Transformer (machine learning model)A transformer is a deep learning architecture that relies on the parallel multi-head attention mechanism. The modern transformer was proposed in the 2017 paper titled 'Attention Is All You Need' by Ashish Vaswani et al., Google Brain team. It is notable for requiring less training time than previous recurrent neural architectures, such as long short-term memory (LSTM), and its later variation has been prevalently adopted for training large language models on large (language) datasets, such as the Wikipedia corpus and Common Crawl, by virtue of the parallelized processing of input sequence.
NanolithographyNanolithography (NL) is a growing field of techniques within nanotechnology dealing with the engineering (patterning e.g. etching, depositing, writing, printing etc) of nanometer-scale structures on various materials. The modern term reflects on a design of structures built in range of 10−9 to 10−6 meters, i.e. nanometer scale. Essentially, the field is a derivative of lithography, only covering very small structures. All NL methods can be categorized into four groups: photo lithography, scanning lithography, soft lithography and other miscellaneous techniques.
Attention (machine learning)Machine learning-based attention is a mechanism mimicking cognitive attention. It calculates "soft" weights for each word, more precisely for its embedding, in the context window. It can do it either in parallel (such as in transformers) or sequentially (such as recursive neural networks). "Soft" weights can change during each runtime, in contrast to "hard" weights, which are (pre-)trained and fine-tuned and remain frozen afterwards. Multiple attention heads are used in transformer-based large language models.
Dip-pen nanolithographyDip pen nanolithography (DPN) is a scanning probe lithography technique where an atomic force microscope (AFM) tip is used to create patterns directly on a range of substances with a variety of inks. A common example of this technique is exemplified by the use of alkane thiolates to imprint onto a gold surface. This technique allows surface patterning on scales of under 100 nanometers.
Nanoimprint lithographyNanoimprint lithography (NIL) is a method of fabricating nanometer scale patterns. It is a simple nanolithography process with low cost, high throughput and high resolution. It creates patterns by mechanical deformation of imprint resist and subsequent processes. The imprint resist is typically a monomer or polymer formulation that is cured by heat or UV light during the imprinting. Adhesion between the resist and the template is controlled to allow proper release.
Electron-beam lithographyElectron-beam lithography (often abbreviated as e-beam lithography, EBL) is the practice of scanning a focused beam of electrons to draw custom shapes on a surface covered with an electron-sensitive film called a resist (exposing). The electron beam changes the solubility of the resist, enabling selective removal of either the exposed or non-exposed regions of the resist by immersing it in a solvent (developing). The purpose, as with photolithography, is to create very small structures in the resist that can subsequently be transferred to the substrate material, often by etching.
CodecA codec is a device or computer program that encodes or decodes a data stream or signal. Codec is a portmanteau of coder/decoder. In electronic communications, an endec is a device that acts as both an encoder and a decoder on a signal or data stream, and hence is a type of codec. Endec is a portmanteau of encoder/decoder. A coder or encoder encodes a data stream or a signal for transmission or storage, possibly in encrypted form, and the decoder function reverses the encoding for playback or editing.
PhotolithographyIn integrated circuit manufacturing, photolithography or optical lithography is a general term used for techniques that use light to produce minutely patterned thin films of suitable materials over a substrate, such as a silicon wafer, to protect selected areas of it during subsequent etching, deposition, or implantation operations. Typically, ultraviolet light is used to transfer a geometric design from an optical mask to a light-sensitive chemical (photoresist) coated on the substrate.
Multi-core processorA multi-core processor is a microprocessor on a single integrated circuit with two or more separate processing units, called cores, each of which reads and executes program instructions. The instructions are ordinary CPU instructions (such as add, move data, and branch) but the single processor can run instructions on separate cores at the same time, increasing overall speed for programs that support multithreading or other parallel computing techniques.
NanoelectronicsNanoelectronics refers to the use of nanotechnology in electronic components. The term covers a diverse set of devices and materials, with the common characteristic that they are so small that inter-atomic interactions and quantum mechanical properties need to be studied extensively. Some of these candidates include: hybrid molecular/semiconductor electronics, one-dimensional nanotubes/nanowires (e.g. silicon nanowires or carbon nanotubes) or advanced molecular electronics.
High-voltage direct currentA high-voltage direct current (HVDC) electric power transmission system (also called a power superhighway or an electrical superhighway) uses direct current (DC) for electric power transmission, in contrast with the more common alternating current (AC) transmission systems. Most HVDC links use voltages between 100 kV and 800 kV. However, a 1,100 kV link in China was completed in 2019 over a distance of with a power capacity of 12 GW. With this dimension, intercontinental connections become possible which could help to deal with the fluctuations of wind power and photovoltaics.
High voltageHigh voltage electricity refers to electrical potential large enough to cause injury or damage. In certain industries, high voltage refers to voltage above a certain threshold. Equipment and conductors that carry high voltage warrant special safety requirements and procedures. High voltage is used in electrical power distribution, in cathode ray tubes, to generate X-rays and particle beams, to produce electrical arcs, for ignition, in photomultiplier tubes, and in high-power amplifier vacuum tubes, as well as other industrial, military and scientific applications.
Memory protectionMemory protection is a way to control memory access rights on a computer, and is a part of most modern instruction set architectures and operating systems. The main purpose of memory protection is to prevent a process from accessing memory that has not been allocated to it. This prevents a bug or malware within a process from affecting other processes, or the operating system itself. Protection may encompass all accesses to a specified area of memory, write accesses, or attempts to execute the contents of the area.
Surge protectorA surge protector (or spike suppressor, surge suppressor, surge diverter, surge protection device (SPD) or transient voltage surge suppressor (TVSS) is an appliance or device intended to protect electrical devices from voltage spikes in alternating current (AC) circuits. A voltage spike is a transient event, typically lasting 1 to 30 microseconds, that may reach over 1,000 volts.
Shared memoryIn computer science, shared memory is memory that may be simultaneously accessed by multiple programs with an intent to provide communication among them or avoid redundant copies. Shared memory is an efficient means of passing data between programs. Depending on context, programs may run on a single processor or on multiple separate processors. Using memory for communication inside a single program, e.g. among its multiple threads, is also referred to as shared memory.
Uniform memory accessUniform memory access (UMA) is a shared memory architecture used in parallel computers. All the processors in the UMA model share the physical memory uniformly. In an UMA architecture, access time to a memory location is independent of which processor makes the request or which memory chip contains the transferred data. Uniform memory access computer architectures are often contrasted with non-uniform memory access (NUMA) architectures. In the NUMA architecture, each processor may use a private cache.
Memory managementMemory management is a form of resource management applied to computer memory. The essential requirement of memory management is to provide ways to dynamically allocate portions of memory to programs at their request, and free it for reuse when no longer needed. This is critical to any advanced computer system where more than a single process might be underway at any time. Several methods have been devised that increase the effectiveness of memory management.
