The existing algorithm generates a good length matching routing pattern in practical time. However, due to the limited searching range, there are length matching routing patterns that cannot find due to the limited searching range of the algorithm. Also, it needs heavy iterative steps to improve a solution, and the computation time is practical but not fast. In the set-pair routing, although pin-pairs to be connected is flexible, it is expected that combinations of pin-pairs which realize length matching are restricted. In our method, such a combination of pin-pairs is selected in advance, then routing is performed to realize the connection of the selected pin-pairs. Heavy iterative steps are not used for both the selection and the routing, then a routing pattern is generated in a short time. In the experiments, we confirm that the quality of routing patterns generated by our method is almost equivalent to the existing algorithm. Furthermore, our method finds length matching routing patterns that the existing algorithm cannot find. The computation time is about 360 times faster than the existing algorithm.
Foisal AHMED Michihiro SHINTANI Michiko INOUE
Analyzing aging-induced delay degradations of ring oscillators (ROs) is an effective way to detect recycled field-programmable gate arrays (FPGAs). However, it requires a large number of RO measurements for all FPGAs before shipping, which increases the measurement costs. We propose a cost-efficient recycled FPGA detection method using a statistical performance characterization technique called virtual probe (VP) based on compressed sensing. The VP technique enables the accurate prediction of the spatial process variation of RO frequencies on a die by using a very small number of sample RO measurements. Using the predicted frequency variation as a supervisor, the machine-learning model classifies target FPGAs as either recycled or fresh. Through experiments conducted using 50 commercial FPGAs, we demonstrate that the proposed method achieves 90% cost reduction for RO measurements while preserving the detection accuracy. Furthermore, a one-class support vector machine algorithm was used to classify target FPGAs with around 94% detection accuracy.