Related papers: Experimentally Realizing Convolution Processing in…
High-dimensional entanglement offers a variety of advantages for both fundamental and applied applications in quantum information science. A central building block for such applications is a programmable processor of entangled states, which…
Convolutional Neural Networks (CNNs) have achieved remarkable success in various computer vision tasks but rely on tremendous computational cost. To solve this problem, existing approaches either compress well-trained large-scale models or…
The convolutional neural network (CNN) is one of the most commonly used architectures for computer vision tasks. The key building block of a CNN is the convolutional kernel that aggregates information from the pixel neighborhood and shares…
The concept of synthetic dimensions, which has enabled the study of higher-dimensional physics on lower-dimensional physical structures, has generated significant recent interest in many branches of science ranging from ultracold-atomic…
Photonic convolutional accelerators have emerged as low-energy alternatives to power-demanding digital convolutional neural networks, though they often face limitations in scalability. In this work, we introduce a convolutional photonic…
Synthetic dimensions based on particles' internal degrees of freedom, such as frequency, spatial modes and arrival time, have attracted significant attention. They offer ideal large-scale lattices to simulate nontrivial topological…
We generalize the concept of photonic gauge potential in real space, by introducing an additional "synthetic" frequency dimension in addition to the real space dimensions. As an illustration we consider a one-dimensional array of ring…
Recent work in neural networks for image classification has seen a strong tendency towards increasing the spatial context. Whether achieved through large convolution kernels or self-attention, models scale poorly with the increased spatial…
Convolutional Neural Networks (CNN) have been the centerpiece of many applications including but not limited to computer vision, speech processing, and Natural Language Processing (NLP). However, the computationally expensive convolution…
In recent years there has been significant interest in the concepts of synthetic dimensions, where one couples the internal degrees of freedom of a particle to form higher-dimensional lattices in lower-dimensional physical structures. For…
We introduce the concept of compressed convolution, a technique to convolve a given data set with a large number of non-orthogonal kernels. In typical applications our technique drastically reduces the effective number of computations. The…
Arbitrary manipulation of light across multiple physical dimensions is essential for harnessing its parallelism in fundamental research and advanced applications, such as optical interconnects, computing, imaging, sensing, and quantum…
Optical frequency combs are utilized in a wide range of optical applications, including atomic clocks, interferometers, and various sensing technologies. They are often generated via four-wave mixing in chip-integrated microring resonators,…
An efficient simulator for quantum systems is one of the original goals for the efforts to develop a quantum computer [1]. In recent years, synthetic dimension in photonics [2] have emerged as a potentially powerful approach for simulation…
Conventional optical synthesis, the manipulation of the phase and amplitude of spectral components to produce an optical pulse in different temporal modes, is revolutionizing ultrafast optical science and metrology. These technologies rely…
Transposed convolution is crucial for generating high-resolution outputs, yet has received little attention compared to convolution layers. In this work we revisit transposed convolution and introduce a novel layer that allows us to place…
In this paper, we propose an end-to-end mixed-resolution image compression framework with convolutional neural networks. Firstly, given one input image, feature description neural network (FDNN) is used to generate a new representation of…
Dilated Convolutions have been shown to be highly useful for the task of image segmentation. By introducing gaps into convolutional filters, they enable the use of larger receptive fields without increasing the original kernel size. Even…
The frequency conversion of light has proved to be a crucial technology for communication, spectroscopy, imaging, and signal processing. In the quantum regime, it also offers great potential for realizing quantum networks incorporating…
The ability to manipulate the frequency of light is of great importance in both fundamental quantum sciences and practical applications. Traditional method for frequency conversion relies on nonlinear optical processes, which are faced with…