Silicon-based polarization-encoding quantum key distribution (QKD) has been widely studied, owing to its low cost and robustness. However, prior studies have utilized off-chip devices demodulate the states or perform polarization compensation, given difficulty of fabricating polarized independent components on chip. In this paper we propose a fully chip-based decoder for QKD. The chip realizes state analyzer compensates BB84 protocol without requiring additional hardware. It is based...

Simile is a special type of metaphor, where comparators such as like and are used to compare two objects. recognition recognize simile sentences extract components, i.e., the tenor vehicle. This paper presents study in Chinese. We construct an annotated corpus for this research, which consists 11.3k that contain comparator. propose neural network framework jointly optimizing three tasks: sentence classification, component extraction language modeling. The experimental results show based...

As the computation bottleneck in lattice-based cryptography (LBC), polynomial multiplication based on number theoretic transform (NTT) has been continuously studied for flexible hardware implementations with high area-efficiency. This paper presents an area-efficient and configurable NTT-based multiplier (AC-PM) incorporating algorithmic architectural level optimization techniques. For core operation of multiplication, two low-complexity fast modular algorithms are introduced loose...

Recently, the National Institute of Standards and Technology (NIST) has identified first four quantum-resistant algorithms for post-quantum cryptography (PQC) standardization. CRYSTALS-Kyber (Kyber) is only public-key encryption key-establishment algorithm among them. In this article, we propose a reconfigurable, high-speed, area-efficient polynomial multiplication accelerator Kyber to facilitate its practical applications. The cornerstone butterfly unit (BU) structure, composed modular...

Number theoretic transform (NTT) is widely applied as a fundamental component of next-generation cryptosystems. This brief introduces novel high-low interactive memory access pattern for an out-of-place NTT design, which can be configurable in degree parallelism. To achieve high area efficiency, the proposed flexibly selected referring to design parameters. Then first time, we present quantitative analysis about relevance between parallelism and number computing cycles. More importantly, put...

Since strong convective weather is closely related to heavy precipitation, the nowcasting of weather, especially based on radar data, plays an essential role in meteorological operations for disaster prevention and mitigation. The traditional optical flow method cross-correlation have a low forecast accuracy short leading time, while deep learning methods show remarkable advantages nowcasting. However, most current forecasting suffer from drawback that results become increasingly blurred as...

Named entity recognition (NER) is particularly challenging for medical texts due to the high domain specificity, abundance of technical terms, and sparsity data in this field. In work, we propose a novel attention layer, called “ontology layer”, that enhances NER performance language model clinical text by utilizing an ontology consisting conceptual classes related target set. The proposed layer computes relevance between each input token then fuses encoded vectors class enhance explicit...

Abstract We investigated the optical feedback effects on static and dynamic characteristics of 1.3 μ m quantum-dot (QD) Fabry–Pérot laser under reflection from −40 dB up to −8 dB. The onset coherence collapse is determined as −14 electrical spectra. Although degradation in small signal modulation reported above this critical level, transmission operation with available eye diagram higher demonstrated. Under 10 Gb s −1 modulation, there no obvious regarding shape extinction ratio tolerance QD...

Due to the rapid progress made in quantum computers, modern cryptography faces great challenges. Many digital signature schemes that have resistance computing are studied and standardized by several influential international organizations. The Leighton-Micali (LMS) protocol, one of hash-based schemes, is both Internet Engineering Task Force (IETF) National Institute Standards Technology (NIST) due its well-studied security relatively small size. However, heavy computation load high latency...

Due to the rapid progress made in quantum computers, modern cryptography faces great challenges. Many new digital signature schemes that have resistance computing are being presented for Post-Quantum Cryptography (PQC) standardization. The Leighton-Micali (LMS), a kind of hash-based scheme, is selected as promising candidate PQC protocols by Internet Engineering Task Force (IETF) because its small private and public key sizes. However, low-efficiency generation forms bottleneck practical...

Recently, secure neural network (NN) inference, a combination of homomorphic encryption (HE) and NN, has attracted much attention. Nevertheless, large number computations, mainly brought by the HE scheme, form bottleneck in real-time applications. In this article, we present hardware accelerator on field-programmable gate array (FPGA) for convolution layer (HomConvL), which is most computation-intensive part HE-based inference. First, propose new HomConvL algorithm called packed rotations at...

Grammatical error diagnosis is an important task in natural language processing. This paper introduces our system at NLPTEA-2020 Task: Chinese Error Diagnosis (CGED). CGED aims to diagnose four types of grammatical errors which are missing words (M), redundant (R), bad word selection (S) and disordered (W). Our built on the model multi-layer bidirectional transformer encoder ResNet integrated into improve performance. We also explore two ensemble strategies including weighted averaging...

The lattice-based cryptography (LBC) has been widely used recently in many compute-intensive applications, such as the post-quantum (PQC) and privacy-preserving deep learning, where main task for applications is to improve computational efficiency. modular multiplication operations, mainly involved number theoretic transform (NTT), comprise a large proportion of whole computations required by an LBC. This paper presents novel "decompose-and-reduce" algorithm (DARM), considering primes with...

Recently, the secure neural network inference, an organic combination of homomorphic encryption (HE) and deep (DNN), has attracted much attention. Nevertheless, large number computations, brought by HE scheme, form bottleneck for real-time applications. A significant portion is permutation (Perm), which mainly made up theoretic transform (NTT). In this paper, first time, we propose efficient architecture Perm incorporating algorithmic transformations architectural level optimizations. First,...

The homomorphic encryption over the torus (TFHE) is a promising fully (FHE) scheme that allows arbitrary computations with programmable bootstrapping (PBS) algorithm. However, PBS suffers from prohibitive computation complexity and latency, which hinders practical applications of TFHE. To address these challenges, we propose ALT, field-programmable gate array (FPGA) accelerator for exhibits high area efficiency low latency. Our approach involves modifying parameters algorithm to strike...

Supersingular isogeny key encapsulation (SIKE) protocol is a promising candidate for the standard of post quantum cryptography (PQC), but it suffers from high computational complexity. Since modular multiplication takes up large proportion computations in SIKE protocol, accelerating this operation can efficiently speed entire protocol. In paper, we propose new algorithm, which achieve lower complexity than prior arts. The SIKE-friendly prime with form p = 2 <sup...

With the deepening of research on language modeling, automatic metaphor identification task has become a new challenge in field natural processing, especially machine translation. This paper is about simile, one kind most frequent metaphorical way daily life. More specifically, recognition sentences with obvious components. We divided this into two sub-tasks, first classifying containing word "like", and then labelling tagged components labels. In paper, we used Random Forest classifier...

E-bussiness has grown rapidly in the last decade and massive amount of data on customer purchases, browsing pattern preferences been generated. Classification electronic plays a pivotal role to mine valuable information thus become one most important applications E-bussiness. Support Vector Machines are popular powerful machine learning techniques, they offer state-of-the-art performance. Rough set theory is formal mathematical tool deal with incomplete or imprecise its feature selection. In...

A novel median filter based manifold ranking approach for robust image annotation was presented in this paper. Firstly, we propose a similarity matching of to improve the robustness traditional label propagation with filter. Then refined affinity matrix applied propagation. Thus, pair-wise is updated scores using its neighbors space. Finally, proposed propagated labels from labeled images whole database. The experiments over Corel have shown that embedding space beneficial annotation.

Graph based learning has been an active research topic in machine community as well many application areas including image annotation recently. In order to exploit the correlation between keywords and images, we proposed a novel method via graph semantic fusion estimate probability of being caption image, present new framework solve problem. The experiments over Corel images have shown that this approach outperforms other methods is effective for annotation.

We propose a fully chip-based decoder for polarization-encoding quantum key distribution. The chip realizes polarization state analyzer and compensates the BB84 protocol without requiring additional hardware.

As a typical machine learning algorithm, convolutional neural network (CNN) has drawn great interests in academic research and industrial applications. However, traditional CPU can no longer meet the computation requirement of CNN due to CPU's sequential computing nature. Heterogeneous combines together with GPGPU or FPGAs form much more powerful platform. In this paper, we present our study on an implementation heterogeneous systems, it shows than 3x runtime speedup. Our systematically...