Related papers: 20-GHz bandwidth optical activation function based…
A wide variety of activation functions have been proposed for neural networks. The Rectified Linear Unit (ReLU) is especially popular today. There are many practical reasons that motivate the use of the ReLU. This paper provides new…
The reason behind CNNs capability to learn high-dimensional complex features from the images is the non-linearity introduced by the activation function. Several advanced activation functions have been discovered to improve the training…
Reliable, versatile galaxy activity diagnostics are essential for understanding galaxy evolution. Traditional methods frequently necessitate extensive preprocessing, such as starlight subtraction and emission line deblending (e.g.,…
Silicon photonics in conjunction with complementary metal-oxide-semiconductor (CMOS) fabrication has greatly enhanced the development of integrated optical phased arrays. This facilitates a dynamic control of light in a compact form factor…
We report a single-frequency, narrow-linewidth semiconductor pulsed laser based on pump current modulation and optical injection locking technique. A monolithic non-planar ring oscillator laser is employed as the seed source to guarantee…
Convolutional Neural Networks (CNNs) have been widely applied. But as the CNNs grow, the number of arithmetic operations and memory footprint also increase. Furthermore, typical non-linear activation functions do not allow associativity of…
Electro-optic modulation, the imprinting of a radio-frequency (RF) waveform on an optical carrier, is one of the most important photonics functions, being crucial for high-bandwidth signal generation, optical switching, waveform shaping,…
We demonstrate a free-space amplitude modulator for mid-infrared radiation (lambda=9.6 um) that operates at room temperature up to at least 20 GHz (above the -3dB cutoff frequency measured at 8.2 GHz). The device relies on the ultra-fast…
In contrast to radio frequency (RF), where the modulation bandwidth is restricted by regulations to avoid interference, the available bandwidth in optical wireless communication (OWC) is primarily constrained by system components. To…
Oscillatory dynamics have recently proven highly effective in machine learning (ML), particularly through State-Space-Models (SSM) that leverage structured linear recurrences for long-range temporal processing. Resonate-and-Fire neurons…
Activation functions play a key role in providing remarkable performance in deep neural networks, and the rectified linear unit (ReLU) is one of the most widely used activation functions. Various new activation functions and improvements on…
The rapid expansion of generative AI drives unprecedented demands for high-performance computing. Training large-scale AI models now requires vast interconnected GPU clusters across multiple data centers. Multi-scale AI training and…
Optomechanical systems combine extreme sensitivity and bandwidth in the control of mechanical motion, of interest for various applications. Integrated on a chip, actuated and detected all-optically by a single laser, they could disrupt…
Successive linear transforms followed by nonlinear "activation" functions can approximate nonlinear functions to arbitrary precision given sufficient layers. The number of necessary layers is dependent on, in part, by the nature of the…
This paper presents a theoretical and experimental investigation of optical signal regeneration properties of a non-linear optical loop mirror using a semiconductor optical amplifier as the active element (SOA-NOLM). While this device has…
Deep neural networks have been successfully used in diverse emerging domains to solve real world complex problems with may more deep learning(DL) architectures, being developed to date. To achieve these state-of-the-art performances, the DL…
Deep neural networks have achieved remarkable breakthroughs by leveraging multiple layers of data processing to extract hidden representations, albeit at the cost of large electronic computing power. To enhance energy efficiency and speed,…
In neural networks, non-linearity is introduced by activation functions. One commonly used activation function is Rectified Linear Unit (ReLU). ReLU has been a popular choice as an activation but has flaws. State-of-the-art functions like…
One of the fundamental limits of classical optical microscopy is the diffraction limit of optical resolution. It results from the finite bandwidth of the optical transfer function (or OTF) of an optical microscope, which restricts the…
Optical beamforming networks (OBFNs) based on optical true time delay lines (OTTDLs) are well-known as the promising candidate to solve the bandwidth limitation of traditional electronic phased array antennas (PAAs) due to beam squinting.…