AI Assistant Extension

AI Assistant Extension — independent reviews, comparisons, pricing and step-by-step guides on Aizhi.

• 

  Neural processing unit

  A neural processing unit (NPU), also known as an AI accelerator or deep learning processor, is a class of specialized hardware accelerator or computer system designed to accelerate artificial intelligence and machine learning applications, including artificial neural networks and computer vision. == Use == Their purpose is either to efficiently execute already trained AI models (inference) or to train AI models. NPUs can be more efficient in terms of speed or power consumption. NPU applications include algorithms for robotics, Internet of things, and data-intensive or sensor-driven tasks. They are often manycore or spatial designs and focus on low-precision arithmetic, novel dataflow architectures, or in-memory computing capability. As of 2024, a widely used datacenter-grade AI integrated circuit chip, the Nvidia H100 GPU, contains tens of billions of MOSFETs. === Consumer devices === AI accelerators are used in Apple silicon, Qualcomm, Samsung, Huawei, and Google Tensor smartphone processors. Vision processing units are accelerators specialized for machine vision algorithms such as CNN (convolutional neural networks) and SIFT (scale-invariant feature transform). They are used in devices that need to keep track of objects visually such as AR headsets and drones. It is more recently (circa 2017) added to processors from Apple and (circa 2022) to processors from Intel and AMD. All models of Intel Meteor Lake processors have a built-in versatile processor unit (VPU) for accelerating inference for computer vision and deep learning. On consumer devices, the NPU is intended to be small, power-efficient, but reasonably fast when used to run small models. To do this they are designed to support low-bitwidth operations using data types such as INT4, INT8, FP8, and FP16. A common metric is trillions of operations per second (TOPS). Although TOPS does not explicitly specify the kind of operations, it is typically INT8 additions and multiplications. === Datacenters === Accelerators are used in cloud computing servers: e.g., tensor processing units (TPU) for Google Cloud Platform, and Trainium and Inferentia chips for Amazon Web Services. Many vendor-specific terms exist for devices in this category, and it is an emerging technology without a dominant design. Since the late 2010s, graphics processing units designed by companies such as Nvidia and AMD often include AI-specific hardware in the form of dedicated functional units for low-precision matrix-multiplication operations. These GPUs are commonly used as AI accelerators, both for training and inference. === Scientific computation === Although NPUs are tailored for low-precision (e.g., FP16, INT8) matrix multiplication operations, they can be used to emulate higher-precision matrix multiplications in scientific computing. As modern GPUs place much focus on making the NPU part fast, using emulated FP64 (Ozaki scheme) on NPUs can potentially outperform native FP64. This has been demonstrated using FP16-emulated FP64 on NVIDIA TITAN RTX and using INT8-emulated FP64 on NVIDIA consumer GPUs and the A100 GPU. Consumer GPUs especially benefited as they have limited FP64 hardware capacity, showing a 6× speedup. Since CUDA Toolkit 13.0 Update 2, cuBLAS automatically uses INT8-emulated FP64 matrix multiplication of the equivalent precision if it is faster than native. This is in addition to the FP16-emulated FP32 feature introduced in version 12.9. == Programming == An operating system or a higher-level library may provide application programming interfaces such as TensorFlow with LiteRT Next (Android), CoreML (iOS, macOS) or DirectML (Windows). Formats such as ONNX are used to represent trained neural networks. Consumer CPU-integrated NPUs are accessible through vendor-specific APIs. AMD (Ryzen AI), Intel (OpenVINO), Apple silicon (CoreML), and Qualcomm (SNPE) each have their own APIs, which can be built upon by a higher-level library. GPUs generally use existing GPGPU pipelines such as CUDA and OpenCL adapted for lower precisions and specialized matrix-multiplication operations. Vulkan is also being used. Custom-built systems such as the Google TPU use private interfaces. There are a large number of separate underlying acceleration APIs and compilers/runtimes in use in the AI field, causing a great increase in software development effort due to the many combinations involved. As of 2025, the open standard organization Khronos Group is pursuing standardization of AI-related interfaces to reduce the amount of work needed. Khronos is working on three separate fronts: expansion of data types and intrinsic operations in OpenCL and Vulkan, inclusion of compute graphs in SPIR-V, and a NNEF/SkriptND file format for describing a neural network.

  Read more →
• 

  Vector quantization

  Vector quantization (VQ) is a classical quantization technique from signal processing that allows the modeling of probability density functions by the distribution of prototype vectors. Developed in the early 1980s by Robert M. Gray, it was originally used for data compression. It works by dividing a large set of points (vectors) into groups having approximately the same number of points closest to them. Each group is represented by its centroid point, as in k-means and some other clustering algorithms. In simpler terms, vector quantization chooses a set of points to represent a larger set of points. The density matching property of vector quantization is powerful, especially for identifying the density of large and high-dimensional data. Since data points are represented by the index of their closest centroid, commonly occurring data have low error, and rare data high error. This is why VQ is suitable for lossy data compression. It can also be used for lossy data correction and density estimation. Vector quantization is based on the competitive learning paradigm, so it is closely related to the self-organizing map model and to sparse coding models used in deep learning algorithms such as autoencoder. == Training == One simple training algorithm for vector quantization is: Pick a sample point at random Move the nearest quantization vector centroid towards this sample point, by a small fraction of the distance Repeat A more sophisticated algorithm reduces the bias in the density matching estimation and ensures that all points are used, by including an extra sensitivity parameter: Increase each centroid's sensitivity s i {\displaystyle s_{i}} by a small amount Pick a sample point P {\displaystyle P} at random For each quantization vector centroid c i {\displaystyle c_{i}} , let d ( P , c i ) {\displaystyle d(P,c_{i})} denote the distance of P {\displaystyle P} and c i {\displaystyle c_{i}} Find the centroid c i {\displaystyle c_{i}} for which d ( P , c i ) − s i {\displaystyle d(P,c_{i})-s_{i}} is the smallest Move c i {\displaystyle c_{i}} towards P {\displaystyle P} by a small fraction of the distance Set s i {\displaystyle s_{i}} to zero Repeat It is desirable to use a cooling schedule to produce convergence: see Simulated annealing. Another simple method is LBG, which is based on k-means. The algorithm can be iteratively updated with "live" data, rather than by picking random points from a data set, but this will introduce some bias if the data are temporally correlated over many samples. == Applications == Vector quantization is used for lossy data compression, lossy data correction, pattern recognition, density estimation and clustering. Lossy data correction, or prediction, is used to recover data missing from some dimensions. It is done by finding the nearest group with the data dimensions available, then predicting the result based on the values for the missing dimensions, assuming that they will have the same value as the group's centroid. For density estimation, the area/volume that is closer to a particular centroid than to any other is inversely proportional to the density (due to the density matching property of the algorithm). === Use in data compression === Vector quantization, also called "block quantization" or "pattern matching quantization" is often used in lossy data compression. It works by encoding values from a multidimensional vector space into a finite set of values from a discrete subspace of lower dimension. A lower-space vector requires less storage space, so the data is compressed. Due to the density matching property of vector quantization, the compressed data has errors that are inversely proportional to density. The transformation is usually done by projection or by using a codebook. In some cases, a codebook can be also used to entropy code the discrete value in the same step, by generating a prefix coded variable-length encoded value as its output. The set of discrete amplitude levels is quantized jointly rather than each sample being quantized separately. Consider a k-dimensional vector [ x 1 , x 2 , . . . , x k ] {\displaystyle [x_{1},x_{2},...,x_{k}]} of amplitude levels. It is compressed by choosing the nearest matching vector from a set of n-dimensional vectors [ y 1 , y 2 , . . . , y n ] {\displaystyle [y_{1},y_{2},...,y_{n}]} , with n < k. All possible combinations of the n-dimensional vector [ y 1 , y 2 , . . . , y n ] {\displaystyle [y_{1},y_{2},...,y_{n}]} form the vector space to which all the quantized vectors belong. Only the index of the codeword in the codebook is sent instead of the quantized values. This conserves space and achieves more compression. Twin vector quantization (VQF) is part of the MPEG-4 standard dealing with time domain weighted interleaved vector quantization. === Video codecs based on vector quantization === Bink video Cinepak Daala is transform-based but uses pyramid vector quantization on transformed coefficients Digital Video Interactive: Production-Level Video and Real-Time Video Indeo Microsoft Video 1 QuickTime: Apple Video (RPZA) and Graphics Codec (SMC) Sorenson SVQ1 and SVQ3 Smacker video VQA format, used in many games The usage of video codecs based on vector quantization has declined significantly in favor of those based on motion compensated prediction combined with transform coding, e.g. those defined in MPEG standards, as the low decoding complexity of vector quantization has become less relevant. === Audio codecs based on vector quantization === AMR-WB+ CELP CELT (now part of Opus) is transform-based but uses pyramid vector quantization on transformed coefficients Codec 2 DTS G.729 iLBC Ogg Vorbis TwinVQ === Use in pattern recognition === VQ was also used in the eighties for speech and speaker recognition. Recently it has also been used for efficient nearest neighbor search and on-line signature recognition. In pattern recognition applications, one codebook is constructed for each class (each class being a user in biometric applications) using acoustic vectors of this user. In the testing phase the quantization distortion of a testing signal is worked out with the whole set of codebooks obtained in the training phase. The codebook that provides the smallest vector quantization distortion indicates the identified user. The main advantage of VQ in pattern recognition is its low computational burden when compared with other techniques such as dynamic time warping (DTW) and hidden Markov model (HMM). The main drawback when compared to DTW and HMM is that it does not take into account the temporal evolution of the signals (speech, signature, etc.) because all the vectors are mixed up. In order to overcome this problem a multi-section codebook approach has been proposed. The multi-section approach consists of modelling the signal with several sections (for instance, one codebook for the initial part, another one for the center and a last codebook for the ending part). === Use as clustering algorithm === As VQ is seeking for centroids as density points of nearby lying samples, it can be also directly used as a prototype-based clustering method: each centroid is then associated with one prototype. By aiming to minimize the expected squared quantization error and introducing a decreasing learning gain fulfilling the Robbins-Monro conditions, multiple iterations over the whole data set with a concrete but fixed number of prototypes converges to the solution of k-means clustering algorithm in an incremental manner. === Generative adversarial networks (GAN) === VQ has been used to quantize a feature representation layer in the discriminator of generative adversarial networks. The feature quantization (FQ) technique performs implicit feature matching. It improves the GAN training, and yields an improved performance on a variety of popular GAN models: BigGAN for image generation, StyleGAN for face synthesis, and U-GAT-IT for unsupervised image-to-image translation.

  Read more →
• 

  AI Content Generators: Free vs Paid (2026)

  Shopping for the best AI content generator? An AI content generator is software that uses machine learning to help you get more done — it keeps getting smarter as the underlying models improve. Pricing, accuracy, and the size of the model behind the tool are the three factors that most affect daily usefulness. Whether you are a beginner or a pro, the right AI content generator slots into your workflow and pays for itself fast. We tested the leading options and ranked them by quality, value, and ease of use.

  Read more →
• 

  FrameNet

  FrameNet is a group of online lexical databases based upon the theory of meaning known as Frame semantics, developed by linguist Charles J. Fillmore. The project's fundamental notion is simple: most words' meanings may be best understood in terms of a semantic frame, which is a description of a certain kind of event, connection, or item and its actors. As an illustration, the act of cooking usually requires the following: a cook, the food being cooked, a container to hold the food while it is being cooked, and a heating instrument. Within FrameNet, this act is represented by a frame named Apply_heat, and its components (Cook, Food, Container, and Heating_instrument), are referred to as frame elements (FEs). The Apply_heat frame also lists a number of words that represent it, known as lexical units (LUs), like fry, bake, boil, and broil. Other frames are simpler. For example, Placing only has an agent or cause, a theme—something that is placed—and the location where it is placed. Some frames are more complex, like Revenge, which contains more FEs (offender, injury, injured party, avenger, and punishment). As in the examples of Apply_heat and Revenge below, FrameNet's role is to define the frames and annotate sentences to demonstrate how the FEs fit syntactically around the word that elicits the frame. == Concepts == === Frames === A frame is a schematic representation of a situation involving various participants, props, and other conceptual roles. Examples of frame names are Being_born and Locative_relation. A frame in FrameNet contains a textual description of what it represents (a frame definition), associated frame elements, lexical units, example sentences, and frame-to-frame relations. === Frame elements === Frame elements (FE) provide additional information to the semantic structure of a sentence. Each frame has a number of core and non-core FEs which can be thought of as semantic roles. Core FEs are essential to the meaning of the frame while non-core FEs are generally descriptive (such as time, place, manner, etc.) For example: The only core FE of the Being_born frame is called Child; non-core FEs Time, Place, Means, etc. Core FEs of the Commerce_goods-transfer frame include the Seller, Buyer, and Goods, while non-core FEs include a Place, Purpose, etc. FrameNet includes shallow data on syntactic roles that frame elements play in the example sentences. For example, for a sentence like "She was born about AD 460", FrameNet would mark She as a noun phrase referring to the Child frame element, and "about AD 460" as a noun phrase corresponding to the Time frame element. Details of how frame elements can be realized in a sentence are important because this reveals important information about the subcategorization frames as well as possible diathesis alternations (e.g. "John broke the window" vs. "The window broke") of a verb. === Lexical units === Lexical units (LUs) are lemmas, with their part of speech, that evoke a specific frame. In other words, when an LU is identified in a sentence, that specific LU can be associated with its specific frame(s). For each frame, there may be many LUs associated to that frame, and also there may be many frames that share a specific LU; this is typically the case with LUs that have multiple word senses. Alongside the frame, each lexical unit is associated with specific frame elements by means of the annotated example sentences. For example, lexical units that evoke the Complaining frame (or more specific perspectivized versions of it, to be precise), include the verbs complain, grouse, lament, and others. === Example sentences === Frames are associated with example sentences and frame elements are marked within the sentences. Thus, the sentence She was born about AD 460 is associated with the frame Being_born, while She is marked as the frame element Child and "about AD 460" is marked as Time. From the start, the FrameNet project has been committed to looking at evidence from actual language use as found in text collections like the British National Corpus. Based on such example sentences, automatic semantic role labeling tools are able to determine frames and mark frame elements in new sentences. === Valences === FrameNet also exposes statistics on the valence of each frame; that is, the number and position of the frame elements within example sentences. The sentence She was born about AD 460 falls in the valence pattern NP Ext, INI --, NP Dep which occurs twice in the FrameNet's annotation report for the born.v lexical unit, namely: She was born about AD 460, daughter and granddaughter of Roman and Byzantine emperors, whose family had been prominent in Roman politics for over 700 years. He was soon posted to north Africa, and never met their only child, a daughter born 8 June 1941. === Frame relations === FrameNet additionally captures relationships between different frames using relations. These include the following: Inheritance: When one frame is a more specific version of another, more abstract, parent frame. Anything that is true about the parent frame must also be true about the child frame, and a mapping is specified between the frame elements of the parent and the frame elements of the child. Perspectivization: A neutral frame is connected to a frame with a specific perspective of the same scenario. For example, Commerce_transfer-goods is considered from the perspective of the buyer in Commerce_buy and from that of the seller in Commerce_sell. Subframe: Some frames refer to complex scenarios that consist of several individual states or events that can be described by separate frames. For example, Criminal_process is composed of Arrest, Trial, and so on. Precedence: This relation captures the temporal order that holds between subframes of a complex frame. For example, within the Cycle_of_life_and_death frame, the subframe Death is preceded by the subframe Being_born. Causative and Inchoative: These two relations mark, for causative- and inchoative-aspect frames, the separate stative frame they refer to. For example, the stative Position_on_a_scale (e.g. "She had a high salary") is described by the causative Cause_change_of_scalar_position (e.g. "She raised his salary") and by the inchoative Change_position_on_a_scale frame (e.g. "Her salary increased"). Using: This relation marks a frame that in some way involves another frame. For example, Judgment_communication uses both Judgment and Statement, but does not inherit from either of them because there is no clear correspondence of frame elements. See also: Connects frames that bear some resemblance but need to be distinguished carefully. == Applications == FrameNet has proven to be useful in a number of computational applications, because computers need additional knowledge in order to recognize that "John sold a car to Mary" and "Mary bought a car from John" describe essentially the same situation, despite using two quite different verbs, different prepositions and a different word order. FrameNet has been used in applications like question answering, paraphrasing, recognizing textual entailment, and information extraction, either directly or by means of Semantic Role Labeling tools. The first automatic system for Semantic Role Labeling (SRL, sometimes also referred to as "shallow semantic parsing") was developed by Daniel Gildea and Daniel Jurafsky based on FrameNet in 2002. Semantic Role Labeling has since become one of the standard tasks in natural language processing, with the latest version (1.7) of FrameNet now fully supported in the Natural Language Toolkit. Since frames are essentially semantic descriptions, they are similar across languages, and several projects have arisen over the years that have relied on the original FrameNet as the basis for additional non-English FrameNets, for Spanish, Japanese, German, and Polish, among others.

  Read more →
• 

  ArcObjects

  ArcObjects is a development environment of the ArcGIS family of applications. Using Visual Basic for Applications, C# or Java SDK for ArcGIS, it allows developers to extend these applications.ArcObjects is a library of Component Object Model (COM) components that build up the foundation of Esri's ArcGIS platform. ArcObjects is written primarily in the C++ programming language. Since ArcGIS is completely built on top of ArcObjects, the ArcGIS platform can be fully customized and extended by making use of its COM services and capabilities. This allows for easy extension of the ArcObjects data model with any programming language that is compatible with COM, such as Visual Basic, C#, Visual Basic.NET, Java and Python. COM enables components to be reused at a binary level, meaning developers do not require access to the source code of ArcObjects in order to extend the ArcGIS platform. For this reason, an ArcObjects programmer can make use of any type inside the ArcObjects system without knowing the implementation details of the type, only needing to know what the type is able to do. The ArcObjects data model is based on the COM standard, which makes it compatible with other COM objects and applications. This allows for easy integration and collaboration with other systems that are also based on the COM standard. The ArcGIS platform was built using ArcObjects types, such as classes, interfaces, and enumerations. ArcObjects use COM interfaces to organize and communicate properties and methods of its classes, ensuring compatibility with other COM-based objects and systems. When working with an ArcObjects COM class, its properties and methods are accessed solely through one of its implemented interfaces via the process of Query Interface (QI). Multiple interfaces are commonly available for classes in ArcObjects. For example, it is possible to query for additional interfaces implemented by an object after instantiation via the process of QI. Although only one interface can be used when instantiating an object, multiple interfaces are often available for classes in ArcObjects, allowing for greater flexibility and compatibility with other systems based on the COM standard.

  Read more →
• 

  Selmer Bringsjord

  Selmer Bringsjord (born November 24, 1958) is a professor of computer science and cognitive science and a former chair of the Department of Cognitive Science at Rensselaer Polytechnic Institute. He also holds an appointment in the Lally School of Management & Technology and teaches artificial Intelligence (AI), formal logic, human and machine reasoning, and philosophy of AI. == Biography == Bringsjord's education includes a B.A. in philosophy from the University of Pennsylvania and a Ph.D. in philosophy from Brown University, where he studied under Roderick Chisholm. He conducts research in AI as the director of the Rensselaer AI & Reasoning (RAIR) Laboratory. He specializes in the logico-mathematical and philosophical foundations of AI and cognitive science, and in collaboratively building AI systems on the basis of computational logic. Bringsjord believes that "the human mind will forever be superior to AI", and that "much of what many humans do for a living will be better done by indefatigable machines who require not a cent in pay". Bringsjord has stated that the "ultimate growth industry will be building smarter and smarter such machines on the one hand, and philosophizing about whether they are truly conscious and free on the other". Bringsjord has an argument for P = NP using digital physics. Other research includes developing a new computational-logic framework allowing the formalization of deliberative multi-agent "mindreading" as applied to the realm of nuclear strategy, with the goal of creating a model and simulation to enable reliable prediction. He has published an opinion piece advocating for counter-terrorism security ensured by pervasive, all-seeing sensors; automated reasoners; and autonomous, lethal robots. Bringsjord received a National Science Foundation award to research Social Robotics and the Covey Award for the advancement of philosophy of computing awarded by the International Association for Computing And Philosophy, among several others prizes. == Books authored == with Yang, Y. Mental Metalogic: A New, Unifying Theory of Human and Machine Reasoning (Mahwah, NJ: Lawrence Erlbaum).(2007) with Zenzen, M. Superminds: People Harness Hypercomputation, and More (Dordrecht, The Netherlands: Kluwer). (2003) ISBN 978-1402010958 with Ferrucci, D. Artificial Intelligence and Literary Creativity: Inside the Mind of Brutus, A Storytelling Machine (Mahwah, NJ: Lawrence Erlbaum).(2000) Abortion: A Dialogue (Indianapolis, IN: Hackett).(1997) What Robots Can and Can’t Be (Dordrecht, The Netherlands: Kluwer).(1992) Soft Wars (New York, NY: Penguin USA). A novel.(1991)

  Read more →
• 

  Eric Brill

  Eric Brill is a computer scientist specializing in natural language processing. He created the Brill tagger, a supervised part of speech tagger. Another research paper of Brill introduced a machine learning technique now known as transformation-based learning. == Biography == Brill earned a BA in mathematics from the University of Chicago in 1987 and a MS in Computer Science from UT Austin in 1989. In 1994, he completed his PhD at the University of Pennsylvania. He was an assistant professor at Johns Hopkins University from 1994 to 1999. In 1999, he left JHU for Microsoft Research, he developed a system called "Ask MSR" that answered search engine queries written as questions in English, and was quoted in 2004 as predicting the shift of Google's web-page based search to information based search. In 2009 he moved to eBay to head their research laboratories.

  Read more →
• 

  Global Language Monitor

  The Global Language Monitor (GLM) is a company based in Austin, Texas, that analyzes trends in the English language. == History == Founded in Silicon Valley in 2003 by Paul J.J. Payack, the GLM describes its role as "a media analytics company that documents, analyzes and tracks cultural trends in language the world over, with a particular emphasis upon International and Global English". In April 2008, GLM moved its headquarters from San Diego to Austin. In July 2020, GLM announced that the word covid was its Top Word of 2020 for English. The company has been repeatedly criticized by linguists for promoting misinformation about language. Writing on Language Log, the linguist Ben Zimmer accused it of "hoodwink[ing] unsuspecting journalists on a range of pseudoscientific claims".

  Read more →
• 

  ArcObjects

  ArcObjects is a development environment of the ArcGIS family of applications. Using Visual Basic for Applications, C# or Java SDK for ArcGIS, it allows developers to extend these applications.ArcObjects is a library of Component Object Model (COM) components that build up the foundation of Esri's ArcGIS platform. ArcObjects is written primarily in the C++ programming language. Since ArcGIS is completely built on top of ArcObjects, the ArcGIS platform can be fully customized and extended by making use of its COM services and capabilities. This allows for easy extension of the ArcObjects data model with any programming language that is compatible with COM, such as Visual Basic, C#, Visual Basic.NET, Java and Python. COM enables components to be reused at a binary level, meaning developers do not require access to the source code of ArcObjects in order to extend the ArcGIS platform. For this reason, an ArcObjects programmer can make use of any type inside the ArcObjects system without knowing the implementation details of the type, only needing to know what the type is able to do. The ArcObjects data model is based on the COM standard, which makes it compatible with other COM objects and applications. This allows for easy integration and collaboration with other systems that are also based on the COM standard. The ArcGIS platform was built using ArcObjects types, such as classes, interfaces, and enumerations. ArcObjects use COM interfaces to organize and communicate properties and methods of its classes, ensuring compatibility with other COM-based objects and systems. When working with an ArcObjects COM class, its properties and methods are accessed solely through one of its implemented interfaces via the process of Query Interface (QI). Multiple interfaces are commonly available for classes in ArcObjects. For example, it is possible to query for additional interfaces implemented by an object after instantiation via the process of QI. Although only one interface can be used when instantiating an object, multiple interfaces are often available for classes in ArcObjects, allowing for greater flexibility and compatibility with other systems based on the COM standard.

  Read more →
• 

  Philipp Koehn

  Philipp Koehn (born 1 August 1971 in Erlangen, West Germany) is a computer scientist and researcher in the field of machine translation. His primary research interest is statistical machine translation and he is one of the inventors of a method called phrase based machine translation. This is a sub-field of statistical translation methods that employs sequences of words (or so-called "phrases") as the basis of translation, expanding the previous word based approaches. A 2003 paper which he authored with Franz Josef Och and Daniel Marcu called Statistical phrase-based translation has attracted wide attention in Machine translation community and has been cited over a thousand times. Phrase based methods are widely used in machine translation applications in industry. Philipp Koehn received his PhD in computer science in 2003 from the University of Southern California, where he worked at the Information Sciences Institute advised by Kevin Knight. After a year as a postdoctoral fellow under Michael Collins at the Massachusetts Institute of Technology, he joined the University of Edinburgh as a lecturer in the School of Informatics in 2005. He was appointed reader in 2010 and professor in 2012. In 2014, he was appointed professor at the computer science department of The Johns Hopkins University, where he is affiliated with the Center for Language and Speech Processing. == Moses statistical machine translation decoder == The Moses machine translation decoder is an open source project that was created by and is maintained under the guidance of Philipp Koehn. The Moses decoder is a platform for developing Statistical machine translation systems given a parallel corpus for any language pair. The decoder was mainly developed by Hieu Hoang and Philipp Koehn at the University of Edinburgh and extended during a Johns Hopkins University Summer Workshop and further developed under Euromatrix and GALE project funding. The decoder (which is part of a complete statistical machine translation toolkit) is the de facto benchmark for research in the field. Although Koehn continues to play a major role in the development of Moses, the Moses decoder was supported by the European Framework 6 projects Euromatrix, TC-Star, the European Framework 7 projects EuroMatrixPlus, Let's MT, META-NET and MosesCore and the DARPA GALE project, as well as several universities such as the University of Edinburgh, the University of Maryland, ITC-irst, Massachusetts Institute of Technology, and others. Substantial additional contributors to the Moses decoder include Hieu Hoang, Chris Dyer, Josh Schroeder, Marcello Federico, Richard Zens, and Wade Shen. == Europarl corpus == The Europarl corpus is a set of documents that consists of the proceedings of the European Parliament from 1996 to the present. The corpus has been compiled and expanded by a group of researchers led by Philipp Koehn at University of Edinburgh. The data that makes up the corpus was extracted from the website of the European Parliament and then prepared for linguistic research. The latest release (2012) comprised up to 60 million words per language, with 21 European languages represented: Romanic (French, Italian, Spanish, Portuguese, Romanian), Germanic (English, Dutch, German, Danish, Swedish), Slavic (Bulgarian, Czech, Polish, Slovak, Slovene), Finno-Ugric (Finnish, Hungarian, Estonian), Baltic (Latvian, Lithuanian), and Greek. == Other interests and activities in chronological order == Koehn is a professor at Johns Hopkins University where he continues his research into machine translation through his affiliation with the Center for Language and Speech Processing Koehn is a professor and chair of machine translation at the University of Edinburgh School of Informatics and contributes to its statistical machine translation group which organises workshops, seminars and project related to the subject. Koehn has consulted to SYSTRAN periodically between 2006 and 2011. SYSTRAN was acquired by CLSI, a Korean machine translation company in April 2014. Koehn worked for Facebook/META AI Research from 2018 to 2022. Koehn is also chief scientist for Omniscien Technologies and a shareholder in Omniscien Technologies since 2007. Omniscien Technologies is a private company developing and commercialising machine translation technologies. Koehn authored a book titled "Statistical Machine Translation" in 2009 and a book titled "Neural Machine Translation" in 2020. == Awards and recognition == 2013: One of three finalists in the category of Research for the European Patent Office (EPO) 2013 European Inventor Award. Koehn was recognised for patent EP 1488338 B, Phrase-Based Joint Probability Model for Statistical Machine Translations, a translation model that uses mathematical probabilities to determine the most likely interpretation of chunks of text between foreign languages. 2015: Koehn received the Award of Honor of the International Association for Machine Translation. 2024: Koehn was named Fellow of the Association for Computational Linguistics (ACL).

  Read more →
• 

  OCR Systems

  OCR Systems, Inc., was an American computer hardware manufacturer and software publisher dedicated to optical character recognition technologies. The company's first product, the System 1000 in 1970, was used by numerous large corporations for bill processing and mail sorting. Following a series of pitfalls in the 1970s and early 1980s, founder Theodor Herzl Levine put the company in the hands of Gregory Boleslavsky and Vadim Brikman, the company's vice presidents and recent immigrants from the Soviet Ukraine, who were able to turn OCR System's fortunes around and expand its employee base. The company released the software-based OCR application ReadRight for DOS, later ported to Windows, in the late 1980s. Adobe Inc. bought the company in 1992. == History == OCR Systems was co-founded by Theodor Herzl Levine (c. 1923 – May 30, 2005). Levine served in the U.S. Army Signal Corps during World War II in the Solomon Islands, where he helped develop a sonar to find ejected pilots in the ocean. After the war, Levine spent 22 years at the University of Pennsylvania, earning his bachelor's degree in 1951, his master's degree in electrical engineering in 1957, and his doctorate in 1968. Alongside his studies, Levine taught statistics and calculus at Temple University, Rutgers University, La Salle University and Penn State Abington. Sometime in the 1960s, Levine was hired at Philco. He and two of his co-workers decided to form their own company dedicated to optical character recognition, founding OCR Systems in 1969 in Bensalem, Pennsylvania. OCR Systems's first product, the System 1000, was announced in 1970. OCR Systems entered a partnership with 3M to resell the System 1000 throughout the United States in March 1973. This was 3M's entry into the data entry field, managed by the company's Microfilm Products Division and accompanying 3M's suite of data retrieval systems. It soon found use among Texas Instruments, AT&T, Ricoh, Panasonic and Canon for bill processing and mail sorting. Later in the mid-1970s an unspecified Fortune 500 company reneged on a contract to distribute the System 1000; later still a Canadian company distributing the System 1000 in Canada went defunct. Both incidents led OCR Systems to go nearly bankrupt, although it eventually recovered. By the early 1980s, however, the company was almost insolvent. In 1983 Levine had only $8,000 in his savings and became bedridden with an illness. He left the company in the hands of Gregory Boleslavsky and Vadim Brikman, two Soviet Ukraine expats whom Levine had hired earlier in the 1980s. Boleslavsky was hired as a wire wrapper for the System 1000 and as a programmer and beta tester for ReadRight—a software package developed by Levine implementing patents from Nonlinear Technology, another OCR-centric company from Greenbelt, Maryland. Boleslavsky in turn recommended Brikman to Levine. The two soon became vice presidents of the company while Levine was bedridden; in Boleslavsky's case, he worked 14-hour work days for over half a year in pursuit of the title. The two presented OCR Systems' products to the National Computer Conference in Chicago, where they were massively popular. The company soon gained such clients as Allegheny Energy in Pennsylvania and the postal service of Belgium and received an influx of employees—mostly expats from Russia but also Poland and South Korea, as well as American-born workers. To accommodate the company's employee base, which had grown to over 30 in 1988, Levine moved OCR System's headquarters from Bensalem to the Masons Mill Business Park in Bryn Athyn. Chinon Industries of Japan signed an agreement with OCR Systems in 1987 to distribute OCR's ReadRight 1.0 software with Chinon's scanners, starting with their N-205 overhead scanner. In 1988, OCR opened their agreement to distribute ReadRight to other scanner manufacturers, including Canon, Hewlett-Packard, Skyworld, Taxan, Diamond Flower and Abaton. That year, the company posted a revenue of $3 million. OCR Systems extended their agreement with Chinon in 1989 and introduced version 2.0 of ReadRight. OCR Systems faced stiff competition in the software OCR market in the turn of the 1990s. The Toronto-based software firm Delrina signed a letter of intent to purchase the company in November 1991, expecting the deal to close in December and have OCR software available by Christmas. OCR was to receive $3 million worth of Delrina shares in a stock swap, but the deal collapsed in January 1992. Delrine later marketed its own Extended Character Recognition, or XCR, software package to compete with ReadRight. In July 1992, OCR Systems was purchased by Adobe Inc. for an undisclosed sum. == Products == === System 1000 === The System 1000 was based on the 16-bit Varian Data 620/i minicomputer with 4 KB of core memory. The system used the 620/i for controlling the paper feed, interpreting the format of the documents, the optical character recognition process itself, error detection, sequencing and output. The System was initially programmed to recognize 1428 OCR (used by Selectrics); IBM 407 print; and the full character sets of OCR-A, OCR-B and Farrington 7B; as well as optical marks and handwritten numbers. OCR Systems promised added compatibility with more fonts available down the line—per request—in 1970. The number of fonts supported was limited by the amount of core memory, which was expandable in 4 KB increments up to 32 KB. The System 1000 later supported generalized typewriter and photocopier fonts. The rest of the System 1000 comprised the document transport, one or more scanner elements, a CRT display and a Teletype Model 33 or 35. Pages are fed via friction with a rubber belt. Up to three lines could be scanned per document, while the rest of the scanned document could be laid out in any manner granted there was enough space around the fields to be read. The reader initially supported pages as small as 3.25 in by 3.5 in dimension (later supporting 2.6 in by 3.5 in utility cash stubs) all the way to the standard ANSI letter size (8.5 in by 11 in; later 8.5 in by 12 in as used in stock certificates). The initial System 1000 had a maximum throughput of 420 documents per minute per transport (later 500 documents per minute), contingent on document size and content. A feature unique to the System 1000 over other optical character recognition systems of the time was its ability to alert the operator when a field was unreadable or otherwise invalid. This feature, called Document Referral, placed the document in front of the operator and displayed a blank field on the screen of the included CRT monitor for manual re-entry via keyboard. Once input, data could be output to 7- or 9-track tape, paper tape, punched cards and other mass storage media or to System/360 mainframes for further processing. The complete System 1000 could be purchased for US$69,000. Options for renting were $1,800 per month on a three-year lease or $1,600 per month for five years. Computerworld wrote that it was less than half the cost of its competitors while more capable and user-friendly. Competing systems included the Recognition Equipment Retina, the Scan-Optics IC/20 and the Scan-Data 250/350. === ReadRight === ReadRight processes individual letters topographically: it breaks down the scanned letter into parts—strokes, curves, angles, ascenders and descenders—and follows a tree structure of letters broken down into these parts to determine the corresponding character code. ReadRight was entirely software-based, requiring no expansion card to work. Version 2.01, the last version released for DOS, runs in real mode in under 640 KB of RAM. OCR Systems released the Windows-only version 3.0 in 1991 while offering version 2.01 alongside it. The company unveiled a sister product, ReadRight Personal, dedicated to handheld scanners and for Windows only in October 1991. This version adds real-time scanning—each word is updated to the screen while lines are being scanned. ReadRight proper was later made a Windows-only product with version 3.1 in 1992. The inclusion of ReadRight 2.0 with Canon's IX-12F flatbed scanner led PC Magazine to award it an Editor's Choice rating in 1989. Despite this, reviewer Robert Kendall found qualification with ReadRight's ability to parse proportional typefaces such as Helvetica and Times New Roman. Mitt Jones of the same publication found version 2.01 to have improved its ability to read such typefaces and praised its ease of use and low resource intensiveness. Jones disliked the inability to handle uneven page paragraph column widths and graphics, noting that the manual recommended the user block out graphics with a Post-it Note. Version 3.1 for Windows received mixed reviews. Mike Heck of InfoWorld wrote that its "low cost and rich collection of features are hard to ignore" but rated its speed and accuracy average. Barry Simon of PC Magazine called it economical but inaccurate, unable to correct errors it did

  Read more →
• 

  Linguistics Research Center at UT Austin

  The Linguistics Research Center (LRC) at the University of Texas is a center for computational linguistics research & development. It was directed by Prof. Winfred Lehmann until his death in 2007, and subsequently by Dr. Jonathan Slocum. Since its founding, virtually all projects at the LRC have involved processing natural language texts with the aid of computers. The principal activities of the Center at present focus on Indo-European languages and comprise historical study, lexicography, and web-based teaching; staff members engage in several independent but often complementary projects in these fields using a variety of software, almost all of it developed in-house. == History == The LRC was founded by Winfred Lehmann in 1961. In the early days, research efforts at the LRC concentrated on machine translation (MT) -- the translation of texts from one human language to another with the aid of computers, very developed nowadays in the field of language industry—funded by the USAF and other sponsors. The LRC concentrated on German English translation, though a copy of the Russian Master Dictionary was deposited at the LRC after the ALPAC report. After a general hiatus ca. 1975-78, new funding led to the development by Jonathan Slocum and others of a new system with the same name (the METAL MT system), but with new sets of tools for linguists and vastly greater success, resulting in the delivery a production prototype then later a full-fledged commercial MT system. MT R&D continued at the LRC, with funding by various sponsors, until well into the 1990s. From its early years to the present, the LRC has mounted a number of smaller projects resulting in the publication of significant works relating to Indo-European languages and/or their common ancestor, Proto-Indo-European. The hallmark of this work has been the use of computers to transcribe texts and prepare them for publication. A prominent example of the LRC using computers to prepare texts for print publication is the book by Winfred P. Lehmann, A Gothic Etymological Dictionary (Leiden: Brill, 1986). The final print-ready version was produced with the aid of a laser printer (exotic new technology, in those days) using, for the various languages included in the entries, approximately 500 special characters—many of them designed at the Center. This was the first major etymological dictionary for Indo-European languages to be produced with the aid of computers. Current LRC projects have concentrated on transcribing early Indo-European texts, developing language lessons based on them, and publishing on the web these and other materials related to the study of Indo-European languages, of their common ancestor Proto-Indo-European, and of historical linguistics more generally. == Alumni == Winfred Lehmann Rolf A. Stachowitz Jonathan Slocum Winfield S. Bennett John White

  Read more →
• 

  RCUDA

  rCUDA, which stands for Remote CUDA, is a type of middleware software framework for remote GPU virtualization. Fully compatible with the CUDA application programming interface (API), it allows the allocation of one or more CUDA-enabled GPUs to a single application. Each GPU can be part of a cluster or running inside of a virtual machine. The approach is aimed at improving performance in GPU clusters that are lacking full utilization. GPU virtualization reduces the number of GPUs needed in a cluster, and in turn, leads to a lower cost configuration – less energy, acquisition, and maintenance. The recommended distributed acceleration architecture is a high performance computing cluster with GPUs attached to only a few of the cluster nodes. When a node without a local GPU executes an application needing GPU resources, remote execution of the kernel is supported by data and code transfers between local system memory and remote GPU memory. rCUDA is designed to accommodate this client-server architecture. On one end, clients employ a library of wrappers to the high-level CUDA Runtime API, and on the other end, there is a network listening service that receives requests on a TCP port. Several nodes running different GPU-accelerated applications can concurrently make use of the whole set of accelerators installed in the cluster. The client forwards the request to one of the servers, which accesses the GPU installed in that computer and executes the request in it. Time-multiplexing the GPU, or in other words sharing it, is accomplished by spawning different server processes for each remote GPU execution request. == rCUDA v20.07 == The rCUDA middleware enables the concurrent usage of CUDA-compatible devices remotely. rCUDA employs either the InfiniBand network or the socket API for the communication between clients and servers. rCUDA can be useful in three different environments: Clusters. To reduce the number of GPUs installed in High Performance Clusters. This leads to energy savings, as well as other related savings like acquisition costs, maintenance, space, cooling, etc. Academia. In commodity networks, to offer access to a few high performance GPUs concurrently to many students. Virtual Machines. To enable the access to the CUDA facilities on the physical machine. The current version of rCUDA (v20.07) supports CUDA version 9.0, excluding graphics interoperability. rCUDA v20.07 targets the Linux OS (for 64-bit architectures) on both client and server sides. CUDA applications do not need any change in their source code in order to be executed with rCUDA.

  Read more →
• 

  Luca Maria Gambardella

  Luca Maria Gambardella (born 4 January 1962) is an Italian computer scientist and author. He is the former director of the Dalle Molle Institute for Artificial Intelligence Research in Lugano, in the Ticino canton of Switzerland. He is currently the prorector of Università della Svizzera italiana, where he directs the Master of Science in Artificial Intelligence degree course. Several of his papers have been extensively cited, with his collaborators including Marco Dorigo, with whom he has published papers on the application of ant colony optimization theory to the traveling salesman problem, and Jürgen Schmidhuber with whom he has published research on deep neural networks.. Beside working in research, Gambardella explores the potentials of AI applied for the generation of art. Some of his artistic installations received significant media coverage. As a novelist, the genres he approached broad from Bildungsroman of his first book "Sei vite" ("Six lives"), to romance of his second book "Il suono dell'alba" ("The sound of sunrise").

  Read more →
• 

  Forrest N. Iandola

  Forrest N. Iandola is an American computer scientist specializing in efficient AI. == Career == Iandola earned a PhD in Electrical Engineering and Computer Science from UC Berkeley in 2016, advised by Kurt Keutzer. As part of his dissertation, he co-authored SqueezeNet, a deep neural network for image classification optimized for smartphones and other mobile devices. Iandola and Keutzer went on to co-found DeepScale. The firm squeezes deep neural networks onto low-cost automotive-grade processors for use in driver assistance systems. Tesla acquired DeepScale in 2019. In 2020, he co-authored SqueezeBERT, an efficient neural network for natural language processing. In 2022, he joined Meta as an AI research scientist. His research at Meta includes developing efficient AI models, such as EfficientSAM and MobileLLM.

  Read more →