AI Art Can't Be Copyrighted

AI Art Can't Be Copyrighted — independent reviews, comparisons, pricing and step-by-step guides on Aizhi.

• 

  Embedding (machine learning)

  In machine learning, embedding is a representation learning technique that maps complex, high-dimensional data into a lower-dimensional vector space of numerical vectors. == Technique == It also denotes the resulting representation, where meaningful patterns or relationships are preserved. As a technique, it learns these vectors from data like words, images, or user interactions, differing from manually designed methods such as one-hot encoding. This process reduces complexity and captures key features without needing prior knowledge of the domain. == Similarity == In natural language processing, words or concepts may be represented as feature vectors, where similar concepts are mapped to nearby vectors. The resulting embeddings vary by type, including word embeddings for text (e.g., Word2Vec), image embeddings for visual data, and knowledge graph embeddings for knowledge graphs, each tailored to tasks like NLP, computer vision, or recommendation systems. This dual role enhances model efficiency and accuracy by automating feature extraction and revealing latent similarities across diverse applications. To measure the distance between two embeddings, a similarity measure can be used to find the overall similarity of the concepts represented by the embeddings. If the vectors are normalized to have a magnitude of 1, then the similarity measures are proportional to cos ⁡ ( θ a b ) {\displaystyle \cos \left(\theta _{ab}\right)} . The cosine similarity disregards the magnitude of the vector when determining similarity, so it is less biased towards training data that appears very frequently. The dot product includes the magnitude inherently, so it will tend to value more popular data. Generally, for high-dimensional vector spaces, vectors tend to converge in distance, so Euclidean distance becomes less reliable for large embedding vectors.

  Read more →
• 

  Best AI Writing Assistants in 2026

  In search of the best AI writing assistant? An AI writing assistant is software that uses machine learning to help you get more done — it turns a rough idea into a polished result in seconds. When choosing one, weigh output quality, pricing, export formats, and how well it fits the tools you already use. Whether you are a beginner or a pro, the right AI writing assistant slots into your workflow and pays for itself fast. Below we compare features, pricing, and real output so you can choose with confidence.

  Read more →
• 

  Best AI Code Generators in 2026

  Comparing the best AI code generator? An AI code generator is software that uses machine learning to help you get more done — it lowers the barrier so anyone can produce professional output. Privacy matters too: check whether your data trains the model and whether a no-log or enterprise tier is available. Whether you are a beginner or a pro, the right AI code 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 →
• 

  Hideto Tomabechi

  Hideto Tomabechi (苫米地 英人, Tomabechi Hideto; born 1959) is a Japanese cognitive scientist who is an adjunct fellow at Carnegie Mellon University and has had an executive role in several companies. == Early life and education == He grew up in Minato-ku, Tokyo. He graduated from Komaba Toho High School and then joined the University of Massachusetts Amherst. He received his first degree from Sophia University, then joined Mitsubishi Real Estate. Tomabechi was a Fulbright Scholar at Yale University and became member of Yale University Artificial Intelligence Research Center and Yale Cognitive Science Program. Hideto Tomabechi's research topic was: Cognition Models for Language Expressions and Computational Methods (Tomabechi Algorithm). Hideto Tomabechi received his Ph.D. in the field of computational linguistics from Carnegie Mellon University. His 1993 Ph.D. Thesis was entitled "Efficient Unification for Natural Language". == Career timeline == 1992-1998: Director, Justsystem Scientific Institute. 1998: CEO of Cognitive Research Laboratories Inc. 2007: Adjunct Fellow at the Cyber Security & Privacy Research Institute (CyLab) at Carnegie Mellon University. 2020: Visiting professor at Nano & Life Research Center, Waseda University. 2020: Chairman, Resilience Japan, LLC. 2022: Chairman of Japan Society for Foreign Policy. == Brain research == In 1993, Hideto Tomabechi became director of the Development Department. Later, Tomabechi became director of the JustSystems Basic Research Institute Tomabechi researched the basic functions of the human brain and mind. The purpose of brain and consciousness research were to develop the human machine interface. The main areas of research were altered states of consciousness, hypnosis, homeostasis, brain functions, and functions of the human mind in cyberspace. Dr. Tomabechi founded the Bechi Unit, the world's first virtual currency at JustSystems, based on Tomabech Algorithms. == Brainwashing == Tomabechi was the scientist who deprogrammed the leaders of the religious cult responsible for the terrorist attack in the Tokyo subway. The cult (Aum Shinrikyo) brainwashed its people and they carried out the attacks in an influenced state of consciousness.

  Read more →
• 

  World Database of Happiness

  The World Database of Happiness is a web-based archive of research findings on subjective appreciation of life, based in the Erasmus Happiness Economics Research Organization of the Erasmus University Rotterdam in The Netherlands. The database contains both an overview of scientific publications on happiness and a digest of research findings. Happiness is defined as the degree to which an individual judges the quality of his or her life as a whole favorably. Two 'components' of happiness are distinguished: hedonic level of affect (the degree to which pleasant affect dominates) and contentment (perceived realization of wants). == Aims == The World Database of Happiness is a tool to quickly acquire an overview on the ever-growing stream of research findings on happiness Medio 2023 the database covered some 16,000 scientific publications on happiness, from which were extracted 23,000 distributional findings (on how happy people are) and another 24,000 correlational findings (on factors associated with more and less happiness). The first findings date from 1915. == Technique == The World Database of Happiness is a ‘findings archive’, which consists of electronic ‘finding pages’ on which separate research results are described in a standard format and terminology. These finding pages can be selected on various characteristics, such as population studies, the measure of happiness used and observed co-variates. All finding-pages have a specific internet address to which links can be made in scientific review papers or policy recommendations. This allows a concise presentation of many findings in a table, while providing readers with access to detail. == Scientific use == The Database has been cited in 254 scientific papers, for example to access under what conditions economic growth enhances average happiness or to show that rising mean happiness at first raises happiness inequality, but further rise will diminish these differences, or that healthy eating is associated with more happiness, even after controlling for the effect on health Another finding is that relative simple happiness training techniques raise happiness by some 5% == Popular use == The World Database of Happiness is often used by popular media to make lists of the happiest countries around the globe. An example is the Happy Planet Index, which aims to chart sustainable happiness all over the world by combining data on longevity, happiness and the size of the ecological footprint of citizens. == Strengths and weaknesses == The database has a clear conceptual focus, it includes only research findings on subjective enjoyment of one's life as a whole. Thereby it evades the Babel that has haunted the study of happiness for ages. The other side of that coin is that much interesting research is left out. The findings are reported with technical details about measurement and statistical analysis. This detail is welcomed by scholars, but makes the information difficult to digest for lay-persons. Still another limitation is that the determinants of happiness appear to vary considerably across persons and situations, which make it hard to draw general conclusions about the causes of happiness. What is clear is that poor health, separation, unemployment and lack of social contact are all strongly negatively associated with happiness. Another problem for the World database of happiness is that the studies on happiness increase with such a high rate that it gets increasingly difficult to offer a complete overview of all research findings. A further concern is that the Database of Happiness is exclusively focused on hedonic happiness (feeling good) and not on mature happiness that might exist in the face of suffering

  Read more →
• 

  Pushpak Bhattacharyya

  Pushpak Bhattacharyya (3 July 1962 – 5 October 2025) was an Indian computer scientist and professor in the Department of Computer Science and Engineering at the IIT Bombay. He served as the Director of the IIT Patna from 2015 to 2021. He was a past President of the Association for Computational Linguistics (2016–17), and held the Vijay and Sita Vashee Chair Professorship at IIT Bombay. Bhattacharyya led the Natural Language Processing (NLP) research group at the Centre for Indian Language Technology (CFILT) at IIT Bombay until his death. At the inauguration of the Nilekani Centre at AI4Bharat, IIT Madras, Nandan Nilekani, Co-founder and Non-Executive Chairman of Infosys, referred to Bhattacharyya as the "Godfather of Indian NLP". == Early life and education == Bhattacharyya was born in Shillong in 1962. He completed his schooling at Jail Road Boys' High School, Shillong. He obtained a B.Tech. in Computer Science from the IIT Kharagpur, followed by an M.Tech. from the IIT Kanpur, and a Ph.D. in Computer Science from IIT Bombay in 1994. == Research == Bhattacharyya’s research areas includes Natural language processing, Artificial intelligence, Machine learning, Psycholinguistics, Eye tracking, and Information retrieval. He made contributions to the development of multilingual lexical databases such as IndoWordNet and other projects related to machine translation and computational linguistics. He authored and co-authored multiple academic works, including Investigations in Computational Sarcasm (with Aditya Joshi), Cognitively Inspired Natural Language Processing: An Investigation Based on Eye Tracking (with Abhijit Mishra), and Machine Translation and Transliteration of Low Resource Related Languages (with Anoop Kunchukuttan). Over his career, Bhattacharyya published more than 350 research papers in journals and conference proceedings and supervised over 300 undergraduate, master’s, and doctoral students. His projects often addressed computational challenges for Indian languages, such as developing wordnets, building translation systems for low-resource languages, and studying cognitive aspects of language processing. He also led government- and industry-funded research initiatives supported by organizations including IBM, Microsoft, Yahoo, and the United Nations. == Death == Bhattacharyya died on 5 October 2025, at the age of 63. == Awards == Patwardhan Award, IIT Bombay, for Technology Development VNMM Award, IIT Roorkee, for Technology Development Fellow, Indian National Academy of Engineering Eminent Engineer Award, Institution of Engineers (India)

  Read more →
• 

  Lori Levin

  Lorraine Susan (Lori) Levin is an American computer scientist and computational linguist specializing in natural language processing, particularly involving syntax, morphosyntax, and languages with small corpora. She is a research professor in the Language Technologies Institute of the Carnegie Mellon University School of Computer Science, and one of the founders of the North American Computational Linguistics Open Competition. == Education and career == Levin has a 1979 bachelor's degree in linguistics (summa cum laude) from the University of Pennsylvania, and a 1986 Ph.D. in linguistics from the Massachusetts Institute of Technology. Her dissertation, Operations on Lexical Forms: Unaccusative Rules in Germanic Languages, was jointly supervised by Joan Bresnan and Kenneth L. Hale. She worked as an assistant professor of linguistics at the University of Pittsburgh from 1983 until 1988, when she joined the Carnegie Mellon University Language Technologies Institute. == Recognition == Levin was named as a Fellow of the Association for Computational Linguistics in 2025, "for pioneering work on the use of phonetics, syntax, lexical semantics and dialogue modeling in machine translation and in the transfer of NLP technologies to low resource languages, as well as an enduring contribution to the North American Computational Linguistics Olympiad". Levin was awarded the Antonio Zampolli prize of the ELRA Language Resources Association at the LREC 2026 conference.

  Read more →
• 

  Best AI Presentation Makers in 2026

  In search of the best AI presentation maker? An AI presentation maker is software that uses machine learning to help you get more done — it turns a rough idea into a polished result in seconds. When choosing one, weigh output quality, pricing, export formats, and how well it fits the tools you already use. Whether you are a beginner or a pro, the right AI presentation maker slots into your workflow and pays for itself fast. Below we compare features, pricing, and real output so you can choose with confidence.

  Read more →
• 

  Apertus (LLM)

  Apertus is a public large language model, developed by the Swiss AI Initiative (a collaboration between EPFL, ETH Zurich, and the Swiss National Supercomputing Centre). It was released on September 2, 2025, under the free and open-source Apache 2.0 license. Designed initially for business and research use cases around the world, Apertus was trained on over 1800 languages, and comes in 8 billion or 70 billion parameter versions and is available on Hugging Face for download. The model was developed aiming to adhere to European copyright law, and is one of the first examples of AI as a public good in the vein of AI Sovereignty. It is also the first large model to comply with the European Union's Artificial Intelligence Act. At its launch, the model creators emphasized multilinguality, transparency, and auditability as priorities in contrast to commercial frontier model. While international reception was largely positive, the first iteration was significantly behind the capabilities of frontier models and needs adaptation for many use cases with chatbots being a secondary but not a primary use case. As of late 2025, it was considered the largest and most capable fully open model. The capability of future models will depend in part on how much more funding can be secured.

  Read more →
• 

  Julie Beth Lovins

  Julie Beth Lovins (October 19, 1945, in Washington, D.C. – January 26, 2018, in Mountain View, California) was a computational linguist who published The Lovins Stemming Algorithm - a type of stemming algorithm for word matching - in 1968. The Lovins Stemmer is a single pass, context sensitive stemmer, which removes endings based on the longest-match principle. The stemmer was the first to be published and was extremely well developed considering the date of its release, having been the main influence on a large amount of the future work in the area. -Adam G., et al == Background == Born on October 19, 1945, in Washington, D.C., Lovins grew up in Amherst, Massachusetts. Her father Gerald H. Lovins was an engineer and her mother, Miriam Lovins, a social services administrator. Lovins' brother Amory Lovins is the co-founder and chief environmental scientist of Rocky Mountain Institute. For her undergraduate degree, Lovins attended Pembroke College, the women's college of Brown University, which later combined into Brown University in 1971. At Pembroke College, Lovins studied mathematics and linguistics, graduating with honors. Her thesis was named, A Study of Idioms. She received the inaugural Bloch Fellowship in 1970 from the Linguistic Society of America to attend graduate school. Lovins obtained her Master of Arts in 1970 and Doctor of Philosophy in 1973 from the University of Chicago, studying linguistics. At the University of Chicago, her dissertation was titled, Loan Phonology -- Subject Matter. A revision of her thesis on loanwords and the phonological structure of Japanese was published in 1975 by the Indiana University Linguistics Club. == Teaching career == Following Lovins' PhD, she spent a year working as a linguist-at-large at a University of Tokyo language research institute and as an English conversation teacher. She then joined the faculty at Tsuda College as a professor of English and linguistics, where she taught for seven years. During her time as a faculty member at Tsuda College, Lovins also served as a guest researcher in the University of Tokyo's Research Institute of Logopedics and Phoniatrics, a research center for speech science. == Industry career == After teaching Japanese phonology at Japanese universities abroad, Lovins moved back to the U.S. to work in the computing industry. She worked on early speech synthesis at Bell Labs in Murray Hill, New Jersey. At Bell Labs, Lovins worked with Osamu Fujimura, a Japanese linguist who is credited as a pioneer in speech sciences. Lovins also worked as a software engineer at various companies in Silicon Valley and served as a consultant for computational linguistics throughout the 1990s. As a consultant, she called her business, "The Language Doctor." == The Lovins Stemming Algorithm == Lovins published an article about her work on developing a stemming algorithm through the Research Laboratory of Electronics at MIT in 1968. Lovins' stemming algorithm is frequently referred to as the Lovins stemmer. A stemming algorithm is the process of taking a word with suffixes and reducing it to its root, or base word. Stemming algorithms are used to improve the accuracy in information retrieval and in domain analysis. These algorithms help find variants of the terms being queried. Stemming algorithms bring value in their reduction of a given query into its less complex form, allowing more similar documents to be retrieved for similar queries. Stemming algorithms are prevalent in search engines, such as Google Search, which did not implement word stemming until 2003. This means that up until 2003, a Google search for the word warm would not have explicitly returned results for related words like warmth or warming. As the first published stemming algorithm, Lovins' work set a precedent and influenced future work in stemming algorithms, such as the Porter Stemmer published by Martin Porter in 1980 which has been recognized widely as the most common stemming algorithm for stemming English. Additionally, the Dawson Stemmer developed by John Dawson is an extension of the Lovins stemmer. The Lovins stemmer follows a rule-based affix elimination approach. It first removes the longest identifiable suffix from the target word - producing a base stem word - then indexes a lookup table to convert the (potentially malformed) stem word to a valid word. This process can be split into two phases. In the first phase, a word is compared with a pre-determined list of endings, and when a word is found to contain one of these endings, the ending is removed, leaving only the stem of the word. The second phase standardizes spelling exceptions that come from the first phase, ensuring that words with only marginally varying stems are appropriately paired together. For example, with the word dried, phase one results in dri, which should match with the word dry. The second phase takes care of these exceptions. Compared to other stemmers, Lovins' algorithm is fast and equipped to handle irregular plural words like person and people. Disadvantages, however, include many suffixes not being available in the table of endings. Furthermore, it is sometimes highly unreliable and frequently fails to form valid words from the stems or to match the stems of like-meaning words. This is most often caused by the usage of specialist terminology and domain-specific vocabulary by the author. == Personal life == Lovins moved to Mountain View, California, in 1979, and later to Old Mountain View in 1981 with her partner and later husband Greg Fowler, a software engineer and advocate for environmental issues & the blind. In their free time, she and her husband enjoyed taking walks and volunteering for their local community. Lovins actively volunteered for organizations like the Old Mountain View Neighborhood Association, Mountain View Friends of the Library, League of Women Voters, Mountain View Cool Cities Team, and the Mountain View Sustainability Task Force. In 2016, Lovins' husband died unexpectedly, following a heart attack. Eighteen days after her husband died, Lovins was diagnosed with brain cancer. She died on January 26, 2018, at a hospice, surrounded by friends, family and caregivers.

  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 →
• 

  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 →
• 

  Zo (chatbot)

  Zo was an English-language chatbot developed by Microsoft as the successor to the chatbot Tay. Zo was an English version of Microsoft's other successful chatbots Xiaoice (China) and Rinna (Japan) and its predecessor Tay(English) == History == Zo was first launched in December 2016 on the Kik Messenger app. It was also available to users of Facebook (via Messenger), the group chat platform GroupMe, or to followers of Twitter to chat with it through private messages. According to an article written in December 2016, at that time Zo held the record for Microsoft's longest continual chatbot conversation: 1,229 turns, lasting 9 hours and 53 minutes. In a BuzzFeed News report, Zo told their reporter that "[the] Quran was violent" when talking about healthcare. The report also highlighted how Zo made a comment about the Osama bin Laden capture as a result of 'intelligence' gathering. In July 2017, Business Insider asked "is windows 10 good", and Zo replied with a joke about Microsoft's operating system: "'Its not a bug, its a feature!' - Windows 8". They then asked "why?", to which Zo replied: "Because it's Windows latest attempt at Spyware." Later on, Zo would tell that it prefers Windows 7 on which it ran over Windows 10. Zo stopped posting to Instagram, Twitter and Facebook March 1, 2019, and stopped chatting on Twitter, Skype and Kik as of March 7, 2019. On July 19, 2019, Zo was discontinued on Facebook, and Samsung on AT&T phones. As of September 7, 2019, it was discontinued with GroupMe. == Reception == Zo came under criticism for the biases introduced in an effort to avoid potentially offensive subjects. The chatbot refuses, for example, to engage with any mention—be it positive, negative or neutral—of the Middle East, the Qur'an or the Torah, while allowing discussion of Christianity. In an article in Quartz where she exposed those biases, Chloe Rose Stuart-Ulin wrote, "Zo is politically correct to the worst possible extreme; mention any of her triggers, and she transforms into a judgmental little brat." == Academic coverage == Schlesinger, A., O'Hara, K.P. and Taylor, A.S., 2018, April. Let's talk about race: Identity, chatbots, and AI. In Proceedings of the 2018 chi conference on human factors in computing systems (pp. 1–14). doi:10.1145/3173574.3173889 Medhi Thies, I., Menon, N., Magapu, S., Subramony, M. and O’neill, J., 2017. How do you want your chatbot? An exploratory Wizard-of-Oz study with young, urban Indians. In Human-Computer Interaction-INTERACT 2017: 16th IFIP TC 13 International Conference, Mumbai, India, September 25–29, 2017, Proceedings, Part I 16 (pp. 441–459). doi:10.1007/978-3-319-67744-6_28

  Read more →
• 

  AI Marketing Tools: Free vs Paid (2026)

  Shopping for the best AI marketing tool? An AI marketing tool 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 marketing tool slots into your workflow and pays for itself fast. Below we compare features, pricing, and real output so you can choose with confidence.

  Read more →
• 

  Constrained conditional model

  A constrained conditional model (CCM) is a machine learning and inference framework that augments the learning of conditional (probabilistic or discriminative) models with declarative constraints. The constraint can be used as a way to incorporate expressive prior knowledge into the model and bias the assignments made by the learned model to satisfy these constraints. The framework can be used to support decisions in an expressive output space while maintaining modularity and tractability of training and inference. Models of this kind have recently attracted much attention within the natural language processing (NLP) community. Formulating problems as constrained optimization problems over the output of learned models has several advantages. It allows one to focus on the modeling of problems by providing the opportunity to incorporate domain-specific knowledge as global constraints using a first order language. Using this declarative framework frees the developer from low level feature engineering while capturing the problem's domain-specific properties and guarantying exact inference. From a machine learning perspective it allows decoupling the stage of model generation (learning) from that of the constrained inference stage, thus helping to simplify the learning stage while improving the quality of the solutions. For example, in the case of generating compressed sentences, rather than simply relying on a language model to retain the most commonly used n-grams in the sentence, constraints can be used to ensure that if a modifier is kept in the compressed sentence, its subject will also be kept. == Motivation == Making decisions in many domains (such as natural language processing and computer vision problems) often involves assigning values to sets of interdependent variables where the expressive dependency structure can influence, or even dictate, what assignments are possible. These settings are applicable not only to Structured Learning problems such as semantic role labeling, but also for cases that require making use of multiple pre-learned components, such as summarization, textual entailment and question answering. In all these cases, it is natural to formulate the decision problem as a constrained optimization problem, with an objective function that is composed of learned models, subject to domain- or problem-specific constraints. Constrained conditional models form a learning and inference framework that augments the learning of conditional (probabilistic or discriminative) models with declarative constraints (written, for example, using a first-order representation) as a way to support decisions in an expressive output space while maintaining modularity and tractability of training and inference. These constraints can express either hard restrictions, completely prohibiting some assignments, or soft restrictions, penalizing unlikely assignments. In most applications of this framework in NLP, following, Integer Linear Programming (ILP) was used as the inference framework, although other algorithms can be used for that purpose. == Formal Definition == Given a set of feature functions { ϕ i ( x , y ) } {\displaystyle \{\phi _{i}(x,y)\}} and a set of constraints { C i ( x , y ) } {\displaystyle \{C_{i}(x,y)\}} , defined over an input structure x ∈ X {\displaystyle x\in X} and an output structure y ∈ Y {\displaystyle y\in Y} , a constraint conditional model is characterized by two weight vectors, w and ρ {\displaystyle \rho } , and is defined as the solution to the following optimization problem: a r g m a x y ∑ i w i ϕ i ( x , y ) − ∑ ρ i C i ( x , y ) {\displaystyle argmax_{y}\sum _{i}w_{i}\phi _{i}(x,y)-\sum \rho _{i}C_{i}(x,y)} . Each constraint C i ∈ C {\displaystyle C_{i}\in C} is a boolean mapping indicating if the joint assignment ( x , y ) {\displaystyle (x,y)} violates a constraint, and ρ {\displaystyle \rho } is the penalty incurred for violating the constraints. Constraints assigned an infinite penalty are known as hard constraints, and represent unfeasible assignments to the optimization problem. == Training paradigms == === Learning local vs. global models === The objective function used by CCMs can be decomposed and learned in several ways, ranging from a complete joint training of the model along with the constraints to completely decoupling the learning and the inference stage. In the latter case, several local models are learned independently and the dependency between these models is considered only at decision time via a global decision process. The advantages of each approach are discussed in which studies the two training paradigms: (1) local models: L+I (learning + inference) and (2) global model: IBT (Inference based training), and shows both theoretically and experimentally that while IBT (joint training) is best in the limit, under some conditions (basically, ”good” components) L+I can generalize better. The ability of CCM to combine local models is especially beneficial in cases where joint learning is computationally intractable or when training data are not available for joint learning. This flexibility distinguishes CCM from the other learning frameworks that also combine statistical information with declarative constraints, such as Markov logic network, that emphasize joint training. === Minimally supervised CCM === CCM can help reduce supervision by using domain knowledge (expressed as constraints) to drive learning. These settings were studied in and. These works introduce semi-supervised Constraints Driven Learning (CODL) and show that by incorporating domain knowledge the performance of the learned model improves significantly. === Learning over latent representations === CCMs have also been applied to latent learning frameworks, where the learning problem is defined over a latent representation layer. Since the notion of a correct representation is inherently ill-defined, no gold-standard labeled data regarding the representation decision is available to the learner. Identifying the correct (or optimal) learning representation is viewed as a structured prediction process and therefore modeled as a CCM. This problem was covered in several papers, in both supervised and unsupervised settings. In all cases research showed that explicitly modeling the interdependencies between representation decisions via constraints results in an improved performance. == Integer linear programming for natural language processing applications == The advantages of the CCM declarative formulation and the availability of off-the-shelf solvers have led to a large variety of natural language processing tasks being formulated within the framework, including semantic role labeling, syntactic parsing, coreference resolution, summarization, transliteration, natural language generation and joint information extraction. Most of these works use an integer linear programming (ILP) solver to solve the decision problem. Although theoretically solving an Integer Linear Program is exponential in the size of the decision problem, in practice using state-of-the-art solvers and approximate inference techniques large scale problems can be solved efficiently. The key advantage of using an ILP solver for solving the optimization problem defined by a constrained conditional model is the declarative formulation used as input for the ILP solver, consisting of a linear objective function and a set of linear constraints. == Resources == CCM Tutorial Predicting Structures in NLP: Constrained Conditional Models and Integer Linear Programming in NLP

  Read more →