Micro stuttering is a visual artifact in real-time computer graphics in which the time intervals between consecutively displayed frames are uneven, even though the average frame rate reported by benchmarking software appears adequate. Tools such as 3DMark typically compute frame rates over intervals of one second or more, which can conceal momentary drops in the instantaneous frame rate that the viewer perceives as hitching or jerking of on-screen motion. At low frame rates the effect is visible as a stutter in moving images, degrading the experience in interactive applications such as video games. In severe cases a lower but more consistent frame rate can appear smoother than a higher but more erratic one. The term gained prominence in the late 2000s in discussions of multi-GPU rendering (see History), but micro stuttering also affects single-GPU systems. Common causes on modern hardware include real-time shader compilation, asset streaming from storage, VRAM exhaustion, and driver bugs. == Causes == === Shader compilation === A common cause of micro stuttering on modern PCs is real-time shader compilation. Shaders are small programs that instruct the GPU on how to render visual effects such as lighting, shadows, and reflections. On consoles, developers can pre-compile all shaders for the known, fixed hardware. On PCs, the variety of GPU architectures means shaders must often be compiled at run time, either when the game launches or during gameplay itself. When the rendering engine encounters a shader that has not yet been compiled, the CPU must finish the compilation before the GPU can draw the affected object. This causes a spike in frame time that the player perceives as a hitch. The problem has been particularly associated with games built on Unreal Engine 4 running under DirectX 12, because DX12 shifts more shader management responsibility to the application. Several techniques exist to reduce shader compilation stutter. Pipeline State Object (PSO) pre-caching records the shader permutations used at runtime so that they can be compiled in advance on subsequent launches. Asynchronous shader compilation moves the work to background CPU threads to avoid blocking the main rendering thread. Platform-level services such as Steam's shader pre-caching distribute previously compiled shaders to users with matching GPU hardware. The Steam Deck, which contains a single fixed GPU, benefits from pre-compiled shader caches because all units share the same hardware configuration. === Other causes === Micro stuttering on single-GPU systems can have several additional causes. CPU bottlenecks or scheduling interruptions from background tasks can prevent the processor from preparing frames at regular intervals. Asset streaming during gameplay (loading textures, geometry, or audio from storage) can produce hitches sometimes called traversal stutter; the use of solid-state drives and technologies such as DirectStorage has reduced but not eliminated this. VRAM exhaustion forces data to be swapped between video memory and system memory over the PCI Express bus, which is slower. Graphics driver bugs can also introduce stutter; Nvidia released hotfix driver 551.46 in February 2024 to correct intermittent micro stuttering when V-Sync was enabled. == Measurement == Micro stuttering drew attention to the limitations of average frame rate as a performance metric. In 2013, Scott Wasson at The Tech Report published a series of articles advocating frame time analysis, in which the delivery time of every individual frame is recorded and plotted rather than collapsed into a single frames-per-second figure. This approach was adopted by other hardware review publications in the following years. GPU reviews now routinely report 1% low and 0.1% low frame rates alongside the average. The 1% low is the average frame rate of the slowest 1% of frames in a sample; it serves as an indicator of worst-case smoothness. A large gap between the average and the 1% low suggests poor frame pacing. Tools for capturing per-frame timing data include FRAPS, PresentMon, OCAT, CapFrameX, and MSI Afterburner with RivaTuner Statistics Server. == Mitigation == === Frame pacing === Frame pacing is a software technique that regulates the timing of frame delivery to produce even intervals between displayed frames. Game engines, GPU drivers, and platform libraries all implement frame pacing strategies to varying degrees. On mobile platforms, Google provides the Android Frame Pacing library (Swappy) as part of the Android Game Development Kit. In December 2025, the Khronos Group published the VK_EXT_present_timing Vulkan extension, giving developers explicit control over presentation timing in a cross-platform graphics API for the first time. === Variable refresh rate === Variable refresh rate (VRR) display technologies allow a monitor's refresh rate to change to match the GPU's frame output. Implementations include Nvidia G-Sync (2013), AMD FreeSync (2015), and the VESA Adaptive-Sync standard built into DisplayPort 1.2a and later. VRR eliminates the screen tearing that results from a mismatch between frame rate and refresh rate, and avoids the frame-holding behaviour of V-Sync that can itself cause stutter. It is effective at smoothing moderate frame rate fluctuations but cannot compensate for large sudden spikes in frame time such as those caused by shader compilation or heavy asset streaming. VRR support has become standard in gaming monitors, televisions (via HDMI 2.1), and the Xbox Series X/S and PlayStation 5 consoles. === Frame generation === Beginning with DLSS 3 on the GeForce RTX 40 series in 2022, Nvidia introduced AI-based frame generation, which uses dedicated optical flow hardware and a neural network to create new frames between traditionally rendered ones. AMD followed with FSR 3 in 2023, using an algorithmic approach, and the AI-based FSR 4 for the Radeon RX 9000 series in 2025. DLSS 4, released in January 2025 for the GeForce RTX 50 series, can generate up to three frames per rendered frame using a technique called Multi Frame Generation. Frame generation increases the displayed frame rate but introduces its own frame pacing concerns. If the underlying rendered frames are unevenly timed, the interpolated frames can make the unevenness more apparent rather than less. DLSS 4 addresses this with hardware-level flip metering on the GPU's display engine, which controls the timing of frame presentation more precisely than the CPU-based pacing used in DLSS 3. Both vendors pair frame generation with latency-reduction features (Nvidia Reflex and AMD Anti-Lag+) to offset the additional input latency that results from inserting synthetic frames into the pipeline. === Frame rate limiters === Capping the frame rate below the display's maximum refresh rate, using tools such as RivaTuner Statistics Server, in-game limiters, or driver-level settings, is a common way to improve frame pacing. Preventing the GPU from running ahead of the display reduces variability in frame delivery times and can produce a smoother result than an uncapped but more irregular frame rate. == History == === Multi-GPU configurations === Micro stuttering was first widely documented in the late 2000s as a side effect of multi-GPU configurations using Alternate Frame Rendering (AFR), in which consecutive frames are assigned to alternating GPUs. Because each GPU may take a different amount of time to complete its assigned frame — due to varying scene complexity, driver scheduling, or inter-GPU communication overhead — the resulting frame delivery is irregular even when the average frame rate is high. Both Nvidia SLI and AMD CrossFireX were affected, with dual-GPU setups exhibiting the worst frame pacing irregularities. In 2012 benchmarks using Battlefield 3, dual Radeon HD 7970 cards in CrossFire showed 85% variation in frame delivery times compared with 7% for a single card, while dual GeForce GTX 680 cards in SLI showed only 7% variation compared with 5% for a single card. Multi-GPU micro stuttering became a significant factor in the eventual decline and discontinuation of consumer multi-GPU gaming. Nvidia restricted SLI to a handful of enthusiast-class cards from the GeForce 10 series onward, then replaced it with NVLink on the GeForce RTX 20 series, which saw limited gaming adoption. AMD ceased active CrossFire development around 2017. By the mid-2020s, neither vendor's current consumer GPUs support multi-GPU rendering for games. Other factors that contributed to the decline include DirectX 12 placing multi-GPU support in the hands of game developers rather than driver authors, the incompatibility of temporal anti-aliasing and other temporal rendering techniques with AFR, and the increasing size, power draw, and cost of individual GPUs. The third-party utility RadeonPro could reduce CrossFire micro stuttering through dynamic V-Sync and frame pacing adjustments, and AMD later introduced a driver-level frame paci
CinePlayer
CinePlayer is a software based media player used to review Digital Cinema Packages (DCP) without the need for a digital cinema server by Doremi Labs. CinePlayer can play back any DCP, not just those created by Doremi Mastering products. In addition to playing DCPs, CinePlayer can also playback JPEG2000 image sequences and many popular multimedia file types. There are two versions of CinePlayer available, standard and Pro. The standard version supports playback of non-encrypted, 2D DCP's up to 2K resolution. The Pro version supports playback of encrypted, 2D or 3D DCP's with subtitles up to 4K resolution. == Supported formats == === Containers === AVI MOV MXF MPG TS WMV M2TS MTS MP4 MKV === Video codecs === JPEG2000 ProRes 422 DNxHD YUV Uncompressed 8-10 bits DIVX XVID MPEG4 AVC / H-264 VC-1 MPEG2 === Supported image sequences === BMP TIFF TGA DPX JPG J2C === Supported audio files === WAV MP3 WMA MP2
Cryptosystem
In cryptography, a cryptosystem is a suite of cryptographic algorithms needed to implement a particular security service, such as confidentiality (encryption). Typically, a cryptosystem consists of three algorithms: one for key generation, one for encryption, and one for decryption. The term cipher (sometimes cypher) is often used to refer to a pair of algorithms, one for encryption and one for decryption. Therefore, the term cryptosystem is most often used when the key generation algorithm is important. For this reason, the term cryptosystem is commonly used to refer to public key techniques; however both "cipher" and "cryptosystem" are used for symmetric key techniques. == Formal definition == Mathematically, a cryptosystem or encryption scheme can be defined as a tuple ( P , C , K , E , D ) {\displaystyle ({\mathcal {P}},{\mathcal {C}},{\mathcal {K}},{\mathcal {E}},{\mathcal {D}})} with the following properties. P {\displaystyle {\mathcal {P}}} is a set called the "plaintext space". Its elements are called plaintexts. C {\displaystyle {\mathcal {C}}} is a set called the "ciphertext space". Its elements are called ciphertexts. K {\displaystyle {\mathcal {K}}} is a set called the "key space". Its elements are called keys. E = { E k : k ∈ K } {\displaystyle {\mathcal {E}}=\{E_{k}:k\in {\mathcal {K}}\}} is a set of functions E k : P → C {\displaystyle E_{k}:{\mathcal {P}}\rightarrow {\mathcal {C}}} . Its elements are called "encryption functions". D = { D k : k ∈ K } {\displaystyle {\mathcal {D}}=\{D_{k}:k\in {\mathcal {K}}\}} is a set of functions D k : C → P {\displaystyle D_{k}:{\mathcal {C}}\rightarrow {\mathcal {P}}} . Its elements are called "decryption functions". For each e ∈ K {\displaystyle e\in {\mathcal {K}}} , there is d ∈ K {\displaystyle d\in {\mathcal {K}}} such that D d ( E e ( p ) ) = p {\displaystyle D_{d}(E_{e}(p))=p} for all p ∈ P {\displaystyle p\in {\mathcal {P}}} . Note; typically this definition is modified in order to distinguish an encryption scheme as being either a symmetric-key or public-key type of cryptosystem. == Examples == A classical example of a cryptosystem is the Caesar cipher. A more contemporary example is the RSA cryptosystem. Another example of a cryptosystem is the Advanced Encryption Standard (AES). AES is a widely used symmetric encryption algorithm that has become the standard for securing data in various applications. Paillier cryptosystem is another example used to preserve and maintain privacy and sensitive information. It is featured in electronic voting, electronic lotteries and electronic auctions.
Content management
Content management (CM) are a set of processes and technologies that support the collection, managing, and publishing of information in any form or medium. When stored and accessed via computers, this information may be more specifically referred to as digital content, or simply as content. Digital content may take the form of text (such as electronic documents), images, multimedia files (such as audio or video files), or any other file type that follows a content lifecycle requiring management. The process of content development and management is complex enough that various commercial software vendors (large and small), such as Interwoven and Microsoft, offer content management software to control and automate significant aspects of the content lifecycle. == Process == Content management practices and goals vary by mission and by organizational governance structure. News organizations, e-commerce websites, and educational institutions all use content management, but in different ways. This leads to differences in terminology and in the names and number of steps in the process. For example, some digital content is created by one or more authors. Over time that content may be edited. One or more individuals may provide some editorial oversight, approving the content for publication. Publishing may take many forms: it may be the act of "pushing" content out to others, or simply granting digital access rights to certain content to one or more individuals. Later that content may be superseded by another version of the content and thus retired or removed from use (as when this wiki page is modified). Content management is an inherently collaborative process. It often consists of the following basic roles and responsibilities: Creator – responsible for creating and editing content. Editor – responsible for tuning the content message and the style of delivery, including translation and localization. Publisher – responsible for releasing the content for use. Administrator – responsible for managing access permissions to folders, collections and files, usually accomplished by assigning access rights to user groups or roles. Admins may also assist and support users in various ways. Consumer, viewer or guest – the person who reads or otherwise consumes the content after it is published or shared. A critical aspect of content management is the ability to manage versions of content as it evolves (see also version control). Authors and editors often need to restore older versions of edited products due to a process failure or an undesirable series of edits. Time-sensitive content may also require updates as the subject matter evolves over time. Another equally important aspect of content management involves the creation, maintenance, and application of review standards. Each member of the content creation and review process has a unique role and set of responsibilities in the development or publication of the content. Each review team member requires clear and concise review standards. These must be maintained on an ongoing basis to ensure the long-term consistency and health of the knowledge base. A content management system is a set of automated processes that may support the following features: Import and creation of documents and multimedia material Identification of all key users and their roles The ability to assign roles and responsibilities to different instances of content categories or types Definition of workflow tasks often coupled with messaging so that content managers are alerted to changes in content The ability to track and manage multiple versions of a single instance of content The ability to publish the content to a repository to support access The ability to personalize content based on a set of rules Increasingly, the repository is an inherent part of the system, and incorporates enterprise search and retrieval. Content management systems take the following forms: Web content management system—software for web site management (often what content management implicitly means) Output of a newspaper editorial staff organization Workflow for article publication Document management systems Knowledge management software Single source content management system—content stored in chunks within a relational database Variant management system—where personnel tag source content (usually text and graphics) to represent variants stored as single source "master" content modules, resolved to the desired variant at publication (for example: automobile owners manual content for 12 model years stored as single master content files and "called" by model year as needed)—often used in concert with database chunk storage (see above) for large content objects == Governance structures == Content management expert Marc Feldman defines three primary content management governance structures: localized, centralized, and federated—each having its unique strengths and weaknesses. === Localized governance === By putting control in the hands of those closest to the content, the context experts, localized governance models empower and unleash creativity. These benefits come, however, at the cost of a partial-to-total loss of managerial control and oversight. === Centralized governance === When the levers of control are strongly centralized, content management systems are capable of delivering an exceptionally clear and unified brand message. Moreover, centralized content management governance structures allow for a large number of cost-savings opportunities in large enterprises, realized, for example, through (1) the avoidance of duplicated efforts in creating, editing, formatting, repurposing and archiving content; (2) process management and the streamlining of all content related labor; and/or (3) an orderly deployment or updating of the content management system. === Federated governance === Federated governance models potentially realize the benefits of both localized and centralized control while avoiding the weaknesses of both. While content management software systems are inherently structured to enable federated governance models, realizing these benefits can be difficult because it requires, for example, negotiating the boundaries of control with local managers and content creators. In the case of larger enterprises, in particular, the failure to fully implement or realize a federated governance structure equates to a failure to realize the full return on investment and cost savings that content management systems enable. == Implementation == Content management implementations must be able to manage content distributions and digital rights in content life cycle. Content management systems are usually involved with digital rights management in order to control user access and digital rights. In this step, the read-only structures of digital rights management systems force some limitations on content management, as they do not allow authors to change protected content in their life cycle. Creating new content using managed (protected) content is also an issue that gets protected contents out of management controlling systems. A few content management implementations cover all these issues.
European Grid Infrastructure
EGI (originally an initialism for European Grid Infrastructure) is a federation of computing and storage resource providers that deliver advanced computing and data analytics services for research and innovation. The Federation is governed by its participants represented in the EGI Council and coordinated by the EGI Foundation. As of 2024, the EGI Federation supports 160 scientific communities worldwide and over 95,000 users in their intensive data analysis. The most significant scientific communities supported by EGI in 2022 were Medical and Health Sciences, High Energy Physics, and Engineering and Technology. The EGI Federation provideds services through over 150 data centres, of which 25 are cloud sites, in 43 countries and 64 Research Infrastructures (4 of which are members of the Federation). == Name == Originally, EGI stood for European Grid Infrastructure. This reflected its focus on providing access to high-throughput computing resources across Europe using Grid computing techniques. However, as EGI's service offerings expanded beyond traditional grid computing, particularly with the incorporation of federated cloud services, the original meaning of the acronym became less accurate. To emphasise the broader scope of EGI's services and avoid any confusion associated with the outdated term "grid," it is recommended to refer to EGI simply as EGI. == Structure == === EGI Federation === The EGI Federation delivers a scalable digital research infrastructure (e-infrastructure), empowering tens of thousands of researchers across diverse scientific disciplines. Through the EGI Federation, researchers gain access to advanced computing and data analytics capabilities, including large-scale data analysis, while benefiting from the collaborative efforts of hundreds of service providers from both public and private sectors, consolidating resources from Europe and beyond. Overall, the EGI Federation offers a range of services, encompassing distributed high-throughput computing and cloud computing, storage and data management capabilities, co-development of new solutions, expert support, and comprehensive training opportunities. This ecosystem propels collaboration, scientific progress and innovation. === EGI Foundation === The EGI Foundation is the coordinating body of the EGI Federation. It was established in 2010 with headquarters in Amsterdam, Netherlands. The Foundation coordinates the research and innovation efforts of its members, spanning technical areas critical to data-intensive science, including large-scale data processing and analysis, distributed Artificial Intelligence/Machine Learning, federated Identity and access management and the application of digital twins for research. The day-to-day running of the EGI Foundation is supervised by the Executive Board. The board’s members work closely with the EGI Director on operational, technical and financial issues. The Executive Board’s members are appointed by the EGI Council for a two-year term. === EGI Council === The EGI Council is responsible for defining the strategic direction of the EGI Federation. The Council acts as the senior decision-making and supervisory authority of the EGI Foundation, with a mandate to define the strategic direction of the entire EGI ecosystem. === EGI Services === EGI offers a suite of services to support data-intensive research. These services include compute resources, orchestration tools, storage and data management solutions, training programmes, security and identity services, and applications. Compute resources encompass cloud compute, cloud container compute, high-throughput compute, and software distribution. Orchestration tools include the Workload Manager and infrastructure manager. Storage and data management solutions include online storage, data transfer, and DataHub. Training programmes cover FitSM, ISO 27001, and general training infrastructure. EGI Check-in and Secrets Store are key security and identity services, while applications such as Notebooks and Replay enhance research productivity. In addition to services for Research, EGI also provides services for Federation and Business. Services for Federation are designed to help resource providers and user communities collaborate and share resources. EGI also offers a range of services to support businesses in their digital transformation. Through the EGI Digital Innovation Hub (EGI DIH), companies can access advanced computing resources, networking, funding and training opportunities, collaborate with research institutions, and test solutions before investing. == History == In 2002, the first large-scale experimental facility was successfully demonstrated by the DataGrid project under the lead of CERN with tens of technical architects from the major High Energy Physics institutes in the world. For the first time, distributed computing was applied to data-intensive processing. It aimed at developing a large-scale computational grid to facilitate distributed data-intensive scientific computing across High Energy Physics, Earth Observation, and Biology science applications. On 28 February 2003, the first software release of LCG-MW was published. gLite, the Lightweight Middleware for Grid Computing and LCG, Large Hadron Collider Computing Grid, are the cornerstone of the Worldwide LHC Computing Grid, which expanded over time towards the EGI Federation. 2004 marks the year of the first pilot infrastructure, seeing the participation of CERN and data centres in the United Kingdom, Spain, Germany, the Netherlands, France, Canada, Russia, Bulgaria, the Asia-Pacific region and Switzerland. Over the years, the infrastructure has grown into a federation of 128 data centres and 25 cloud providers serving more than 95,000 users worldwide. In 2004, the first data processing tasks started being formally recorded in a central accounting system. The EGI Accounting Portal provides the accounting data for Compute, Storage and Data services gathered from the data centres of the EGI Federation. A few years later, in 2010, EGI was established as the coordinating body of the EGI Federation to build an integrated pan-European infrastructure to support European research communities primarily. In the same year, EGI launched the flagship project EGI Inspire. That project brought together European organisations to establish a sustainable European Grid Infrastructure for large-scale data analysis. The success of the project was due to the adoption of a distributed computing model to solve big data problems. Moreover, EGI-Inspire harmonised operational policies across its federation of affiliated data centres and cloud service providers worldwide, integrating e-infrastructures from 57 countries. The EGI Federation was the first to apply federation to cloud provisioning, opening a new avenue in large-scale interactive data analysis. In 2015, within EGI Engage, opening a new avenue in large-scale interactive data analysis. The EGI Federated Cloud is an IaaS-type cloud, incorporating academic and private clouds and virtualised resources built using open standards. Its development is driven by the needs of the scientific community, resulting in a novel research e-infrastructure that relies on well-established federated operational services, making EGI a dependable resource for scientific endeavours. In 2015, EGI, EUDAT, GÉANT, LIBER and OpenAIRE published a position paper on a 'European Open Science Cloud for Research'. With the EOSC-hub project in 2016, EGI started contributing in practice to shaping the services for the EOSC. The work continued with a series of projects, like EOSC Enhance, EOSC Life and EOSC Synergy. With EGI-ACE and its contribution to EOSC Future, EGI has continued developing the EOSC Core. In early 2024, EGI started providing services to the EOSC EU Node, and with EOSC Beyond it will provide new EOSC Core capabilities and pilot additional national and thematic nodes. In October 2024, EUDAT, GÉANT, OpenAIRE, PRACE and EGI signed a Memorandum of Understanding establishing the European e-Infrastructures Assembly. This collaboration will bolster the position and promote the services of e-Infrastructures, empowering researchers across Europe to drive innovation and advance scientific discovery.
QANDA
QANDA (stands for 'Q and A') is an AI-based learning platform developed by Mathpresso Inc., a South Korea-based education technology company. Its best known feature is a solution search, which uses optical character recognition technology to scan problems and provide step-by-step solutions and learning content. As of March 2024, QANDA solved over 6.3 billion questions. QANDA has 90 million total registered users and has reached 8 million monthly active users (MAU) in 50 countries. 90% of the cumulative users are from overseas such as Vietnam and Indonesia. In January 2024, its MathGPT, a math-specific small large language model set a new world record, surpassed Microsoft's 'ToRA 13B', the previous record holder in benchmarks assessing mathematical performance such as 'MATH' (high school math) and 'GSM8K' (grade school math). 'MathGPT' was co-developed with Upstage and KT. In March 2024, Mathpresso launched 'Cramify' (formerly known as Prep.Pie), an AI-powered study material generator designed to create personalized exam prep materials for U.S. college students. It uses generative AI to create customized study materials uploaded by students. Its features include a range of tools including study summarizer and question solver. == History == Co-founder Jongheun ‘Ray’ Lee first came up with the idea of QANDA during his freshman year in college. While he was tutoring to earn money, Lee realized that the quality of education a student receives is greatly based on their location. Lee saw his K-12 students were regularly asking similar questions and realized that these questions were from a pre-selected number of textbooks currently being used in schools. He decided to team up with his high school friend, Yongjae ‘Jake’ Lee to build a platform whereby, one uses a mobile app to scan and submit questions, and students can ask and receive detailed responses. Lee's school friends, Wonguk Jung and Hojae Jeong, joined the team. In June 2015, Mathpresso, Inc. was founded in Seoul, South Korea. In January 2016, Mathpresso's first product QANDA was launched. It supported a Q&A feature between students and tutors. In October 2017, QANDA introduced an AI-based search capability that permitted users to search for answers in seconds. In April 2020, Jake Yongjae Lee(CEO & co-founder) and Ray Jongheun Lee (co-founder) were selected as Forbes 30 under 30 Asia. In June 2021, QANDA raised $50 million in series C funding. Jake Yongjae Lee was recognized as an Innovator Under 35 by MIT Technology Review. In November 2021, QANDA secured a strategic investment from Google. Since its inception, it has received backing in Series C funding from investors namely Google, Yellowdog, GGV Capital, Goodwater Capital, KDB, and SKS Private Equity with participation from SoftBank Ventures Asia, Legend Capital, Mirae Asset Venture Investment, and Smilegate Investment. In September 2023, Mathpresso has raised $8 million (10 billion KRW) from Korea's telecom giant, KT. The total cumulative investment is about 130 million US dollars. The partnership aims to accelerate the development of an education-specific Large Language Model. The company intends to incorporate the LLM model to fortify its AI tutor, which later will be integrated into the existing services: QANDA App, B2B & B2G Saas, and 1:1 online tutoring (QANDA Tutor). == Features == QANDA features OCR-based solution search, one-on-one Q&A tutoring, a study timer. In 2021, QANDA launched additional features, including the premium subscription model that offers unlimited “byte-sized” micro-video lectures and the community feature that enhances collaborative learning. In 2021, QANDA launched QANDA Tutor, a tablet-based 1:1 tutoring service and QANDA Study, a 1:N online school in Vietnam. In 2022, QANDA launched an exam prep feature that offers past exam materials from school via online. This feature is currently available in South Korea. In August 2023, QANDA launched a beta version of an LLM-powered AI Tutor. == Awards and recognition == Best Hidden Gems of 2017 by Google Playstore 2018 AWS AI Startup Challenge Award National representative for the Google AI for Social Good APAC, 2018 Best Self-Improvement Apps of 2018 by Google Playstore GSV Edtech 150 — the Most Transformational Growth Companies in Digital Learning Speaker at the Google App Summit, 2021 Selected as a prospect unicorn company by Korea Technology Finance Corporation in 2023 Winner of G20-DIA Global Pitching in 2023 2021, 2022, 2023 East Asia EdTech 150 by HolonIQ
Classora
Classora is a knowledge base for the Internet oriented to data analysis. From a practical point of view, Classora is a digital repository that stores structured information and allows it to be displayed in multiple formats: analytically, graphically, geographically (through maps); as well as carry out OLAP analysis. The information contained in Classora comes from public sources and is uploaded into the system through bots and ETL processes. The Knowledge Base has a commercial API for semantic enhancement, and an open web through which any user can access to part of the information collected (it also allows users to complete data and share opinions). Internally, Classora is organized into Knowledge Units and Reports. A «Knowledge Unit» is any element of the World about which information may be stored and presented in the form of a data sheet (a person, a company, a country, etc.) A «Report» is a group of Knowledge Units: a ranking of companies, a sport classification table, a survey about people, etc. In fact, one of the technical capabilities of Classora is that it allows the comparison of reports and knowledge units gathered from different sources, thereby generating an added value for the media in which this information is published: digital media, interactive TV, etc. == Key definitions == === Knowledge unit === The units of knowledge (also known as entries) in Classora are data sheets that have a certain semantic equivalence with the articles on the Wikipedia: they store information about any element of the world, be it a film, a country, a company or an animal. However, they differ from Wikipedia in that Classora stores structured information, enriched with a metadata layer; and therefore it is able to automatically interpret the meaning of each unit of knowledge. === Data report === A report is a group of units of knowledge in which the repetition of elements is not allowed. This definition includes any list, poll, ranking, etc.; and, in general, any consultation that involves more than one unit of knowledge. Classora excels at the reports management due to its visualization capabilities, being able to display data in the form of tables, graphs and maps. Types of reports: Sports scores: Sports competitions results sanctioned by the competent institution. Rankings and lists: All types of interesting and curious lists, whether they have an implicit order or not. Polls: Units of knowledge that are ranked according to users’ votes. Queries to the Knowledge Base: Questions from users using CQL. Networks of connections: automatically calculated from the reports and the taxonomy of each Knowledge Unit. === Organizational taxonomy === An organizational taxonomy (also referred to as entry type) is a data sheet that brings together the common attributes of a set of units of knowledge. For instance, the organizational taxonomy F1 Driver displays attributes such as date of debut, team, etc.; and the organizational taxonomy Football Club presents attributes such as city, stadium, etc. In Classora, taxonomies are hierarchically organized, so that they inherit attributes from their parent taxonomies. For instance, F1 Driver is a subsidiary taxonomy of Sportsperson, which is a subsidiary taxonomy of Person, which in turn is a subsidiary taxonomy of Organism. The simplest type of entry in Classora is Classora Object. All the other taxonomies are its subsidiaries and inherit its attributes. In fact, the only attribute Classora Object possesses is name (all units of knowledge are required to have one name at least). == Architecture of Classora == === Data Extraction Module === The Data Extraction Module consists of a set of robots coordinated by software that also manages the potential incidents. Most of the information available in Classora is automatically uploaded through those robots, which connect to the main online public sources to gather all types of data. There are three categories of robots: Extraction robots: responsible for the massive uploading of reports from official public sources (FIFA, CIA, IMF, Eurostat...). They are used for either absolute or incremental data uploading. Data scanner robots: responsible for looking for and updating the data of a unit of knowledge. They use specific sources to perform this task: Wikipedia, IMDB, World Bank, etc. Content aggregators: they don’t connect to external sources. Instead, they generate new information using Classora’s internal database. === Participatory Module === In Classora’s Open Website, Internet users may participate providing their knowledge as they would on the Wikipedia. There are different ways to participate: adding or correcting data in the Knowledge Base, voting in surveys (participatory rankings) and creating new Knowledge Units and Data Reports. === Connectivity Module === The Knowledge Base is designed to be embedded in multi-platform, multi-channel systems, thus enabling its integration into mobile devices, tablets, interactive TV, etc. This integration may be carried out through specific plugins (for navigators or other devices) or an API REST that provides content in XML or JSON formats. The API is divided into three blocks of operations. The first one is the block of general utility tools (ranging from autosuggest components about geographical hierarchies to operations to obtain the list of today’s celebrity birthdays, using CQL). The second one is the block of operations for widget generation (graphs, maps, rankings) using information from the knowledge base. Finally, there is a block of operations designed for the publication of free-source content. == Project statistics == As of April 2012, 2,000,000 Knowledge Units, 15,000 Reports, around 10,000 Maps and several million potential Comparative Analyses had been added to Classora. According to the site of web metrics Alexa, Classora Open Website is ranked at 100,557 globally and at 2,880 in the Spanish traffic ranking. Users spend an average of 9 ½ minutes in Classora.