AUTINDEX is a commercial text mining software package based on sophisticated linguistics. AUTINDEX, resulting from research in information extraction, is a product of the Institute of Applied Information Sciences (IAI) which is a non-profit institute that has been researching and developing language technology since its foundation in 1985. IAI is an institute affiliated to Saarland University in Saarbrücken, Germany. AUTINDEX is the result of a number of research projects funded by the EU (Project BINDEX), by Deutsche Forschungsgemeinschaft and the German Ministry for Economy. Amongst the latter there are the projects LinSearch, and WISSMER, see also the reference to IAI-Website. The basic functionality of AUTINDEX is the extraction of key words from a document to represent the semantics of the document. Ideally the system is integrated with a thesaurus that defines the standardised terms to be used for key word assignment. AUTINDEX is used in library applications (e.g. integrated in dandelon.com) as well as in high quality (expert) information systems, and in document management and content management environments. Together with AUTINDEX a number of additional software comes along such as an integration with Apache Solr / Lucene to provide a complete information retrieval environment, a classification and categorisation system on the basis of a machine learning software that assigns domains to the document, and a system for searching with semantically similar terms that are collected in so called tag clouds.
Labeled data
Labeled data is a group of samples that have been tagged with one or more labels. Labeling typically takes a set of unlabeled data and augments each piece of it with informative tags called judgments. For example, a data label might indicate whether a photo contains a horse or a cow, which words were uttered in an audio recording, what type of action is being performed in a video, what the topic of a news article is, what the overall sentiment of a tweet is, or whether a dot in an X-ray is a tumor. Labels can be obtained by having humans make judgments about a given piece of unlabeled data. Labeled data is significantly more expensive to obtain than the raw unlabeled data. The quality of labeled data directly influences the performance of supervised machine learning models in operation, as these models learn from the provided labels. == Crowdsourced labeled data == In 2006, Fei-Fei Li, the co-director of the Stanford Human-Centered AI Institute, initiated research to improve the artificial intelligence models and algorithms for image recognition by significantly enlarging the training data. The researchers downloaded millions of images from the World Wide Web and a team of undergraduates started to apply labels for objects to each image. In 2007, Li outsourced the data labeling work on Amazon Mechanical Turk, an online marketplace for digital piece work. The 3.2 million images that were labeled by more than 49,000 workers formed the basis for ImageNet, one of the largest hand-labeled database for outline of object recognition. == Automated data labelling == After obtaining a labeled dataset, machine learning models can be applied to the data so that new unlabeled data can be presented to the model and a likely label can be guessed or predicted for that piece of unlabeled data. == Challenges == === Data-driven bias === Algorithmic decision-making is subject to programmer-driven bias as well as data-driven bias. Training data that relies on bias labeled data will result in prejudices and omissions in a predictive model, despite the machine learning algorithm being legitimate. The labeled data used to train a specific machine learning algorithm needs to be a statistically representative sample to not bias the results. For example, in facial recognition systems underrepresented groups are subsequently often misclassified if the labeled data available to train has not been representative of the population,. In 2018, a study by Joy Buolamwini and Timnit Gebru demonstrated that two facial analysis datasets that have been used to train facial recognition algorithms, IJB-A and Adience, are composed of 79.6% and 86.2% lighter skinned humans respectively. === Human error and inconsistency === Human annotators are prone to errors and biases when labeling data. This can lead to inconsistent labels and affect the quality of the data set. The inconsistency can affect the machine learning model's ability to generalize well. === Domain expertise === Certain fields, such as legal document analysis or medical imaging, require annotators with specialized domain knowledge. Without the expertise, the annotations or labeled data may be inaccurate, negatively impacting the machine learning model's performance in a real-world scenario.
Wonder.land
Wonder.land (stylised as wonder.land) is a musical with music by Damon Albarn and lyrics and book by Moira Buffini. Inspired by Lewis Carroll's novels Alice's Adventures in Wonderland (1865) and Through the Looking-Glass (1871), it had its world premiere at the Palace Theatre in Manchester in July 2015 as part of the Manchester International Festival. The musical moved to London's Royal National Theatre in November 2015 before opening at the Théâtre du Châtelet in Paris in 2016. Licencing for potential future smaller scale productions is held by United Agents UK. == Background == The musical is inspired by the novels Alice in Wonderland and Through the Looking-Glass, written by Lewis Carroll. It was announced on 21 January 2015 that the show would premiere in July of that year as part of the Manchester International Festival, with tickets going on sale the following day. The musical, a co-production by the Manchester International Festival, the Royal National Theatre and the Théâtre du Châtelet in Paris, marks the 150th anniversary of the publication of Alice's Adventures in Wonderland. The idea for a musical based on Alice in Wonderland came from Manchester International Festival artistic director Alex Poots. Damon Albarn had collaborated with the festival on Monkey: Journey to the West and Dr Dee. The musical has a book by Moira Buffini. It was directed by Rufus Norris, with set design by Rae Smith, costume design by Katrina Lindsay, lighting design by Paule Constable, projections by 59 Productions and choreography by Javier De Frutos. The musical's score was composed by Damon Albarn, with lyrics by Moira Buffini, sound design by Paul Arditti and musical direction by David Shrubsole. == Production history == The musical began previews at the Palace Theatre in Manchester on 29 June 2015. It opened on 2 July for a limited run until 12 July. A revised version moved to the Royal National Theatre, where it ran at the Olivier Theatre from 27 November 2015 to 30 April 2016. The production had a limited run, from 7 to 16 June 2016, at the Theatre Du Chatelet in Paris. == Synopsis == This synopsis is based on the final version, as seen at the National Theatre and the Théâtre du Châtelet. Earlier performances significantly differed in songs and plot. === Act 1 === AI, the MC, explains that virtual technology is "a portal to boundless lands" ("Prologue"). Aly's mother, Bianca, is exasperated with her for spending the weekend indoors on her phone. Aly accompanies Bianca to the supermarket, and thinks that her life is being ruined by her parents due to dysfunctional problems ("Who's Ruining Your Life?") Her alcoholic father, Matt, is also at the supermarket; he and Bianca argue about their divorce and his gambling. Aly goes home and picks up her phone. She tries to engage with schoolmates, who bully her ("Network"). Aly begins to wish that she is someone else. She finds the virtual online game Wonder.land. In its strange world, Aly creates an avatar: beautiful, kind Alice ("Wonder.land"). Wonder.land has one rule: malice causes deletion from the game. Aly and Alice become friends and encounter the Cheshire Cat, who explains that you can be anyone you want ("Fabulous"). Aly decides to go on a quest; Alice follows the white rabbit down a hole, falling past unusual objects and musical notes ("Falling"). The next morning, Aly is too distracted by Wonder.land to listen to Bianca's complaints about her baby brother Charlie. She plays the game at school before her phone is confiscated by stern headmistress Ms Manxome, who tells her students that taking pleasures from them is for their own good ("I'm Right"). Aly goes to Ms Manxome's office to retrieve her phone. Ms Manxome returns it, warning that if she catches her with it again, "it's a beheading – I mean, detention." Aly sees the girls who bullied her, and they bully her again until a teacher arrives. Aly's friend, Luke, is late and is given detention. Aly goes on her phone and takes out her frustration and sadness on Alice, whose tears form a pool until she is interrupted by the quarrelsome twins Dum and Dee ("Freaks"). Alice tries to befriend them, but they insult her and Aly makes her fight them. Dum and Dee cry, and Aly and Alice see a large mouse who is attracted by Alice's fighting. They are joined by the Dodo, the Mock Turtle and Humpty, who all have problems. The Dodo is stressed because his parents want him to save the planet; Dum and Dee are dancers who hate pressure; Humpty has problems with her parents; the Mock Turtle lacks self-esteem, and the mouse is lustful. Wonderland is a hiding place from teenage life ("Crap Life"). Aly returns to reality when asked a math question she cannot answer. Confronting the three bullies, Aly mocks the facial hair of one and hides in the bathroom. She again immerses herself in Wonder.land, where Alice meets a Caterpillar who is obsessed with identity ("Who are You?"). Aly is interrupted by the girls, who ridicule her father's gambling addiction and poverty before beating her up. Aly seeks understanding from Alice, who tries to get Aly to tell her what is wrong. Aly tells Alice about her family and how she hates her life, and is surprised that Alice has similar problems ("Secrets"). Luke comes into the girls' bathroom because Kieran has threatened him with violence, and hides in a cubicle when Kieran enters. Aly defends Luke, and makes Kieran leave. Luke reveals that the reason Kieran hates him is because, like himself, he is gay. Aly is amazed, and they skip class and play games on their phones. Luke plays Zombie Swarm, and Aly plays Wonder.land. Ms Manxome enters the bathroom; Luke hides his phone, but Aly does not. Ms Manxome confiscates the phone for three months, and Aly and Luke leave. Ms Manxome finds that Aly did not lock her phone, and Alice is calling her. Ms Manxome begins to talk to her, and Alice thinks she is talking to Aly. Aly complains to Luke about her phone being taken away. Matt then takes them out for tea to celebrate his new job at the local garden centre ("In Clover"). At the tea shop, Matt maniacally dances on the tables and plays with spoons; asked to stop, he punches a waiter. Bianca arrives, and they argue again. Aly begins to notice that Wonder.land is invading reality; the MC emerges from a gigantic teapot, and the landscape outside becomes surreal ("Chances"). === Act 2 === Ms Manxome manipulates Alice around Wonder.land on Aly's phone, buys many things, and makes Alice's hair red ("Entre Act"). She tells Alice about her plans to dominate and destroy the online world, and Alice thinks she is talking to Aly ("Me"). Aly, Matt, Bianca, and Charlie are at the police station. PC Rook unsuccessfully tries to get Matt to make a statement (since he is charged with assault and affray), but Matt and Bianca argue again. Aly laments the loss of her family's unity ("Heartless Useless"). In Wonder.land, Ms Manxome is hostile when she meets Dum and Dee, the Mock Turtle, the Dodo, Humpty and the Mouse. She makes Alice chase them away, but Alice and Ms Manxome are driven away by Alice's friends, who are worried about the change in her ("Me (Reprise)"). Bianca learns that Aly missed a detention and had her phone confiscated. Concerned that she is losing Aly to technology, she bans her from the internet ("Gadget"). Charlie vomits, and Aly is left to clean it up. She looks for an internet cafe to go to Wonder.land, the only place she is truly happy ("Everyone Loves Charlie"). At the cafe, Aly cannot log into Wonder.land and her avatar seems to be in use. She sees Alice receive a Vorpal sword, bought by Ms Manxome with the money on Aly's phone. Alice is no longer Alice but the Red Queen, and Ms Manxome tells her to kill her friends. Alice, knowing the person controlling her is not Aly, cannot rebel; she lashes out at her friends, bullying and trying to hurt them. The MC warns that Alice has a deletion warning – any more malice, and she will be deleted. Aly now knows that Ms Manxome controls her phone and avatar ("O Children"). Aly enlists Luke to help and decides to break into Ms. Manxome's office to retrieve the phone. Luke agrees to meet her at the school gates. Matt and Bianca wonder if they should reconcile ("Man of Broken Glass"). At the school, Luke is reluctant to get involved; Aly decides to break into the office anyway. Luke contacts the girls who bullied Aly and tells them about Ms Manxome playing on Aly's stolen phone. They decide to spread the word that it is not Aly ("Fabulous (Reprise)"). Bianca goes to the police because Aly is missing, and gives her phone to Matt. Aly is likely to also be in Wonder.land. The avatars prepare for war against Alice but disagree about a strategy. At the police station, Matt hacks into Wonder.land sees Alice, and realizes that she is controlled by someone other than Aly. The White Rabbit appears (delighting Alice), but Ms Manxome makes Alice push him aside. The borderline between Wonder.land and
It's the Most Terrible Time of the Year
It's the Most Terrible Time of the Year is an AI-generated television commercial created for McDonald's Netherlands by TBWA\Neboko and The Sweetshop. It was released on 6 December 2025 before being pulled four days later due to negative reception over its use of generative artificial intelligence and its cynical, negative depiction of the holiday season. == Plot == On a bleak, snowy day, various people in the city experience different kinds of mishaps during the Christmas season. Among other incidents, families struggle with their huge loads of presents; Santa Claus gets stuck in traffic; a Christmas tree "redecorates" a man's home, sending him through the window; another family puts up with annoying relatives and a burnt Christmas dinner. Because of all this chaos, a man decides to find refuge in a McDonald's outlet. A Christmas choir finishes singing the jingle "It's the Most Terrible Time of the Year" with the call to action to "hide out in McDonald's till January's here". == Campaign == It's the Most Terrible Time of the Year is a 45-second television commercial made by Dutch agency TBWA\Neboko with involvement of United States-based film production studio The Sweetshop. The advertisement was produced heavily with generative artificial intelligence (AI) following the trend set by other brands such as Coca-Cola and Toys "R" Us. McDonald's Netherlands, the client, released a statement that the commercial was meant to depict "the stressful moments during the holidays in the Netherlands". The commercial also used Andy Williams's "It's the Most Wonderful Time of the Year" with lyrics changed to fit with the concept of the advertisement. According to The Sweetshop, the production of the advertisement took "seven weeks". It also added that much effort was put into the commercial compared to the traditional process. Ten people of its in-house AI engine The Gardening Club worked on the project. Los Angeles-based directors Mark Potoka and Matt Spicer were initially credited to be involved in the film but they resigned due to being sidelined from the production process. == Reception == The advertisement was released on McDonald's Netherlands' YouTube channel on 6 December 2025. It had a negative reception over the use of generative AI and the "cynical" concept of the work's story. The video was made private on 9 December 2025. The Sweetshop stated that the production of the advertisement took human effort. McDonald's Netherlands, while stating the original intent of the commercial, released a statement after its pullout that, for many of its customers, the holiday season is the "most wonderful time of the year".
Refik Anadol
Refik Anadol (born November 7, 1985) is a Turkish American media artist and the co-founder of Refik Anadol Studio and Dataland. Recognized as a pioneer in the aesthetics of data visualization and AI arts, his work merges art, technology, science, and architecture. Through media embedded into existing architecture, live audio-visual performances, immersive rooms, exhibitions, AI data paintings and sculptures, and digital collections, Anadol explores collective memories, humanity's relationship to nature, the perception of space and time, and human-machine collaborations. His work has been exhibited in more than seventy cities on six continents. == Early life and education == Anadol was born and raised in Istanbul and grew up in a family of teachers. He taught himself basic programming on a Commodore 64 when he was eight. His connection to machines began with coding and video games. Anadol saw Blade Runner for the first time when he was eight; his mother said the way he perceived his surroundings shifted the day after he saw the film. He was fascinated with its futuristic depiction of downtown Los Angeles, and transfixed by as a scene during which a replicant discovers that her memories are an implanted component of her machine mind, In a 2024 interview with the Financial Times, he said: "Since that moment, one of my inspirations has been that question: 'What can a machine do with someone else's memories?" Anadol attended Istanbul Bilgi University, where he received a BA in photography and video in 2009 and an MFA in visual communication in 2011. In 2014 he earned an MFA in design media arts at UCLA. He was mentored by Casey Reas, Jennifer Steinkamp, and Christian Moeller. == Career and selected works == === 2008–2012: Data painting, Quadrature and Quadrangle, Istanbul Biennial === As an undergraduate, Anadol read a paper by Lev Manovich on augmented space. Manovich's assertion that collaborations between architects and artists could make the "invisible flow of data visible" triggered Anadol's imagination, and in 2008, he altered built space for the first time. Bringing a projector outside, he projected large-scale images onto a concrete to create the illusion of movement. Coining the term "data painting," the piece inspired Anadol to use light as material and data as pigment. In 2010 he created Quadrature with Alican Aktürk, a fellow graduate student, at the SantralIstanbul Art and Culture Center's main gallery building. A live audio-visual performance that examined the relationship between architecture and media, Quadrature used video projection techniques to manipulate footage of quadrilaterals. He followed Quadrature with Quadrangle at SANAA School of Design in Essen, Germany, using the entire 360 degrees of the building as a canvas. In 2011, he was invited to create a media installation at the Istanbul Biennial on the heavily trafficked İstiklal Avenue. He created a site-specific large-scale interpretation of sounds he recorded during different times of day, and used nine projectors to project reinterpreted images. The work was titled Augmented Structures v1.0. Anadol's first solo exhibition, Sceptical Interventions, was held at the Piveneli Gallery in Istanbul in early 2012. Later that year he moved to Los Angeles to attend UCLA's Design Media Arts program. The first place he went after his arrival was downtown Los Angeles. [6] === 2013–2016: Visions of America: Amériques, Infinity Room, Google AMI === In 2013, at Microsoft Research's annual Design Expo, Anadol presented his idea to use the external walls of Walt Disney Concert Hall as a canvas. His presentation brought him to the attention of Gehry Technologies, and with the support of Gehry and his team, Anadol was offered the use of the original 3D model of the concert hall. For his 2014 thesis project, with assistance from architects and UCLA researchers, he created a site-specific architectural video installation inside the concert hall that accompanied a Los Angeles Philharmonic performance of Edgard Varèse's Amérique. Titled Visions of America: Amériques, Anadol used algorithmic sound analysis to listen and respond to the music in real-time. He tracked conductor Esa-Pekka Salonen's heartbeat with a sensor and used a 3-D camera system to integrate Salonen's movements. He created Infinity Room at the Zorlu PSM for the 2015 Istanbul Biennial. Rather than creating an illusion only with mirrors, Anadol used pixel and 3D projection mapping to transform every surface of the room into an abstract infinite moving space. A temporary immersive environment, Infinity Room was also exhibited at events including South by Southwest in Austin, Texas, the New Zealand Festival in Wellington, New Zealand, and Jeffrey Deitch in Los Angeles. In 2016, Anadol was awarded the first Google Artists and Machine Intelligence Artist Residency; it was just after a team at Google opened up the algorithm for DeepDream, a computer vision program that prompted Anadol's realization that if a machine could learn, it could remember, dream, and hallucinate. === 2017–2018: Winds of Boston, Archive Dreaming, Melting Memories, WDCH Dreams === In 2017, he created the data painting Winds of Boston, a 6' x 13' foot video installation in the lobby of a Boston office building, using software he created to read, analyze and visualize wind speed, direction, and gust patterns along with time and temperature at 20-second intervals recorded over a one-year period at Logan International Airport. Later in the year, he used AI to generate infinite new outputs based on a massive dataset for Archive Dreaming, an immersive installation at Salt Research, a contemporary gallery and library in Istanbul. Inspired by his idea of consciousness and its context within AI, as well as Jorge Luis Borges' The Library of Babel, Anadol used AI and machine learning to look at and discover interactions and correlations between 1.7 million items culled from 40,000 publications covering Turkish contemporary and modern art, architecture, and economics from 1997 to 2010. Archive Dreaming, which could be controlled by users with a joystick, dreamed of unexpected correlations among documents when idle. In 2018, after his uncle was diagnosed with Alzheimer's, Anadol created Melting Memories. Working with scientists from the neuroscape laboratory at the University of California, San Francisco, he used academic data from the neuroscience archives and EEG scans of an anonymous Alzheimer's disease dataset to create AI-generated visuals related to memory, health, degeneration, and decay.Melting Memories was projected on the walls of Pilevneli Gallery; visitors to the exhibition could watch as millions of pixels reconstructed people's memories. Anadol won the Lumen Prize Gold Award for Melting Memories. Anadol was commissioned by the Los Angeles Philharmonic to create an installation to celebrate the orchestra's centennial anniversary in 2018. He worked with Google's Kenric MacDowell to create WDCH Dreams, using algorithmic visualizations of data to mimic the process of human dreaming. Projected across the exterior walls of Walt Disney Concert Hall using 42 large-scale projectors with 50K visual resolution, 8-channel sound, and 1.2M luminance, Anadol painted with data points culled from the orchestra's archives, including 587,763 images, 1,880 videos, 1,483 metadata files, and 17,773 audio files. Because Gehry gave him access to the 3D architectural files of Walt Disney Concert Hall, Anadol knew the exact contours of the building. WDCH Dreams debuted in September 2018. A 12-minute performance in three parts staged every 30 minutes over ten nights, "Centennial Memories,” the first piece, used 44.5 terabytes of historical data from the Phil's archives. It was followed by "Consciousness", which processed every note the orchestra has ever recorded, using billions of data points to generate connections; and "Dream," which merged "Centennial Memories" and "Consciousness" to create hallucinations that were described in the New York Times as "a sort of combinatorial Fantasia. === 2019–2021: Machine Hallucinations: NYC, Machine Hallucinations: Nature Dreams, Machine Memories: Space, Quantum Memories === In 2019, Refik Anadol presented Latent History at Fotografiska Stockholm. The site specific installation transformed photographic archives of Stockholm into a large scale, machine generated visual projection displayed in the museum’s main exhibition hall. Drawing on thousands of archival images spanning approximately 150 years, the work used artificial intelligence to reinterpret the city’s historical imagery as a continuously evolving visual narrative.. Anadol began thinking about the work that would become the Machine Hallucinations series while in residence at Google. In 2019, he completed the first work in the series, Machine Hallucinations: NYC, which used 300 million photos of New York City and 113 million additional data points, including subway sounds, ra
NAPLPS
NAPLPS (North American Presentation Layer Protocol Syntax) is a graphics language for use originally with videotex and teletext services. NAPLPS was developed from the Telidon system developed in Canada, with a small number of additions from AT&T Corporation. The basics of NAPLPS were later used as the basis for several other microcomputer-based graphics systems. == History == The Canadian Communications Research Centre (CRC), based in Ottawa, had been working on various graphics systems since the late 1960s, much of it led by Herb Bown. Through the 1970s they turned their attention to building out a system of "picture description instructions", which encoded graphics commands as a text stream. Graphics were encoded as a series of instructions (graphics primitives) each represented by a single ASCII character. Graphic coordinates were encoded in multiple 6-bit strings of XY coordinate data, flagged to place them in the printable ASCII range so that they could be transmitted with conventional text transmission techniques. ASCII SI/SO characters were used to differentiate the text from graphic portions of a transmitted "page". These instructions were decoded by separate programs to produce graphics output, on a plotter for instance. Other work produced a fully interactive version. In 1975, the CRC gave a contract to Norpak to develop an interactive graphics terminal that could decode the instructions and display them on a color display. During this period, a number of companies were developing the first teletext systems, notably the BBC's Ceefax system. Ceefax encoded character data into the lines in the vertical blanking interval of normal television signals where they could not be seen on-screen, and then used a buffer and decoder in the user's television to convert these into "pages" of text on the display. The Independent Broadcasting Authority quickly introduced their own ORACLE system, and the two organizations subsequently agreed to use a single standard, the "Broadcast Teletext Specification". This later became World System Teletext. At about the same time, other organizations were developing videotex systems, similar to teletext except they used modems to transmit their data instead of television signals. This was potentially slower and used up a telephone line, but had the major advantage of allowing the user to transmit data back to the sender. The UK's General Post Office developed a system using the Ceefax/ORACLE standard, launching it as Prestel, while France prepared the first steps for its ultimately very successful Minitel system, using a rival display standard called Antiope. By 1977, the Norpak system was running, and from this work the CRC decided to create their own teletext/videotext system. Unlike the systems being rolled out in Europe, the CRC decided from the start that the system should be able to run on any combination of communications links. For instance, it could use the vertical blanking interval to send data to the user, and a modem to return selections to the servers. It could be used in a one-way or two-way system. In teletext mode, character codes were sent to users' televisions by encoding them as dot patterns in the vertical blanking interval of the video signal. Various technical "tweaks" and details of the NTSC signals used by North American televisions allowed the downstream videotex channel to increase to 600 bit/s, about twice that used in the European systems. In videotext mode, Bell 202 modems were typical, offering a 1,200 bit/s download rate. A set top box attached to the TV decoded these signals back into text and graphics pages, which the user could select among. The system was publicly launched as Telidon on August 15, 1978. Compared to the European standards, the CRC system was faster, bi-directional, and offered real graphics as opposed to simple character graphics. The downside of the system was that it required much more advanced decoders, typically featuring Zilog Z80 or Motorola 6809 processors with RGB and/or RF output. The Innovation, Science and Economic Development Canada (then Department of Communications) launched a four-year plan to fund public roll-outs of the technology in an effort to spur the development of a commercial Telidon system. AT&T Corporation was so impressed by Telidon that they decided to join the project. They added a number of useful extensions, notably the ability to define original graphics commands (macro) and character sets (DRCS). They also tabled algorithms for proportionally spaced text, which greatly improved the quality of the displayed pages. A joint CSA/ANSI working group (X3L2.1) revised the specifications, which were submitted for standardization. In 1983, they became CSA T500 and ANSI X3.110, or NAPLPS. The data encoding system was also standardized as the NABTS (North American Broadcast Teletext Specification) protocol. Business models for Telidon services were poorly developed. Unlike the UK, where teletext was supported by one of only two large companies whose whole revenue model was based on a read-only medium (television), in North America Telidon was being offered by companies who worked on a subscriber basis. == One-way systems == Telidon-based teletext was tested in a few North American trials in the early 1980s — CBC IRIS, TVOntario, MTS-sponsored Project IDA, to name a few. NAPLPS was also part of the NABTS teletext standard, for the encoding and display of teletext pages. In the late 1980s and early 1990s, affiliates of the regional sports network group SportsChannel ran a service called Sports Plus Network, which ran sports news and scores while SportsChannel was not otherwise on the air. The screens, which frequently featured team logos or likenesses of players in addition to text, were drawn entirely with NAPLPS graphics and resembled the loading of Prodigy pages over a modem, though slightly faster. == Two-way systems == Various two-way systems using NAPLPS appeared in North America in the early 1980s. The biggest North American examples were Knight Ridder's Viewtron (based in Miami) and the Los Angeles Times' Gateway service (based in Orange County). Both used the Sceptre NAPLPS terminal from AT&T. The Sceptre contained a slow modem that connected over the consumer's telephone line to host computers. The Sceptre was expensive whether purchased or rented. Despite huge investments by their parent companies, neither Viewtron nor Gateway lasted into the second half of the decade. Another system, Keyfax, was developed by Keycom Electronic Publishing, a joint venture of Honeywell, Centel (since acquired by Sprint) and Field Enterprises, then-owner of the Chicago Sun-Times newspaper. Keyfax had originally been a WST teletext service, broadcast overnights on Field's Chicago television station WFLD-32 and through the VBI of both WFLD and national superstation WTBS; the decision was made to convert Keyfax into a subscription service, using a proprietary NAPLPS terminal device in a last-ditch effort to save the service. It did not work and Keyfax had ceased operations by the end of 1986. Other early-1980s NAPLPS technology was deployed in Canada, both as a way for rural Canadians to get news and weather information and as the platform for touchscreen information kiosks. In Vancouver these were featured at Expo 86. The kiosks became ubiquitous in Toronto under the name Teleguide, and were deployed in many shopping centres and at major tourist attractions. The latter city was the North American nexus of NAPLPS and the home of Norpak, the most successful of NAPLPS-oriented developers. Norpak created and sold hardware and software for NAPLPS development and display. TVOntario also developed NAPLPS content creation software. London, Ontario - based Cableshare used NAPLPS as the basis of touch-screen information kiosks for shopping malls, the flagship of which was deployed at Toronto's Eaton Centre. The system relied on an 8085-based microcomputer which drove several NAPLPS terminals fitted with touch screens, all communicating via Datapac to a back end database. The system offered news, weather and sports information along with shopping mall guides and coupons. Cableshare also developed and sold a leading NAPLPS page creation utility called the "Picture Painter." In the late 1980s, Tribune Media Services (TMS) and the Associated Press operated a cable television channel called AP News Plus that provided NAPLPS-based news screens to cable television subscribers in many U.S. cities. The news pages were created and edited by TMS staffers working on an Atex editing system in Orlando, Florida, and sent by satellite to NAPLPS decoder devices located at the local cable television companies. Among the firms providing technology to TMS and the Associated Press for the AP News Plus channel was Minneapolis-based Electronic Publishers Inc. (1985–1988). In 1981, two amateur radio operators (VE3FTT and VE3GQW) received special permission from the Canad
Batch normalization
In artificial neural networks, batch normalization (also known as batch norm) is a normalization technique used to make training faster and more stable by adjusting the inputs to each layer—re-centering them around zero and re-scaling them to a standard size. It was introduced by Sergey Ioffe and Christian Szegedy in 2015. Experts still debate why batch normalization works so well. It was initially thought to tackle internal covariate shift, a problem where parameter initialization and changes in the distribution of the inputs of each layer affect the learning rate of the network. However, newer research suggests it doesn’t fix this shift but instead smooths the objective function—a mathematical guide the network follows to improve—enhancing performance. In very deep networks, batch normalization can initially cause a severe gradient explosion—where updates to the network grow uncontrollably large—but this is managed with shortcuts called skip connections in residual networks. Another theory is that batch normalization adjusts data by handling its size and path separately, speeding up training. == Internal covariate shift == Each layer in a neural network has inputs that follow a specific distribution, which shifts during training due to two main factors: the random starting values of the network’s settings (parameter initialization) and the natural variation in the input data. This shifting pattern affecting the inputs to the network’s inner layers is called internal covariate shift. While a strict definition isn’t fully agreed upon, experiments show that it involves changes in the means and variances of these inputs during training. Batch normalization was first developed to address internal covariate shift. During training, as the parameters of preceding layers adjust, the distribution of inputs to the current layer changes accordingly, such that the current layer needs to constantly readjust to new distributions. This issue is particularly severe in deep networks, because small changes in shallower hidden layers will be amplified as they propagate within the network, resulting in significant shift in deeper hidden layers. Batch normalization was proposed to reduced these unwanted shifts to speed up training and produce more reliable models. Beyond possibly tackling internal covariate shift, batch normalization offers several additional advantages. It allows the network to use a higher learning rate—a setting that controls how quickly the network learns—without causing problems like vanishing or exploding gradients, where updates become too small or too large. It also appears to have a regularizing effect, improving the network’s ability to generalize to new data, reducing the need for dropout, a technique used to prevent overfitting (when a model learns the training data too well and fails on new data). Additionally, networks using batch normalization are less sensitive to the choice of starting settings or learning rates, making them more robust and adaptable. == Procedures == === Transformation === In a neural network, batch normalization is achieved through a normalization step that fixes the means and variances of each layer's inputs. Ideally, the normalization would be conducted over the entire training set, but to use this step jointly with stochastic optimization methods, it is impractical to use the global information. Thus, normalization is restrained to each mini-batch in the training process. Let us use B to denote a mini-batch of size m of the entire training set. The empirical mean and variance of B could thus be denoted as μ B = 1 m ∑ i = 1 m x i {\displaystyle \mu _{B}={\frac {1}{m}}\sum _{i=1}^{m}x_{i}} and σ B 2 = 1 m ∑ i = 1 m ( x i − μ B ) 2 {\displaystyle \sigma _{B}^{2}={\frac {1}{m}}\sum _{i=1}^{m}(x_{i}-\mu _{B})^{2}} . For a layer of the network with d-dimensional input, x = ( x ( 1 ) , . . . , x ( d ) ) {\displaystyle x=(x^{(1)},...,x^{(d)})} , each dimension of its input is then normalized (i.e. re-centered and re-scaled) separately, x ^ i ( k ) = x i ( k ) − μ B ( k ) ( σ B ( k ) ) 2 + ϵ {\displaystyle {\hat {x}}_{i}^{(k)}={\frac {x_{i}^{(k)}-\mu _{B}^{(k)}}{\sqrt {\left(\sigma _{B}^{(k)}\right)^{2}+\epsilon }}}} , where k ∈ [ 1 , d ] {\displaystyle k\in [1,d]} and i ∈ [ 1 , m ] {\displaystyle i\in [1,m]} ; μ B ( k ) {\displaystyle \mu _{B}^{(k)}} and σ B ( k ) {\displaystyle \sigma _{B}^{(k)}} are the per-dimension mean and standard deviation, respectively. ϵ {\displaystyle \epsilon } is added in the denominator for numerical stability and is an arbitrarily small positive constant. The resulting normalized activation x ^ ( k ) {\displaystyle {\hat {x}}^{(k)}} have zero mean and unit variance, if ϵ {\displaystyle \epsilon } is not taken into account. To restore the representation power of the network, a transformation step then follows as y i ( k ) = γ ( k ) x ^ i ( k ) + β ( k ) {\displaystyle y_{i}^{(k)}=\gamma ^{(k)}{\hat {x}}_{i}^{(k)}+\beta ^{(k)}} , where the parameters γ ( k ) {\displaystyle \gamma ^{(k)}} and β ( k ) {\displaystyle \beta ^{(k)}} are subsequently learned in the optimization process. Formally, the operation that implements batch normalization is a transform B N γ ( k ) , β ( k ) : x 1... m ( k ) → y 1... m ( k ) {\displaystyle BN_{\gamma ^{(k)},\beta ^{(k)}}:x_{1...m}^{(k)}\rightarrow y_{1...m}^{(k)}} called the Batch Normalizing transform. The output of the BN transform y ( k ) = B N γ ( k ) , β ( k ) ( x ( k ) ) {\displaystyle y^{(k)}=BN_{\gamma ^{(k)},\beta ^{(k)}}(x^{(k)})} is then passed to other network layers, while the normalized output x ^ i ( k ) {\displaystyle {\hat {x}}_{i}^{(k)}} remains internal to the current layer. === Backpropagation === The described BN transform is a differentiable operation, and the gradient of the loss l {\displaystyle l} with respect to the different parameters can be computed directly with the chain rule. Specifically, ∂ l ∂ y i ( k ) {\displaystyle {\frac {\partial l}{\partial y_{i}^{(k)}}}} depends on the choice of activation function, and the gradient against other parameters could be expressed as a function of ∂ l ∂ y i ( k ) {\displaystyle {\frac {\partial l}{\partial y_{i}^{(k)}}}} : ∂ l ∂ x ^ i ( k ) = ∂ l ∂ y i ( k ) γ ( k ) {\displaystyle {\frac {\partial l}{\partial {\hat {x}}_{i}^{(k)}}}={\frac {\partial l}{\partial y_{i}^{(k)}}}\gamma ^{(k)}} , ∂ l ∂ γ ( k ) = ∑ i = 1 m ∂ l ∂ y i ( k ) x ^ i ( k ) {\displaystyle {\frac {\partial l}{\partial \gamma ^{(k)}}}=\sum _{i=1}^{m}{\frac {\partial l}{\partial y_{i}^{(k)}}}{\hat {x}}_{i}^{(k)}} , ∂ l ∂ β ( k ) = ∑ i = 1 m ∂ l ∂ y i ( k ) {\displaystyle {\frac {\partial l}{\partial \beta ^{(k)}}}=\sum _{i=1}^{m}{\frac {\partial l}{\partial y_{i}^{(k)}}}} , ∂ l ∂ σ B ( k ) 2 = ∑ i = 1 m ∂ l ∂ y i ( k ) ( x i ( k ) − μ B ( k ) ) ( − γ ( k ) 2 ( σ B ( k ) 2 + ϵ ) − 3 / 2 ) {\displaystyle {\frac {\partial l}{\partial \sigma _{B}^{(k)^{2}}}}=\sum _{i=1}^{m}{\frac {\partial l}{\partial y_{i}^{(k)}}}(x_{i}^{(k)}-\mu _{B}^{(k)})\left(-{\frac {\gamma ^{(k)}}{2}}(\sigma _{B}^{(k)^{2}}+\epsilon )^{-3/2}\right)} , ∂ l ∂ μ B ( k ) = ∑ i = 1 m ∂ l ∂ y i ( k ) − γ ( k ) σ B ( k ) 2 + ϵ + ∂ l ∂ σ B ( k ) 2 1 m ∑ i = 1 m ( − 2 ) ⋅ ( x i ( k ) − μ B ( k ) ) {\displaystyle {\frac {\partial l}{\partial \mu _{B}^{(k)}}}=\sum _{i=1}^{m}{\frac {\partial l}{\partial y_{i}^{(k)}}}{\frac {-\gamma ^{(k)}}{\sqrt {\sigma _{B}^{(k)^{2}}+\epsilon }}}+{\frac {\partial l}{\partial \sigma _{B}^{(k)^{2}}}}{\frac {1}{m}}\sum _{i=1}^{m}(-2)\cdot (x_{i}^{(k)}-\mu _{B}^{(k)})} , and ∂ l ∂ x i ( k ) = ∂ l ∂ x ^ i ( k ) 1 σ B ( k ) 2 + ϵ + ∂ l ∂ σ B ( k ) 2 2 ( x i ( k ) − μ B ( k ) ) m + ∂ l ∂ μ B ( k ) 1 m {\displaystyle {\frac {\partial l}{\partial x_{i}^{(k)}}}={\frac {\partial l}{\partial {\hat {x}}_{i}^{(k)}}}{\frac {1}{\sqrt {\sigma _{B}^{(k)^{2}}+\epsilon }}}+{\frac {\partial l}{\partial \sigma _{B}^{(k)^{2}}}}{\frac {2(x_{i}^{(k)}-\mu _{B}^{(k)})}{m}}+{\frac {\partial l}{\partial \mu _{B}^{(k)}}}{\frac {1}{m}}} . === Inference === During the training stage, the normalization steps depend on the mini-batches to ensure efficient and reliable training. However, in the inference stage, this dependence is not useful any more. Instead, the normalization step in this stage is computed with the population statistics such that the output could depend on the input in a deterministic manner. The population mean, E [ x ( k ) ] {\displaystyle E[x^{(k)}]} , and variance, Var [ x ( k ) ] {\displaystyle \operatorname {Var} [x^{(k)}]} , are computed as: E [ x ( k ) ] = E B [ μ B ( k ) ] {\displaystyle E[x^{(k)}]=E_{B}[\mu _{B}^{(k)}]} , and Var [ x ( k ) ] = m m − 1 E B [ ( σ B ( k ) ) 2 ] {\displaystyle \operatorname {Var} [x^{(k)}]={\frac {m}{m-1}}E_{B}[\left(\sigma _{B}^{(k)}\right)^{2}]} . The population statistics thus is a complete representation of the mini-batches. The BN transform in the inference step thus becomes y ( k ) = B N γ ( k ) , β ( k ) inf ( x ( k ) ) = γ ( k ) x ( k ) − E [ x ( k ) ] Var [ x ( k ) ] + ϵ + β