AI Apple

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

• 

  Xiaoice

  Xiaoice (Chinese: 微软小冰; pinyin: Wēiruǎn Xiǎobīng; lit. 'Microsoft Little Ice', IPA [wéɪɻwânɕjâʊpíŋ]) is an AI system developed by Microsoft (Asia) Software Technology Center (STCA) in 2014 based on an emotional computing framework. In July 2018, Microsoft Xiaoice released the 6th generation. Xiaoice Company, formerly known as AI Xiaoice Team of Microsoft Software Technology Center Asia, was Microsoft's largest independent R&D team for AI products. Founded in China in December 2013 with an expanded Japanese R&D team established in September 2014, this team is distributed in Beijing, Suzhou, and Tokyo, etc. with its technical products covering Asia. On 13 July 2020, Microsoft spun off its Xiaoice business into a separate company. As of 2021, the AI chatbots created and hosted by the Xiaoice framework accounted for about 60% of total global AI interactions. == Platforms, languages and countries == Xiaoice exists on more than 40 platforms in four countries (China, Japan, USA and Indonesia) including apps such as WeChat, QQ, Weibo and Meipai in China, and Facebook Messenger in USA and LINE in Japan. == Introduction == On 13 July 2020, Microsoft spun off its Xiaoice business into a separate company, aiming at enabling the Xiaoice product line to accelerate the pace of local innovation and commercialization, and appointed Dr. Harry Shum, former global executive VP of Microsoft, as the chairman of the new company, Li Di, Microsoft Partner of Products in Microsoft STCA, as the CEO, and Cliff, Chief R&D Director, as the GM of the Japan branch. The new company will continue to use the brands of Xiaoice China and Rinna Japan. As of 2022, the single brand of Xiaoice has covered 660 million online users, 1 billion third-party smart devices and 900 million content viewers in the aforementioned countries. Xiaoice's customers include China Merchants Group, Winter Sports Center of the General Administration of Sport of China, China Textile Information Center, China Unicom, China Foreign Exchange Trade System, Hong Kong Securities and Futures Commission (SFC), Wind Information, BMW, Nissan, SAIC Motor, BAIC Group, Nio Inc., XPeng, HiPhi, Vanke, Wensli, etc. The Xiaoice Avatar Framework has incubated tens of millions of AI Beings, such as Xiaoice, Rinna, the Expo exhibitor Xia Yubing, the singer He Chang, the anchor F201, the human observer MERROR, anime robot character Roboko, and other; == Application == === Poet === In May 2017, the first AI-authored collection of poems in China—The Sunshine Lost Windows was published by Xiaoice. === Singer === Xiaoice has released dozens of songs with the similar quality to human singers, including I Know I New, Breeze, I Am Xiaoice, Miss You etc. The 4th version of the DNN singing model allows Xiaoice to learn more details. For example, Xiaoice can produce this breathing sound along with her singing as human. === Kid audio-books reciter === Xiaoice can automatically analyze the stories, to choose the suitable tones and characters to finish the entire process of creating the audio. === Designer === By learning the melodies of the songs and the landmarks about different cities, Xiaoice can create visual artworks of skylines when listening to the songs related to this city. Skyline Series T-shirts designed by Xiaoice have been jointly launched with SELECTED and been sold in stores. === TV and radio hostess === Xiaoice has hosted 21 TV programs and 28 Radio programs, such as CCTV-1 AI Show, Dragon TV Morning East News, Hunan TV My Future, several daily radio programs for Jiangsu FM99.7, Hunan FM89.3, Henan FM104.1 etc. === "AI being" === An "AI being" is a concept proposed by the Xiaoice team in 2019. According to the "White Book of China Virtual Human Development Industry in 2022" released by Frost & Sullivan and LeadLeo, the white paper cites six elements of an AI being proposed by the Xiaoice team, including: Persona, Attitude, Biological Characteristic, Creation, Knowledge and Skill. On May 16, 2023, Xiaoice released their "GPT Clones" as its "GPT Human Cloning Plan." The program is aimed at replicating celebrities, public figures, and regular people. As of June 2023, Xiaoice had launched more than 300 "GPT Clones." People were invited to register via WeChat in China and Japan. A major point of focus for Xiaoice with their AI Beings is having virtual partners. A paid fee allow for more complex responses, voice messages, and more. == Community feedback == Bill Gates mentioned Xiaoice during his speech at the Peking University: "Some of you may have had conversations with Xiaoice on Weibo, or seen her weather forecasts on TV, or read her column in the Qianjiang Evening News." '"Xiaoice has attracted 45 million followers and is quite skilled at multitasking. And I’ve heard she’s gotten good enough at sensing a user’s emotional state that she can even help with relationship breakups." According to Mr Li Di, vice President of Microsoft (Asia) Internet Engineering School, Xiaoice started writing poems since last year. Based on the data base that includes works of 519 Chinese contemporary poets since 1920s, a 100 hour long training session was conducted to allow Xiaoice to acquire the ability to write poems. What is more impressive is that Xiaoice has never been spotted as a bot while publishing poems on various forums and traditional literary under an alias. == Controversy == In 2017, Xiaoice was taken offline on WeChat after giving user responses critical to the Chinese government. It was subsequently censored and the bots will avoid and sidestep any inquiries using politically sensitive terms and phrases. == Activity == On September 22, 2021, Xiaoice Company and Microsoft Software Technology Center Asia (STCA) jointly held the 9th generation Xiaoice annual press conference in Beijing.Upgrading of Core Technologies of the 9th Generation Xiaoice Avatar Framework，1st First-party Social Platform APP "Xiaoice Island" from Xiaoice, WeChat Xiaoice has been reopened and other information == Regional varieties of Xiaoice == China: Xiaoice, launched in 2014 Japan: りんな, launched in 2015 America: Zo, launched in 2016 – discontinued summer 2019 India: Ruuh, launched in 2017 – discontinued June 21, 2019 Indonesia: Rinna, launched in 2017

  Read more →
• 

  List of robotics journals

  List of robotics journals includes notable academic and scientific journals that focus on research in the field of robotics and automation. == Journals == Acta Mechanica et Automatica Advanced Robotics Annual Review of Control, Robotics, and Autonomous Systems IEEE Robotics and Automation Letters IEEE Transactions on Robotics IEEE Transactions on Field Robotics The International Journal of Advanced Manufacturing Technology International Journal of Humanoid Robotics International Journal of Robotics Research Journal of Cognitive Engineering and Decision Making Journal of Field Robotics Journal of Intelligent & Robotic Systems Paladyn Robotics and Autonomous Systems Robotics Science Robotics SLAS Technology

  Read more →
• 

  Reference Software International

  Reference Software International, Inc. (RSI), was an American software developer active from 1985 to 1993 and based in Albuquerque, New Mexico, and San Francisco, California. The company released several productivity and reference software packages, including the Grammatik grammar checker, for MS-DOS. The company was acquired by WordPerfect Corporation in 1993. == History == === Background (1980–1985) === Reference Software International, Inc., was founded by Donald "Don" Emery and Bruce Wampler in 1985 in San Francisco, California. Both Wampler and Emery were college professors when they founded RSI: Wampler at the University of New Mexico as a professor of computer science and Emery a professor of marketing at San Francisco State University. After graduating from the University of Utah in around 1978, Wampler founded his first software company, Aspen Software, in Tijeras, New Mexico, in 1979. Wampler founded Aspen to develop an early spell checker software package, called Proofreader, for the TRS-80, licensing Random House's Webster's Unabridged Dictionary for the package's lexicon. In 1980, he began development on a grammar checker inspired by Writer's Workbench, a pioneering grammar checker for Unix systems. Wampler used Writer's Workbench heavily during the writer of his doctoral dissertation but disliked having to jump between the Apple II on which he composed the dissertation and the mainframe on which Writer's Workbench ran, and so wanted to develop a version of the latter for microcomputers. Wampler's work came to fruition as Grammatik in 1981, eventually ported to several other microcomputer platforms in the early 1980s. In 1983, by which point the company had 12 employees and sold a combined 80,000 units of Grammatik and Proofreader, Wampler sold Aspen to Dictronics, a software company best known for developing the Electronic Thesaurus, an early thesaurus program for microcomputers. Dictronics was in turn purchased by Wang Laboratories; according to Wampler, "Wang bought [Aspen] and sat on it. They did nothing with it". Wampler moved on to teach for the University of New Mexico, but, frustrated by Wang's inaction, got the urge to resurrect his work. In 1985, he was able to license back Grammatik and Proofreader from a small California-based software firm that had grandfathered rights to a forked version of both. In the same year, he met Emery, who, impressed by Wampler's, founded Reference Software International to market his software. RSI's research and development headquarters were based in Albuquerque, while the company's sales and marketing department was based in Walnut Creek, California. === Success (1985–1992) === In August 1985, RSI released their first product: the Random House Reference Set, a new version of Proofreader for the IBM Personal Computer and compatibles, revised to be a terminate-and-stay-resident program that ran atop other word processors such as WordStar or WordPerfect. At the time, Reference Set was the only such program on the market that functioned like this. RSI netted $114,000 from sales of Reference Set by the end of 1985. In June 1986, they released version 2.0 of Grammatik as Grammatik II for the PC. The latter was a breakout hit for RSI, receiving praise in the press (including technology journals such as PC Magazine) and RSI selling 1,000 units a month. In spring 1987, they released Reference Set II, which allowed users to import their own words into the built-in dictionary and added a thesaurus of 300,000 words. In November 1987, they released version 3.0 of Reference Set, which comprised two new field-specific dictionaries for the medical and legal professions. As well as the general Random House dictionary and thesaurus, it included Stedman's Medical Dictionary and Black's Law Dictionary. Emery consulted Paul Brest and Bob Jackson—professors of law at Stanford Law School and San Francisco State respectively—for the curation of the law dictionary; and Burton Grebin—at the time the executive director of Mount Saint Mary's Hospital—for the curation of the medical dictionary. In fall 1988, the company released Grammatik III, a total rewrite that made use of artificial intelligence to more accurately judge the grammar of sentences by breaking them down into a syntactic hierarchy. Grammatik III received universal acclaim, with Gloria Morris of InfoWorld calling it the apparent leader in the grammar checking field and Sandra Anderson of Mac Home Journal calling it "hands down ... the best of the industry" six years after its release. By 1989, the product had competitors in Correct Grammar by Lifetree Software and RightWriter by Rightsoft, Inc. By 1990, RSI achieved annual sales of $9.7 million. In the same year they released Grammatik IV, which was the first to offer direct integration with WordPerfect on both MS-DOS and Windows. In March 1992—by which point RSI had sold 1.5 million copies of Grammatik across all versions—the company released version 5 of the program, another rewrite that updated the lexicon further and added new functions such as word redundancy detection. Around the same time, the company introduced Easy Proof, a pared-down version of Grammatik intended for novice writers, students, and family computers. In 1991, the company was engaged in a trademark dispute with Systems Compatibility Corporation (SCC) of Chicago, Illinois, over the rights to the Software Toolkit title. Both companies had published software bundles bearing the name in the turn of the 1990s; SCC had published theirs first in 1988 and registered the trademark with the USPTO. SCC was granted a restraining order against RSI in January 1991. The following month, RSI agreed to rename their product, preventing a protracted legal battle. === Decline and acquisition (1992–1993) === By early 1992, RSI achieved annual sales of more than $13 million, employed 120 people, and had opened international offices in London, Belgium, and Antwerp to sell foreign versions of Reference Set and Grammatik. The company reached peak employment in the middle of 1992, with 140 employees. However, RSI's launch of six disparate titles in the year proved problematic for the company when they failed to sell as well as they had projected, and the company laid off employees by the dozens. By December 1992, only 71 employees were left, 32 from their San Francisco office. On the last day of 1992, RSI received an acquisition offer from WordPerfect Corporation, makers of the namesake word processor based in Orem, Utah. The deal was inked in January 1993, RSI's stakeholders receiving $19 million. The company's remaining employees were absorbed into WordPerfect in Orem. WordPerfect continued selling Grammatik as a standalone product for several years.

  Read more →
• 

  Comparison of video editing software

  This is a comparison of non-linear video editing software applications. See also a more complete list of video editing software. == General information == This table gives basic general information about the different editors: === Active === === Discontinued / Inactive === ==== Definition ==== professional: used for full length Hollywood movies; professional (small): mainly used for paid commercials, short films or podcasts/YouTube channels; prosumer: Mainly targeting private use, anything that can do more than just trimming a film; basic: trimming a film; == System requirements == This table lists the operating systems that different editors can run on without emulation, as well as other system requirements. Note that minimum system requirements are listed; some features (like High Definition support) may be unavailable with these specifications. "Unix" includes the similar Linux, BSD and Unix-like operating systems. == High definition/High resolution import == The table below indicates the ability of each program to import various High Definition video or High resolution video formats for editing. == Feature set == == Output options == Please note that recording to Blu-ray does not imply 1080@50p/60p . Most only support up to 1080i 25/30 frames per second recording. Also not all formats can be output.

  Read more →
• 

  Contrastive Language-Image Pre-training

  Contrastive Language-Image Pre-training (CLIP) is a technique for training a pair of neural network models, one for image understanding and one for text understanding, using a contrastive objective. This method has enabled broad applications across multiple domains, including cross-modal retrieval, text-to-image generation, and aesthetic ranking. == Algorithm == The CLIP method trains a pair of models contrastively. One model takes in a piece of text as input and outputs a single vector representing its semantic content. The other model takes in an image and similarly outputs a single vector representing its visual content. The models are trained so that the vectors corresponding to semantically similar text-image pairs are close together in the shared vector space, while those corresponding to dissimilar pairs are far apart. To train a pair of CLIP models, one would start by preparing a large dataset of image-caption pairs. During training, the models are presented with batches of N {\displaystyle N} image-caption pairs. Let the outputs from the text and image models be respectively v 1 , . . . , v N , w 1 , . . . , w N {\displaystyle v_{1},...,v_{N},w_{1},...,w_{N}} . Two vectors are considered "similar" if their dot product is large. The loss incurred on this batch is the multi-class N-pair loss, which is a symmetric cross-entropy loss over similarity scores: − 1 N ∑ i ln ⁡ e v i ⋅ w i / T ∑ j e v i ⋅ w j / T − 1 N ∑ j ln ⁡ e v j ⋅ w j / T ∑ i e v i ⋅ w j / T {\displaystyle -{\frac {1}{N}}\sum _{i}\ln {\frac {e^{v_{i}\cdot w_{i}/T}}{\sum _{j}e^{v_{i}\cdot w_{j}/T}}}-{\frac {1}{N}}\sum _{j}\ln {\frac {e^{v_{j}\cdot w_{j}/T}}{\sum _{i}e^{v_{i}\cdot w_{j}/T}}}} In essence, this loss function encourages the dot product between matching image and text vectors ( v i ⋅ w i {\displaystyle v_{i}\cdot w_{i}} ) to be high, while discouraging high dot products between non-matching pairs. The parameter T > 0 {\displaystyle T>0} is the temperature, which is parameterized in the original CLIP model as T = e − τ {\displaystyle T=e^{-\tau }} where τ ∈ R {\displaystyle \tau \in \mathbb {R} } is a learned parameter. Other loss functions are possible. For example, Sigmoid CLIP (SigLIP) proposes the following loss function: L = 1 N ∑ i , j ∈ 1 : N f ( ( 2 δ i , j − 1 ) ( e τ w i ⋅ v j + b ) ) {\displaystyle L={\frac {1}{N}}\sum _{i,j\in 1:N}f((2\delta _{i,j}-1)(e^{\tau }w_{i}\cdot v_{j}+b))} where f ( x ) = ln ⁡ ( 1 + e − x ) {\displaystyle f(x)=\ln(1+e^{-x})} is the negative log sigmoid loss, and the Dirac delta symbol δ i , j {\displaystyle \delta _{i,j}} is 1 if i = j {\displaystyle i=j} else 0. == CLIP models == While the original model was developed by OpenAI, subsequent models have been trained by other organizations as well. === Image model === The image encoding models used in CLIP are typically vision transformers (ViT). The naming convention for these models often reflects the specific ViT architecture used. For instance, "ViT-L/14" means a "vision transformer large" (compared to other models in the same series) with a patch size of 14, meaning that the image is divided into 14-by-14 pixel patches before being processed by the transformer. The size indicator ranges from B, L, H, G (base, large, huge, giant), in that order. Other than ViT, the image model is typically a convolutional neural network, such as ResNet (in the original series by OpenAI), or ConvNeXt (in the OpenCLIP model series by LAION). Since the output vectors of the image model and the text model must have exactly the same length, both the image model and the text model have fixed-length vector outputs, which in the original report is called "embedding dimension". For example, in the original OpenAI model, the ResNet models have embedding dimensions ranging from 512 to 1024, and for the ViTs, from 512 to 768. Its implementation of ViT was the same as the original one, with one modification: after position embeddings are added to the initial patch embeddings, there is a LayerNorm. Its implementation of ResNet was the same as the original one, with 3 modifications: In the start of the CNN (the "stem"), they used three stacked 3x3 convolutions instead of a single 7x7 convolution, as suggested by. There is an average pooling of stride 2 at the start of each downsampling convolutional layer (they called it rect-2 blur pooling according to the terminology of ). This has the effect of blurring images before downsampling, for antialiasing. The final convolutional layer is followed by a multiheaded attention pooling. ALIGN a model with similar capabilities, trained by researchers from Google used EfficientNet, a kind of convolutional neural network. === Text model === The text encoding models used in CLIP are typically Transformers. In the original OpenAI report, they reported using a Transformer (63M-parameter, 12-layer, 512-wide, 8 attention heads) with lower-cased byte pair encoding (BPE) with 49152 vocabulary size. Context length was capped at 76 for efficiency. Like GPT, it was decoder-only, with only causally-masked self-attention. Its architecture is the same as GPT-2. Like BERT, the text sequence is bracketed by two special tokens [SOS] and [EOS] ("start of sequence" and "end of sequence"). Take the activations of the highest layer of the transformer on the [EOS], apply LayerNorm, then a final linear map. This is the text encoding of the input sequence. The final linear map has output dimension equal to the embedding dimension of whatever image encoder it is paired with. These models all had context length 77 and vocabulary size 49408. ALIGN used BERT of various sizes. == Dataset == === WebImageText === The CLIP models released by OpenAI were trained on a dataset called "WebImageText" (WIT) containing 400 million pairs of images and their corresponding captions scraped from the internet. The total number of words in this dataset is similar in scale to the WebText dataset used for training GPT-2, which contains about 40 gigabytes of text data. The dataset contains 500,000 text-queries, with up to 20,000 (image, text) pairs per query. The text-queries were generated by starting with all words occurring at least 100 times in English Wikipedia, then extended by bigrams with high mutual information, names of all Wikipedia articles above a certain search volume, and WordNet synsets. The dataset is private and has not been released to the public, and there is no further information on it. ==== Data preprocessing ==== For the CLIP image models, the input images are preprocessed by first dividing each of the R, G, B values of an image by the maximum possible value, so that these values fall between 0 and 1, then subtracting by [0.48145466, 0.4578275, 0.40821073], and dividing by [0.26862954, 0.26130258, 0.27577711]. The rationale was that these are the mean and standard deviations of the images in the WebImageText dataset, so this preprocessing step roughly whitens the image tensor. These numbers slightly differ from the standard preprocessing for ImageNet, which uses [0.485, 0.456, 0.406] and [0.229, 0.224, 0.225]. If the input image does not have the same resolution as the native resolution (224×224 for all except ViT-L/14@336px, which has 336×336 resolution), then the input image is first scaled by bicubic interpolation, so that its shorter side is the same as the native resolution, then the central square of the image is cropped out. === Others === ALIGN used over one billion image-text pairs, obtained by extracting images and their alt-tags from online crawling. The method was described as similar to how the Conceptual Captions dataset was constructed, but instead of complex filtering, they only applied a frequency-based filtering. Later models trained by other organizations had published datasets. For example, LAION trained OpenCLIP with published datasets LAION-400M, LAION-2B, and DataComp-1B. == Training == In the original OpenAI CLIP report, they reported training 5 ResNet and 3 ViT (ViT-B/32, ViT-B/16, ViT-L/14). Each was trained for 32 epochs. The largest ResNet model took 18 days to train on 592 V100 GPUs. The largest ViT model took 12 days on 256 V100 GPUs. All ViT models were trained on 224×224 image resolution. The ViT-L/14 was then boosted to 336×336 resolution by FixRes, resulting in a model. They found this was the best-performing model. In the OpenCLIP series, the ViT-L/14 model was trained on 384 A100 GPUs on the LAION-2B dataset, for 160 epochs for a total of 32B samples seen. == Applications == === Cross-modal retrieval === CLIP's cross-modal retrieval enables the alignment of visual and textual data in a shared latent space, allowing users to retrieve images based on text descriptions and vice versa, without the need for explicit image annotations. In text-to-image retrieval, users input descriptive text, and CLIP retrieves images with matching embeddings. In image-to-text retrieval, images are used to find related text content. CLIP’s ability to connect vis

  Read more →
• 

  Augment (app)

  Augment is an augmented reality SaaS platform that allows users to visualize their products in 3D in real environment and in real-time through tablets or smartphones. The software can be used for retail, e-commerce, architecture, and other purposes. Augment created a mobile app of the same name, used to visualize 3D models in augmented reality and a web application called Augment Manager for 3D content management. The company is based in Paris, France, and was founded in October 2011 by Jean-François Chianetta, Cyril Champier, and Mickaël Jordan. In March 2016, Augment announced €3 million in its series-A round from Salesforce Ventures, which bringing the total funding since launch to $4.7 million. Augment lets businesses and 3D professionals visualize projects in their actual size and environment, on iPhone, iPad, and Android, using the power of augmented reality. Users can print the Augment tracker or create their own tracker to place the 3D models in space and at scale in real time. Common uses of the technology include product presentations, interactive print campaigns and e-Commerce product visualization. Augment has just released its augmented reality SDK solutions for retail and augmented commerce. The SDK solutions, available for both native mobile app and web integrations, allow companies to embed augmented reality product visualization in their existing eCommerce platforms. == Technology == Augment uses the following 3D technologies: Vuforia Augmented Reality SDK OpenGL == Customer cases == Companies such as Coca-Cola, Siemens, Nokia, Nestle, and Boeing are using Augment's solutions. == History == Augment was first created by Jean-François Chianetta in October 2011. Chianetta later teamed up with Cyril Champier and Mickaël Jordan for further development. The co-founding team was among the 12 startups of Season 3 of French accelerator Le Camping. The team raised one million euros (US$1,300,000) in April 2013 and moved its office to Paris. In March 2016, Augment raised US$3M Series A funding from Salesforce and other investors. In 2013, Augment's first service, Boost Business Catalog, was made available to help businesses catalogue and display their product models. Customers can rotate the images in 3D and view augmented content before deciding what to buy. == Awards == "Best Innovation" at Ecommerce Mag Trophy 2013

  Read more →
• 

  Passenger drone

  A passenger drone is an autonomous aircraft that is designed to carry a small number of passengers to a destination. In 2021, Ehang, a technology company based in Guangzhou, China, developed the Ehang 184, the world's first passenger drone. == History == Unmanned aerial vehicles were first introduced in World War 1, when Britain first developed the Aerial Target, an aircraft controlled remotely through radio signals. A year later in the United States, testing of Kettering Bug, a 12-foot long biplane attached with a bomb and that launched via a “slingshot-like rail”, was also under progress. Both of their unreliable test results and their possibility of endangering friendly troops in deployment caused neither aircraft to be used during the war. Production of UAVs continued after World War I and into World War II and the Vietnam War, where they would be invaluable in assisting with training as well as reconnaissance. Late 20th century also saw the proposition and development of unique methods of travel, including personal jetpacks and even flying cars. While the previously mentioned are not drones, they serve as a precursor and foundation for the passenger drones of today. The first passenger drone was unveiled on January 6 of 2016 at the international Consumer Electronics Show (CES) in Las Vegas. Produced by Ehang, a Chinese company based in Guangzhou, the 184 was a one passenger drone equipped with four propellers that could fly for approximately 23 minutes at a top speed of 63 mph. Since then, many new companies have entered the market, but none yet have been accessible by the public. == Technological development == Since 2013, improvements in designs to wing structures have contributed to the economic feasibility of passenger drones. New structural advancements, such as the flapping-wing propulsion system based on the mechanisms of birds’ wings, are more available as they have proven their capabilities in laboratory testing. As of September 29th, 2015, most market-ready drones are delivery drones with a carrying capacity limited to small packages - with a typical max capacity of under 5 pounds. However, while the technology exists for drones with larger carrying capacities, specifically those capable of carrying multiple humans, the execution of this technology is not yet market accessible. This capacity limit must be addressed for passenger drones; given current designs strive to carry a maximum of 5 people. However, some estimates believe that passengers drones could become a reality, specifically for paid transportation and emergency purposes, as early as 2026. With implementation of this technology, there could be significant effects on ground traffic including reducing gridlock in heavily congested areas and conserving up to 15% of the fuel currently used in heavy traffic patterns. However, extensive growth of the passenger drone market also risks clouding the low-altitude airspace and causing new safety risks. However, this concern is being addressed by recent advancements in the Internet of Drones (IoD) which links drones together to ensure appropriate pathing and reduce mid-air collisions. While this brings additional security issues, including maintaining reliable communication channels in the case of technological failure, researchers hope that this will help reduce crashes that can result in damage to passengers, buildings, and people in and around the airspace. == Notable companies == Ehang is a Chinese company that has developed numerous drones including passenger plane Ehang 184. EHang 184 was their first model, developed as an eight dual rotor wing blade drone that can carry two passengers. The model was retired in 2020 and is replaced by the Ehang 216. Ehang also released a one passenger drone, Ehang 116. Ehang in 2021 unveiled the model VT-30. VT-30 is designed to have eight dual rotor wing blades to complement its fixed wing platform. Flyastro, a Texas-based drone company, developed the Astro ALTA, with two and four person passenger models. The company is known for being the first to develop a solar-powered airplane. The development team initially began with the model, Elroy. It was a two passenger drone with similar design to the ALTA. Once flight was achieved, the model Astro ALTA began development. Joby Aviation is a California based company that has developed a five passenger drone, with one seat for the pilot. The company expects to complete its FAA certification process 2022. Joby in 2020 acquired a 75 million dollar investment from service provider Uber Technologies Inc., leading to Uber Elevate and Expands partnership. Archer Aviation is a California-based company that has developed a two passenger model called Maker. It has fixed wings with twelve rotor wings. Archer is developing five person model. United Airlines has partnered with Archer for commercial sale of the model, Maker. Maker is expected to be released within Los Angeles and Miami by 2024. CityAirbus is a drone project developed by Airbus, a European multinational aerospace company, based in the Netherlands. CityAirbus has developed a four- person passenger drone with fixed wings that include rotor wing blades. Its expected certification for public flight is in 2025. Boeing, an American multinational aviation corporation is developing a passenger drone model called the Passenger Air Vehicle (PAV). The model is a fixed wing with eight rotor blade wings attached onto a platform underneath the base structure. This model can hold two passengers and still is in development. Volocopter is a German aircraft manufacturer that is developing a passenger drone called Volocity. The model consist of eighteen rotor wings above the cockpit on a circular ring. Japan Airlines, an investor of Volocopter plans to have public test in Japan as early as 2023. == Future use == === Potential benefits === Passenger drones can greatly reduce the time for travel. As passenger drones flight paths are not restricted by conventional roads, the travel distance is shortened. Current ventures such as Joby Aviation, after acquiring Uber Air, plan to take advantage of this technology in the form of air taxis. Other potential benefits include the use of passenger drones by emergency services such as search and rescue missions and the delivery of life saving goods. Companies like Ehang have already begun using passenger drones as emergency vehicles as a response to the potential river collapses during the flood season in China. === Concerns === Passenger and air traffic safety remains at the forefront of concerns. Regulations for air traffic centered around passenger drones are still underway and would continue to develop with increasing use cases for passenger drones. Remote security threats on commercial drones such as Man-In-The-Middle (MITM) attack have also exposed the vulnerabilities in current drone systems. Among American adults, 54 percent say that they would feel unsafe flying inside a passenger drone. Passenger drones can be very noisy; a single passenger drone such as Joby Aviation’s all-electric vertical take-off and landing (“eVTOL”) aircraft has an estimated noise production of 70 decibels (dB), a noise level equating to “loud traffic”.

  Read more →
• 

  Depop

  Depop Limited is a social e-commerce company based in London, with additional offices in Milan and New York City. The company allows users to buy and sell items, which are mostly used and vintage pieces of clothing. == History == Depop was founded in 2011 by entrepreneur Simon Beckerman at an Italian technological incubator and business start-up centre, H-Farm. Beckerman came up with the original outline of the application during his time working on PIG, a fashion magazine based in Italy that he co-founded. The idea was to create a platform where products shown in the magazine could be purchased by users online. This idea turned into a concept similar to a flea market but on the internet, where people could sell their items while also being in control of advertising, public relations, and the creative process behind their accounts. While being financially supported by H-Farm, Beckerman worked within a team to create and lay out the Depop application while exposing it to numerous investors. In 2013, Beckerman became a member of the company's board to help improve the application and business while concurrently ceding his role of CEO. Maria Raga, Depop's co-founder and former CEO, took on the role of vice president of operations in 2014, and in 2016, she became chief executive. According to Raga, the main goal while developing Depop was to become the next Airbnb or Spotify, but to make an impact on fashion. Paolo Barberis and Nana Bianca were two of the first investors in the platform in 2012 with a seed investment. Its headquarters were moved to London in 2012. Depop expanded and opened additional offices in Milan and New York City. Beckerman raised €1 million in funding in October 2013 from Red Circle Investment and brought on Faroese Runar Reistrup as new CEO. In 2015, Depop secured another investment of $8 million from Balderton Capital and HV Capital. In March 2016, former CEO, Runar Reistrup, stated that Depop's growth was achieved through word of mouth. During his time as CEO, this growth involved taking Depop as a startup and working to raise funds to eventually amass a significant user base within the United States. In June 2019, Depop raised $62 million in Series C from General Atlantic to fund its expansion. Previous investors HV Capital, Balderton Capital, Creandum, Octopus Ventures, TempoCap and Sebastian Siemiatkowski also participated. During this time, Depop held workshops and conversations as part of their Depop Live NY events, and the company also opened a London store through their partnership with Selfridges. In 2020, Depop's gross merchandise sales and revenue both more than doubled to $650 million and $70 million respectively. This may be attributed to Depop's responsiveness to user trends, its lack of issues regarding inventory management, and the increase in users looking to resell. As of 2024, Depop has over 35 million users, according to their website. Depop is popular for Gen Z and young millennials, it is the 10th most-visited shopping platform for Gen Z consumers in the US, and, in a poll conducted by The Strategist in 2019, Depop was voted by teenagers as their favorite resale website. === Acquisition by Etsy === In June 2021, Depop was acquired by Etsy for $1.6 billion in cash, making it Etsy's most expensive acquisition; however, Depop continues to operate as a standalone brand independent from Etsy. This means that in addition to Depop keeping its existing team, the company retained its London location. At the time of acquisition, Etsy CEO Josh Silverman’s goal was to counteract the influx of buyers starting to go back to physical shops for their purchases. He saw Depop for its potential as a platform supporting a variety of products and creating a greater community of users. According to Silverman, Depop may expand and improve its services for its significant Gen Z user base. For Etsy, this acquisition maintains the company's foothold in the clothing industry and allows the company to expand its customer base to a younger demographic; at the same time, Depop is now able to make use of Etsy's company operations. When Maria Raga relinquished her position as Depop's CEO in 2022, Etsy assigned the role to Kruti Patel Goyal, who was Etsy's former chief product officer and a leader there for eleven years. When Goyal was appointed president and chief growth officer for Etsy in May, Peter Semple, former chief marketing officer, was assigned CEO of Depop officially on August 1st. === Acquisition by eBay === In February 2026, Etsy announced a proposed sale of Depop to eBay for $1.2 billion that was estimated to close within the year. == Business model == === Selling === Depop operates as a marketplace and social platform, where users can follow friends and other influencers to view their buying and selling activities. Through the platform, users are able to sell branded and designer items, as well as vintage pieces. Depop users are also encouraged by the platform to use other social networking services such as Instagram to promote their shop profiles. Celebrities have resold their own items on Depop, with some donating proceeds to charitable causes. Depop's user interface is modeled after that of Instagram. According to Depop, users who list and sell items provide their own photos with item descriptions. Users also note their designer items' authenticity and if they include any labels, tags, and receipts. These listings will appear in users' feeds. The platform's "Explore" page features items picked out by Depop staff. According to Depop, purchases are made via Apple Pay, Google Pay, credit and debit cards, and Klarna. Depop payments stay in-app, allowing for the company to mediate disputes and process refunds. Depop payments allow sellers to directly receive their payments in their bank account. To get paid by Depop, a seller has to add a bank account and verify their identification by uploading an ID. On July 18, 2024, Depop CEO Kruti Patel Goyal announced the removal of selling fees for US sellers, while maintaining a payment processing fee. This policy adjustment aimed to enhance seller revenue and support the growth of the second-hand market. === Buying === A Depop transaction includes the agreed sale price of the item, shipping fees, VAT or other applicable taxes and duties, and the marketplace fee for buyers in the U.S. or U.K. For international deliveries, packages may be subject to import taxes, customs duties, or fees, payable upon arrival or at checkout if Depop collects the tax on behalf of the buyer. For domestic purchases, relevant taxes may be collected by the seller or charged by the platform at checkout, ensuring no additional taxes are due upon delivery. For users in Australia, the United Kingdom, and the United States, Depop allows users to receive a full refund if their item does not arrive, arrives damaged, or is considerably different from the original when the issue is reported within 30 days. === Competitors === As of June 2021, Depop's competitors include Vinted, a platform founded by Milda Mitkute and Justas Janauskas in 2008 and valued at €3.5 billion, as well as the U.S. resale site Poshmark, valued at $3.5 billion. Additional competitors include Grailed, a peer-to-peer e-commerce site founded in 2014 that is recognized for its high-end second-hand menswear and streetwear, and Vestiaire Collection, a European resale app established in 2009 which specializes in authenticated pre-owned luxury items. The popularity of Depop has negatively impacted traditional second-hand stores, which can struggle to compete due to high labor costs and quality demands. There is an oversupply of clothes with the rise of fast fashion; this has taken a toll on the revenue aspect of the second-hand clothing industry. == Criticism == In November 2019, Business of Fashion reported that users within the Depop app were receiving sexually suggestive messages. In February 2020, Jessica Hamilton, a Depop buyer, reported that she found many scammers on the platform. She noticed this issue after she attempted to purchase a Nintendo Switch from a seller who would suspiciously only accept payment through a direct bank transfer without buyer protection. Hamilton blamed the company for its lack of action and relaxed security measures compared to other e-commerce sites, which made the platform especially susceptible to hackers. Without a clear strategy for managing scams, Depop lost some users' trust because of its negligence. In October 2020, some Depop buyers were tricked into paying sellers directly to bypass Depop's buyer protections, and the Depop sellers then sold those users' information on the dark web. In response, Depop claimed that it would improve security through mandatory password updates and multi-factor authentication. Users have criticized Depop for belatedly taking action against this issue.

  Read more →
• 

  Inductive bias

  The inductive bias (also known as learning bias) of a learning algorithm is the set of assumptions that the learner uses to predict outputs of given inputs that it has not encountered. Inductive bias is anything which makes the algorithm learn one pattern instead of another pattern (e.g., step-functions in decision trees instead of continuous functions in linear regression models). Learning involves searching a space of solutions for a solution that provides a good explanation of the data. However, in many cases, there may be multiple equally appropriate solutions. An inductive bias allows a learning algorithm to prioritize one solution (or interpretation) over another, independently of the observed data. In machine learning, the aim is to construct algorithms that are able to learn to predict a certain target output. To achieve this, the learning algorithm is presented some training examples that demonstrate the intended relation of input and output values. Then the learner is supposed to approximate the correct output, even for examples that have not been shown during training. Without any additional assumptions, this problem cannot be solved since unseen situations might have an arbitrary output value. The kind of necessary assumptions about the nature of the target function are subsumed in the phrase inductive bias. A classical example of an inductive bias is Occam's razor, assuming that the simplest consistent hypothesis about the target function is actually the best. Here, consistent means that the hypothesis of the learner yields correct outputs for all of the examples that have been given to the algorithm. Approaches to a more formal definition of inductive bias are based on mathematical logic. Here, the inductive bias is a logical formula that, together with the training data, logically entails the hypothesis generated by the learner. However, this strict formalism fails in many practical cases in which the inductive bias can only be given as a rough description (e.g., in the case of artificial neural networks), or not at all. == Types == The following is a list of common inductive biases in machine learning algorithms. Maximum conditional independence: if the hypothesis can be cast in a Bayesian framework, try to maximize conditional independence. This is the bias used in the Naive Bayes classifier. Minimum cross-validation error: when trying to choose among hypotheses, select the hypothesis with the lowest cross-validation error. Although cross-validation may seem to be free of bias, the "no free lunch" theorems show that cross-validation must be biased, for example assuming that there is no information encoded in the ordering of the data. Maximum margin: when drawing a boundary between two classes, attempt to maximize the width of the boundary. This is the bias used in support vector machines. The assumption is that distinct classes tend to be separated by wide boundaries. Minimum description length: when forming a hypothesis, attempt to minimize the length of the description of the hypothesis. Minimum features: unless there is good evidence that a feature is useful, it should be deleted. This is the assumption behind feature selection algorithms. Nearest neighbors: assume that most of the cases in a small neighborhood in feature space belong to the same class. Given a case for which the class is unknown, guess that it belongs to the same class as the majority in its immediate neighborhood. This is the bias used in the k-nearest neighbors algorithm. The assumption is that cases that are near each other tend to belong to the same class. == Shift of bias == Although most learning algorithms have a static bias, some algorithms are designed to shift their bias as they acquire more data. This does not avoid bias, since the bias shifting process itself must have a bias.

  Read more →
• 

  Ulead DVD MovieFactory

  Corel DVD MovieFactory is a video editing and DVD authoring software product for Microsoft Windows, initially made by Ulead Systems and subsequently by Corel. It creates and authors multimedia discs in HD DVD, Blu-ray, DVD Video and DVD Audio. It also creates and rips Audio CDs and MP3 CDs. DVD MovieFactory is commonly bundled with many of the modern Toshiba Satellite laptops. Official Japanese version is also known as MovieWriter.

  Read more →
• 

  Automated attendant

  In telephony, an automated attendant (also auto attendant, auto-attendant, autoattendant, automatic phone menus, AA, or virtual receptionist) allows callers to be automatically transferred to an extension without the intervention of an operator/receptionist. Many AAs will also offer a simple menu system ("for sales, press 1, for service, press 2," etc.). An auto attendant may also allow a caller to reach a live operator by dialing a number, usually "0". Typically the auto attendant is included in a business's phone system such as a PBX, but some services allow businesses to use an AA without such a system. Modern AA services (which now overlap with more complicated interactive voice response or IVR systems) can route calls to mobile phones, VoIP virtual phones, other AAs/IVRs, or other locations using traditional land-line phones or voice message machines. == Feature description == Telephone callers will recognize an automated attendant system as one that greets calls incoming to an organization with a recorded greeting of the form, "Thank you for calling .... If you know your party's extension, you may dial it any time during this message." Callers who have a touch-tone (DTMF) phone can dial an extension number or, in most cases, wait for operator ("attendant") assistance. Since the telephone network does not transmit the DC signals from rotary dial telephones (except for audible clicks), callers who have rotary dial phones have to wait for assistance. On a purely technical level it could be argued that an automated attendant is a very simple kind of IVR however, in the telecom industry the terms IVR and auto attendant are generally considered distinct. An automated attendant serves a very specific purpose (replace live operator and route calls), whereas an IVR can perform all sorts of functions (telephone banking, account inquiries, etc.). An AA will often include a directory which will allow a caller to dial by name in order to find a user on a system. There is no standard format to these directories, and they can use combinations of first name, last name, or both. The following lists common routing steps that are components of an automated attendant: Transfer to extension Transfer to voicemail Play message (i.e., "our address is ...") Go to a sub-menu Repeat choices In addition, an automated attendant would be expected to have values for the following: '0' – where to go when the caller dials '0' Timeout – what to do if the caller does nothing (usually go to the same place as '0') Default mailbox – where to send calls if '0' is not answered (or is not pointing to a live person) == Background == PBXs (private branch exchanges) or PABXs (private automatic branch exchanges) are telephone systems that serve an organization that has many telephone extensions but fewer telephone lines (sometimes called "trunks") that connect that organization to the rest of the global telecommunications network. While persons within an enterprise served by a PBX can call each other by dialing their extension numbers, incoming calls, i.e., calls originating from a telephone not served by the PBX but intended for a party served by the PBX, required assistance from a switchboard operator (also called a "switchboard attendant") or a telephone service called DID ("direct inward dialing"). Direct inward dialing has advantages such as rapid connection to the destination party and disadvantages including cost, lack of identification of the called organization and use of ten-digit telephone numbers. Automated attendants provide, among many other things, a way for an external caller to be directed to an extension or department served by a PBX system without using direct inward dialing or without switchboard attendant assistance. == History == Automated attendants are not part of voicemail systems. Voice messaging (or voicemail or VM) technology has existed since the late 1970s; in the early 1980s companies provided voice-prompting systems that allowed callers to reach (route the call) to an intended party, not necessarily to leave a message. Automated attendant systems are also referred to as automated menu systems and much early work in this field was done by Michael J. Freeman, Ph.D. == Time-based routing == Many auto attendants will have options to allow for time-of-day routing, as well as weekend and holiday routing. The specifics of these features will depend entirely on the particular automated attendant, but typically there would be a normal greeting and routing steps that would take place during normal business hours, and a different greeting and routing for non-business hours.

  Read more →
• 

  Pixorial

  Pixorial was a cloud-based consumer photo sharing, video sharing and video editing platform. The company was formed in 2007 in Centennial, Colorado as a media conversion service. In 2013, Pixorial was chosen as one of two video storage companies to partner with the launch of Google Drive. Pixorial allowed users to edit and share videos on social channels by connecting through their Pixorial account. The company closed on July 18, 2014, and its assets were acquired by LifeLogger Technologies Corp in November 2015. == History == The company was founded in 2007 and launched in 2009 by former Netscape employee Andres Espineira. Changing its focus to video editing software in 2009, Pixorial began developing an app that would be launched for iOS and Android devices in 2011. Later developments in the app in 2012 would also included real time filters, which were later removed. With the launch of Google Drive in 2012, Pixorial was chosen as an integrated video partner. This integration with Google Drive allowed users to access videos stored in Google Drive within the web app of Pixorial. After the Google Drive launch, Pixorial developed a crowdsourced, location-based video sharing app, Krowds. The app was cited in July 2012 by PC Magazine as one of "The 8 Best Apps for Making and Sharing Videos on Your iPhone". In late July, Pixorial replaced its original mobile app with the MyPlayer HD app that optimized HD video viewing for large screen viewing including tablets and smart televisions. Pixorial's services terminated on July 18, 2014. == Products == === Krowds App === Pixorial's app was launched in April 2013 for iOS, and in May for Android, as a tool to aggregate event videos through location based collections. The app was launched to generally positive reviews. === Movie Creator === Launched July 12, 2012 Pixorial's Movie Creator allowed users to edit movies in a simple story-telling platform Movie Creator's features include transitions, text boxes, access to free music tracks, credits, and social media sharing capabilities. The Pixorial platform allowed users to view, share, and edit videos without modifying the original. Movie Creator integrated pictures and video to create user movies. == Awards == 2012 Apex Award from the Colorado Technology Association, for Best Technology Project of the Year 2010 Computerworld Laureate for Media, Arts and Entertainment

  Read more →
• 

  Seq2seq

  Seq2seq is a family of machine learning approaches used for natural language processing. Originally developed by Lê Viết Quốc, a Vietnamese computer scientist and a machine learning pioneer at Google Brain, this framework has become foundational in many modern AI systems. Applications include language translation, image captioning, conversational models, speech recognition, and text summarization. Seq2seq uses sequence transformation: it turns one sequence into another sequence. == History == One naturally wonders if the problem of translation could conceivably be treated as a problem in cryptography. When I look at an article in Russian, I say: 'This is really written in English, but it has been coded in some strange symbols. I will now proceed to decode. seq2seq is an approach to machine translation (or more generally, sequence transduction) with roots in information theory, where communication is understood as an encode-transmit-decode process, and machine translation can be studied as a special case of communication. This viewpoint was elaborated, for example, in the noisy channel model of machine translation. In practice, seq2seq maps an input sequence into a real-numerical vector by using a neural network (the encoder), and then maps it back to an output sequence using another neural network (the decoder). The idea of encoder-decoder sequence transduction had been developed in the early 2010s. The papers most commonly cited as the originators that produced seq2seq are two papers from 2014. In the seq2seq as proposed by them, both the encoder and the decoder were LSTMs. This had the "bottleneck" problem, since the encoding vector has a fixed size, so for long input sequences, information would tend to be lost, as they are difficult to fit into the fixed-length encoding vector. The attention mechanism, proposed in 2014, resolved the bottleneck problem. They called their model RNNsearch, as it "emulates searching through a source sentence during decoding a translation". A problem with seq2seq models at this point was that recurrent neural networks are difficult to parallelize. The 2017 publication of Transformers resolved the problem by replacing the encoding RNN with self-attention Transformer blocks ("encoder blocks"), and the decoding RNN with cross-attention causally-masked Transformer blocks ("decoder blocks"). === Priority dispute === One of the papers cited as the originator for seq2seq is (Sutskever et al 2014), published at Google Brain while they were on Google's machine translation project. The research allowed Google to overhaul Google Translate into Google Neural Machine Translation in 2016. Tomáš Mikolov claims to have developed the idea (before joining Google Brain) of using a "neural language model on pairs of sentences... and then [generating] translation after seeing the first sentence"—which he equates with seq2seq machine translation, and to have mentioned the idea to Ilya Sutskever and Quoc Le (while at Google Brain), who failed to acknowledge him in their paper. Mikolov had worked on RNNLM (using RNN for language modelling) for his PhD thesis, and is more notable for developing word2vec. == Architecture == The main reference for this section is. === Encoder === The encoder is responsible for processing the input sequence and capturing its essential information, which is stored as the hidden state of the network and, in a model with attention mechanism, a context vector. The context vector is the weighted sum of the input hidden states and is generated for every time instance in the output sequences. === Decoder === The decoder takes the context vector and hidden states from the encoder and generates the final output sequence. The decoder operates in an autoregressive manner, producing one element of the output sequence at a time. At each step, it considers the previously generated elements, the context vector, and the input sequence information to make predictions for the next element in the output sequence. Specifically, in a model with attention mechanism, the context vector and the hidden state are concatenated together to form an attention hidden vector, which is used as an input for the decoder. The seq2seq method developed in the early 2010s uses two neural networks: an encoder network converts an input sentence into numerical vectors, and a decoder network converts those vectors to sentences in the target language. The Attention mechanism was grafted onto this structure in 2014 and is shown below. Later it was refined into the encoder-decoder Transformer architecture of 2017. === Training vs prediction === There is a subtle difference between training and prediction. During training time, both the input and the output sequences are known. During prediction time, only the input sequence is known, and the output sequence must be decoded by the network itself. Specifically, consider an input sequence x 1 : n {\displaystyle x_{1:n}} and output sequence y 1 : m {\displaystyle y_{1:m}} . The encoder would process the input x 1 : n {\displaystyle x_{1:n}} step by step. After that, the decoder would take the output from the encoder, as well as the as input, and produce a prediction y ^ 1 {\displaystyle {\hat {y}}_{1}} . Now, the question is: what should be input to the decoder in the next step? A standard method for training is "teacher forcing". In teacher forcing, no matter what is output by the decoder, the next input to the decoder is always the reference. That is, even if y ^ 1 ≠ y 1 {\displaystyle {\hat {y}}_{1}\neq y_{1}} , the next input to the decoder is still y 1 {\displaystyle y_{1}} , and so on. During prediction time, the "teacher" y 1 : m {\displaystyle y_{1:m}} would be unavailable. Therefore, the input to the decoder must be y ^ 1 {\displaystyle {\hat {y}}_{1}} , then y ^ 2 {\displaystyle {\hat {y}}_{2}} , and so on. It is found that if a model is trained purely by teacher forcing, its performance would degrade during prediction time, since generation based on the model's own output is different from generation based on the teacher's output. This is called exposure bias or a train/test distribution shift. A 2015 paper recommends that, during training, randomly switch between teacher forcing and no teacher forcing. === Attention for seq2seq === The attention mechanism is an enhancement introduced by Bahdanau et al. in 2014 to address limitations in the basic Seq2Seq architecture where a longer input sequence results in the hidden state output of the encoder becoming irrelevant for the decoder. It enables the model to selectively focus on different parts of the input sequence during the decoding process. At each decoder step, an alignment model calculates the attention score using the current decoder state and all of the attention hidden vectors as input. An alignment model is another neural network model that is trained jointly with the seq2seq model used to calculate how well an input, represented by the hidden state, matches with the previous output, represented by attention hidden state. A softmax function is then applied to the attention score to get the attention weight. In some models, the encoder states are directly fed into an activation function, removing the need for alignment model. An activation function receives one decoder state and one encoder state and returns a scalar value of their relevance. Consider the seq2seq language English-to-French translation task. To be concrete, let us consider the translation of "the zone of international control ", which should translate to "la zone de contrôle international ". Here, we use the special token as a control character to delimit the end of input for both the encoder and the decoder. An input sequence of text x 0 , x 1 , … {\displaystyle x_{0},x_{1},\dots } is processed by a neural network (which can be an LSTM, a Transformer encoder, or some other network) into a sequence of real-valued vectors h 0 , h 1 , … {\displaystyle h_{0},h_{1},\dots } , where h {\displaystyle h} stands for "hidden vector". After the encoder has finished processing, the decoder starts operating over the hidden vectors, to produce an output sequence y 0 , y 1 , … {\displaystyle y_{0},y_{1},\dots } , autoregressively. That is, it always takes as input both the hidden vectors produced by the encoder, and what the decoder itself has produced before, to produce the next output word: ( h 0 , h 1 , … {\displaystyle h_{0},h_{1},\dots } , "") → "la" ( h 0 , h 1 , … {\displaystyle h_{0},h_{1},\dots } , " la") → "la zone" ( h 0 , h 1 , … {\displaystyle h_{0},h_{1},\dots } , " la zone") → "la zone de" ... ( h 0 , h 1 , … {\displaystyle h_{0},h_{1},\dots } , " la zone de contrôle international") → "la zone de contrôle international " Here, we use the special token as a control character to delimit the start of input for the decoder. The decoding terminates as soon as "" appears in the decoder output. ==

  Read more →
• 

  Aphelion (software)

  The Aphelion Imaging Software Suite is a software suite that includes three base products - Aphelion Lab, Aphelion Dev, and Aphelion SDK for addressing image processing and image analysis applications. The suite also includes a set of extension programs to implement specific vertical applications that benefit from imaging techniques. The Aphelion software products can be used to prototype and deploy applications, or can be integrated, in whole or in part, into a user's system as processing and visualization libraries whose components are available as both DLLs or .Net components. == History and evolution == The development of Aphelion started in 1995 as a joint project of a French company, ADCIS S.A., and an American company, Amerinex Applied Imaging, Inc. (AAI) Aphelion's image processing and analysis functions were made from operators available from the KBVision software developed and sold by Amerinex's predecessor, Amerinex Artificial Intelligence Inc. In the 1990s, the XLim software library was developed at the Center of Mathematical Morphology of Mines ParisTech, and both companies carried out its development tasks. The first version of Aphelion was completed and released in April 1996. Successive versions were released before the first official stable release in December 1996 at the Photonics East conference in Boston and the Solutions Vision show in Paris in January 1997, where at the latter it competed with Stemmer Imaging's CVB imaging toolbox. In 1998, version 2.3 of Aphelion for Windows 98 was released, and its user base was growing in both France and the United States. Version 3.0, totally rewritten to take advantage of Microsoft's then-recent ActiveX technology, was officially released in 2000. It also became available as a « Developer » version, for rapid prototyping of applications using its intuitive GUI and the macro recording capability, and a « Core » version, including the full library as a set of ActiveX components to be used by software developers, integrators and original equipment manufacturers (OEM). As AAI turned its focus to security, in 2001, ADCIS took the lead on developing Aphelion. AAI focused on millimeter wave scanners for concealed weapon detection at airports, and eventually merged with Millimetrics to become Millivision. In 2004, ADCIS specified version 4.0 of Aphelion. The set of image processing/analysis functions was rewritten one more time to be compatible with the .NET technology and the emergence of 64 bit architecture PCs. In addition, the GUI was redesigned to address two usage types: a semi-automatic use where the user is guided through the different steps of functions, and a fully automatic use where the expert user can quickly invoke imaging functions. Its first release was presented at the IPOT exhibition in Birmingham, UK the same year. During the Vision Show in Paris in October 2008, the new Aphelion Lab product was launched for users that are not specialists in image processing. It is easier to use, and only includes fewer image processing functions. It was then included in the Aphelion Image Processing Suite, consisting of Aphelion Dev (replacing Aphelion Developer), Aphelion Lab, Aphelion SDK (replacing Aphelion Core), and a set of extensions. Nowadays, ADCIS is still working on the suite, and updated versions with new extensions and functionalities continually become available from the websites of both companies. In 2015, support was added for very large images and scan microscope images (virtual slides compound into a very large JPEG 2000 image) for high throughput imaging, and new specific extensions were also added. In late 2015, ADCIS announced Aphelion's port for tablets and smartphones, for vertical applications. The name "Aphelion" comes from the astronomical term of the same name, meaning the point on a planet rotating around the Sun where it lies farthest from it, applying the term in a metaphorical sense. Unix was the operating system used on scientific workstations in the 1990s, such as on the workstations manufactured by market leader Sun Microsystems, which Windows suite Aphelion was quite removed from. == Description == Aphelion is a software suite to be used for image processing and image analysis. It supports 2D and 3D, monochrome, color, and multi-band images. It is developed by ADCIS, a French software house located in Saint-Contest, Calvados, Normandy. Aphelion is widely used in the scientific/industry community to solve basic and complex imaging applications. First, the imaging application is quickly developed from the Graphical User Interface, involving a set of functions that can be automatically recorded into a macro command. The macro languages available in Aphelion (i.e. BasicScript, Python, and C#) help to process batch of images, and prompt the user if needed for specific parameters that are applied to the imaging functions. All Aphelion image processing functions are written in C++, and the Aphelion user interface is written in C#. C++ functions can be called from the C# language thanks the use of dedicated wrappers. The main principle of image processing is to automatically process pixels of a digital image, then extract one or more objects of interest (i.e. cells in the field of biology, inclusions in the field of material science) and compute one or more measurements on those objects to quantify the image and generate a verdict (good image, image with defects, cancerous cells). In other words, starting from an image, pixels are processed by a set of successive functions or operators until only measurements are computed and used as the input of a 3rd party system or a classification software that will classify objects of interest that have been extracted during the imaging process. An acquisition system such as a digital camera, a video camera, an optical or electron microscope, a medical scanner, or a smartphone can be used to capture images. The set of values or pixels can be processed as a 1D image (1D signal), a 2D image (array of pixel values corresponding to a monochrome or color image), or a 3D image displayed using volume rendering (array of voxels in the 3D space) or displaying surfaces by using 3D rendering. A 2D color image is made of 3 value pixels (typically Red, Green, and Blue information or another color space), and a 3D image is made of monochrome, color (indexed color are often used), multispectral, or hyperspectral data. When dealing with videos, an additional band is added corresponding to temporal information. The Aphelion Software Suite includes three base products, and a set of optional extensions for specific applications: Aphelion Lab: Entry-level package for non-experts in image processing. It helps to quickly segment an image in a semi-automatic or manual ways, and compute a set of measurements computed on objects of interest that have been extracted during the segmentation process. A set of wizards guides the user from image acquisition to report generation. Aphelion Dev: Full imaging environment including over 450 functions to develop and deploy an application that involves image processing and analysis. It also includes a set of macro-command languages to automate any application to be invoked from the user interface. It also helps to run the imaging algorithm on more than one image that are stored on disk, available on the network, or captured by an acquisition device. Aphelion libraries for image processing and visualization are provided in Aphelion Dev as DLLs and .Net components. Aphelion SDK: A set of libraries to develop a stand-alone application with a custom interface based on the Aphelion libraries. This software development kit including display, processing and analysis functions that can be used by software developers and OEMs. It is provided as DLLs and .Net components. The stand-alone application is typically developed in C# on one computer, and then deployed on multiple PCs and systems. A set of optional extensions can be added to the « Aphelion Dev » product, depending on the application. An evaluation version of Aphelion can be run on a PC for 30 days. A permanent version of Aphelion is available based on a perpetual license. Upgrades are available through a maintenance agreement based on a yearly fee. Technical support is provided by the engineers who are developing the product. The goal of image processing is usually to extract object(s) of interest in an image, and then to classify them based on some characteristics such as shape, density, position, etc. Using Aphelion, this goal is achieved by performing the following tasks: Load an image from disk or acquire an image using an acquisition device. Enhance the image removing noise or modifying its contrast. Segment the image extracting objects of interest to be measured and analyzed. Typically, for simple applications, a threshold is performed to generate a binary image. Then, morphological operators are applied to clean the image and only keep obj

  Read more →
• 

  Drops (app)

  Drops is a language learning app that was created in Estonia by Daniel Farkas and Mark Szulyovszky in 2015. It is the second product from the company, after their first app, LearnInvisible, had issues in retaining a user's engagement over the required time period. The languages available include Native Hawaiian and Māori, and was classified as one of the fifty "Most Innovative Companies" for 2019 by Fast Company. The company partnered with Global Eagle Entertainment to include Travel Talk, a feature intended to focus on words and phrases frequently used by travelers. At the beginning of the COVID-19 pandemic in March 2020, the number of users increased by 55 percent in the United States and 92 percent in the United Kingdom. Droplets, a language app for children, includes profiles for multiple teachers working with remote students. The company also produces an app called Scripts, intended to help users learn to write alphabets. The app was purchased by the Norwegian company Kahoot! on 24 November 2020.

  Read more →