Confidential computing is a security and privacy-enhancing computational technique focused on protecting data in use. Confidential computing can be used in conjunction with storage and network encryption, which protect data at rest and data in transit respectively. It is designed to address software, protocol, cryptographic, and basic physical and supply-chain attacks, although some critics have demonstrated architectural and side-channel attacks effective against the technology. The technology protects data in use by performing computations in a hardware-based trusted execution environment (TEE). Confidential data is released to the TEE only once it is assessed to be trustworthy. Different types of confidential computing define the level of data isolation used, whether virtual machine, application, or function, and the technology can be deployed in on-premise data centers, edge locations, or the public cloud. It is often compared with other privacy-enhancing computational techniques such as fully homomorphic encryption, secure multi-party computation, and Trusted Computing. Confidential computing is promoted by the Confidential Computing Consortium (CCC) industry group, whose membership includes major providers of the technology. == Properties == Trusted execution environments (TEEs) "prevent unauthorized access or modification of applications and data while they are in use, thereby increasing the security level of organizations that manage sensitive and regulated data". Trusted execution environments can be instantiated on a computer's processing components such as a central processing unit (CPU) or a graphics processing unit (GPU). In their various implementations, TEEs can provide different levels of isolation including virtual machine, individual application, or compute functions. Typically, data in use in a computer's compute components and memory exists in a decrypted state and can be vulnerable to examination or tampering by unauthorized software or administrators. According to the CCC, confidential computing protects data in use through a minimum of three properties: Data confidentiality: "Unauthorized entities cannot view data while it is in use within the TEE". Data integrity: "Unauthorized entities cannot add, remove, or alter data while it is in use within the TEE". Code integrity: "Unauthorized entities cannot add, remove, or alter code executing in the TEE". In addition to trusted execution environments, remote cryptographic attestation is an essential part of confidential computing. The attestation process assesses the trustworthiness of a system and helps ensure that confidential data is released to a TEE only after it presents verifiable evidence that it is genuine and operating with an acceptable security posture. It allows the verifying party to assess the trustworthiness of a confidential computing environment through an "authentic, accurate, and timely report about the software and data state" of that environment. "Hardware-based attestation schemes rely on a trusted hardware component and associated firmware to execute attestation routines in a secure environment". Without attestation, a compromised system could deceive others into trusting it, claim it is running certain software in a TEE, and potentially compromise the confidentiality or integrity of the data being processed or the integrity of the trusted code. == Technical approaches == Technical approaches to confidential computing may vary in which software, infrastructure and administrator elements are allowed to access confidential data. The "trust boundary," which circumscribes a trusted computing base (TCB), defines which elements have the potential to access confidential data, whether they are acting benignly or maliciously. Confidential computing implementations enforce the defined trust boundary at a specific level of data isolation. The three main types of confidential computing are: Virtual machine isolation Application isolation, also known as process isolation Function isolation, also known as library isolation Virtual machine isolation removes the elements controlled by the computer infrastructure or cloud provider, but allows potential data access by elements inside a virtual machine running on the infrastructure. Application or process isolation permits data access only by authorized software applications or processes. Function or library isolation is designed to permit data access only by authorized subroutines or modules within a larger application, blocking access by any other system element, including unauthorized code in the larger application. == Threat model == As confidential computing is concerned with the protection of data in use, only certain threat models can be addressed by this technique. Other types of attacks are better addressed by other privacy-enhancing technologies. === In scope === The following threat vectors are generally considered in scope for confidential computing: Software attacks: including attacks on the host’s software and firmware. This may include the operating system, hypervisor, BIOS, other software and workloads. Protocol attacks: including "attacks on protocols associated with attestation as well as workload and data transport". This includes vulnerabilities in the "provisioning or placement of the workload" or data that could cause a compromise. Cryptographic attacks: including "vulnerabilities found in ciphers and algorithms due to a number of factors, including mathematical breakthroughs, availability of computing power and new computing approaches such as quantum computing". The CCC notes several caveats in this threat vector, including relative difficulty of upgrading cryptographic algorithms in hardware and recommendations that software and firmware be kept up-to-date. A multi-faceted, defense-in-depth strategy is recommended as a best practice. Basic physical attacks: including cold boot attacks, bus and cache snooping and plugging attack devices into an existing port, such as a PCI Express slot or USB port. Basic upstream supply-chain attacks: including attacks that would compromise TEEs through changes such as added debugging ports. The degree and mechanism of protection against these threats varies with specific confidential computing implementations. === Out of scope === Threats generally defined as out of scope for confidential computing include: Sophisticated physical attacks: including physical attacks that "require long-term and/or invasive access to hardware" such as chip scraping techniques and electron microscope probes. Upstream hardware supply-chain attacks: including attacks on the CPU manufacturing process, CPU supply chain in key injection/generation during manufacture. Attacks on components of a host system that are not directly providing the capabilities of the trusted execution environment are also generally out-of-scope. Availability attacks: confidential computing is designed to protect the confidentiality and integrity of protected data and code. It does not address availability attacks such as Denial of Service or Distributed Denial of Service attacks. == Use cases == Confidential computing can be deployed in the public cloud, on-premise data centers, or distributed "edge" locations, including network nodes, branch offices, industrial systems and others. === Data privacy and security === Confidential computing protects the confidentiality and integrity of data and code from the infrastructure provider, unauthorized or malicious software and system administrators, and other cloud tenants, which may be a concern for organizations seeking control over sensitive or regulated data. The additional security capabilities offered by confidential computing can help accelerate the transition of more sensitive workloads to the cloud or edge locations. === Multi-party analytics === Confidential computing can enable multiple parties to engage in joint analysis using confidential or regulated data inside a TEE while preserving privacy and regulatory compliance. In this case, all parties benefit from the shared analysis, but no party's sensitive data or confidential code is exposed to the other parties or system host. Examples include multiple healthcare organizations contributing data to medical research, or multiple banks collaborating to identify financial fraud or money laundering. Oxford University researchers proposed the alternative paradigm called "Confidential Remote Computing" (CRC), which supports confidential operations in Trusted Execution Environments across endpoint computers considering multiple stakeholders as mutually distrustful data, algorithm and hardware providers. === Confidential generative AI === Confidential computing technologies can be applied to various stages of a generative AI deployments to help increase data or model privacy, security, and regulatory compliance. TEEs and remote attestation can protect the integrity of data during AI model training, keep
Kubity
Kubity is a cloud-based 3D communication tool that works on desktop computers, the web, smartphones, tablets, augmented reality gear, and virtual reality glasses. Kubity is powered by several proprietary 3D processing engines including "Paragone" and "Etna" that prepare the 3D file for transfer over mobile devices. Kubity has practical applications for architecture, interior design, engineering, product design, film, and video games among others. The majority of its users create 3D models using SketchUp or Autodesk Revit software. Kubity products include the Kubity web app and Kubity Go (a mobile application for iOS and Android). Kubity is compatible across many platforms, devices and operating systems including: iOS, Android, Firefox, Chrome, Windows, MacOS, and Linux. == History == Kubity was created by SPK Technology (ex Kubity S.A.S.), a Paris-based software company specializing in automatic 3D data optimization and visualization. Founded in 2012 by a group of software engineers and an urban projects developer, they united around a simple idea: create a way for anyone, anywhere to simply and intuitively explore 3D models on smartphones and computers. In order to bring architects, engineers and designers together with their clients around a 3D model, it was essential to develop an interactive platform that supported multiple desktop and mobile devices for instantaneous and fluid 3D navigation. With specifications in place, 15 engineers fused together several technologies: 3D design, data compression, decimation and rendering optimization, web and mobile transfer, and virtual reality headset integration. In January 2014, the first public Kubity prototype (1.0 Amethyst) was launched to a small group of beta testers with a plug-in that allowed users to import 3D models from SketchUp to their browser. A global release was announced in April 2014 at the SketchUp Basecamp in Vail, Colorado. In May 2015, Kubity launched a web application that worked using WebGL technology (2.0 Citrine). For the first time, users were able to drag and drop any SketchUp file in a web browser without having to install a plug-in. In December 2015, Kubity launched a mobile application on the App Store for iPhone, iPod, and iPad as well as on Google Play for Android devices (3.0 Druzy). In November 2016, Kubity launched support for Oculus Rift and HTC Vive (4.0 Emerald). Beginning in November 2017, Kubity launched a full suite rollout of mobile applications over six months that included Kubity AR for augmented reality, Kubity VR for virtual reality, and Kubity Mirror for remote presentations and screen mirroring (5.0 Feldspar). In September 2018, a one-click plugin for SketchUp and Revit (Kubity PRO), along with a mobile-first revamp of Kubity Go was launched, allowing PRO-to-Go device pairing for automatic mobile sync (6.0 Gypsum). In early 2019, the Kubity Go application was updated to include fully integrated AR, VR, and screen mirroring functionalities, killing off the dedicated companion apps Kubity AR, Kubity VR and Kubity Mirror in the process (7.0 Heliotrope). In January 2020, support for the Kubity PRO plugin for SketchUp and Revit was migrated to a SketchUp-only web app. == Technology == Kubity is powered by a proprietary 3D crystallization engine known as "Paragone"; a technology developed by SPK Technology. Paragone takes constrained information from a 3D file and runs it through the "BlockWave" algorithm (US Patent 10,482.629), also developed by SPK Technology. BlockWave is a multiphase optimization algorithm that combines 3D design, data compression, decimation and rendering optimization, web and mobile transfer, and mixed reality headset integration to create a crystallized universal format of the original file. One phase of the BlockWave algorithm is based on the quadric-based polygonal surface simplification algorithm, performed using predefined heuristics, and is associated with a plurality of simplified versions of the 3D model, each version being associated with a predefined level of detail adapted to the user specific end device. BlockWave extracts data content, geometry and textures, then sets quadrics for each top of the original 3D model, and identifies pairs of adjacent tops linked by vertices. The algorithm uses a local collapsing operator and a top-plan error metric to obtain a fixed number of faces or a maximum defined error; 3D meshing is simplified by replacing two points with one, then deleting the degrading faces and updating adjacent relations. Once decimation is completed, texture optimization is set using texture target parameters allowing maximized GPU memory to improve computing time. With texture encoding completed, the crystallized universal 3D file can now be easily opened on any user-specific end device and played across most digital devices with real-time rendering. == Features == === 3D Crystallization === A user converts (or crystallizes) a 3D file by exporting it with the Kubity web app. Crystallization adds features like AR/VR and cinematic fly-through tour as well as assigns the model a dedicated QR code. === Automatic Mobile Sync === When a 3D model is exported, it is automatically synced to Kubity Go on the user's mobile device. From there, it can be accessed, explored, and shared with others with or without an internet connection. === Security and Management === User models can be managed all in one place on Kubity Go or in a browser from their account. Models can be renamed, password-protected, shared, and played. === Augmented Reality === Developed using Apple ARKit and Google ARCore technology, Kubity Go's augmented reality feature maps the environment in a room detecting horizontal planes like tables and floors to track and place 3D objects. By blending digital objects and information with the environment, Kubity allows users to interact with 3D models in true augmented reality. Built-in communication features allows users to instantly share 3D models with anyone over text, email, social media, or direct device-to-device with a QR Code. Platform Support AR supports devices running iOS11 including: iPhone SE, iPhone 6s, iPhone 6s Plus, iPhone 7, iPhone 7 Plus, iPhone 8, iPhone X, all iPad Pro models, and iPad (2017). AR for Android requires Android 7.0 or later and access to the Google Play Store. === Virtual Reality === VR allows users to explore SketchUp models and Revit projects on-the-go right from a mobile device using Oculus Go, Google Cardboard, Samsung Gear VR, or the glasses-free Magic Window feature. Kubity's virtual reality feature is compatible with Oculus Go, Google Cardboard viewers and other cardboard compatible devices including clip-on style VR glasses like Homido Mini, as well as the mobile virtual reality headset, Samsung Gear VR. Samsung Gear VR supports: Galaxy S6, Galaxy S6 Edge, Galaxy S6 Edge+, Samsung Galaxy Note 5, Galaxy S7, Galaxy S7 Edge, Galaxy S8, Galaxy S8+, Samsung Galaxy Note Fan Edition, Samsung Galaxy Note 8, Samsung Galaxy A8/A8+ (2018), and Samsung Galaxy S9/Galaxy S9+. === Screen Mirroring === Screen mirroring allows a user to sync the sender device to a receiver on a webpage, then control from the sender device to give a remote presentation of the 3D model. Devices are easily synced by entering a six-digit number displayed on the receiving computer. == Platform support == On iOS, the Kubity application is compatible with devices running on the version 9.0 or higher. On Android, Kubity is compatible with devices running on the version 4.4 “Kit Kat” or higher. The web version of Kubity applications currently support web browsers compatible with WebGL2 : Mozilla Firefox and Google Chrome. AR is compatible with devices running iOS11 including: iPhone SE, iPhone 6s, iPhone 6s Plus, iPhone 7, iPhone 7 Plus, iPhone 8, iPhone X, all iPad Pro models, and iPad (2017), and Android devices. Requires Android 7.0 or later and access to the Google Play Store. VR is compatible with Google Cardboard viewers and other cardboard compatible devices including clip-on style VR glasses like Homido Mini, as well as the Samsung Gear VR and Oculus Go. Samsung Gear VR supports: Galaxy S6, Galaxy S6 Edge, Galaxy S6 Edge+, Samsung Galaxy Note 5, Galaxy S7, Galaxy S7 Edge, Galaxy S8, Galaxy S8+, Samsung Galaxy Note Fan Edition, Samsung Galaxy Note 8, Samsung Galaxy A8/A8+ (2018) and Samsung Galaxy S9/Galaxy S9+.
Gary B. Fogel
Gary Bryce Fogel (born 1968) is an American biologist and computer scientist. He is the Chief Executive Officer of Natural Selection, Inc. He is most known for his applications of computational intelligence and machine learning to bioinformatics, computational biology, and industrial optimization. == Education and Research == Fogel was born and raised in La Jolla, California, graduating from La Jolla High School. He received a B.A. in biology with a minor in earth sciences from the University of California, Santa Cruz in 1991 and a Ph.D. in biology from the University of California, Los Angeles in 1998. Fogel has published over 150 peer-reviewed publications in conferences and journals, 2 edited books, and 11 patents. As CEO of Natural Selection, Inc., his research focuses on the application of computational intelligence, machine learning, and predictive analytics in areas not limited to: Viral evolution, cellular differentiation, drug discovery, RNA structure, cis-regulatory elements, cancer, and evolutionary game theory as well as the development of evolutionary algorithms and other approaches. == Service == Between 2008–2018 Gary Fogel was editor-in-chief of the Elsevier journal BioSystems. He has served previously as an associate editor for IEEE Transactions on Artificial Intelligence, IEEE Computational Intelligence Magazine (2005–2010), IEEE Transactions on Evolutionary Computation (2001–2013), IEEE Transactions on Emerging Topics in Computational Intelligence (2016–2018), IEEE/ACM Transactions on Computational Biology and Bioinformatics (2004–2008), International Journal of Bioinformatics Research and Applications (2004–2007), International Journal of Data Mining and Bioinformatics (2005–2007), as a consulting editor for the Journal of Computational Intelligence in Bioinformatics (2006–2007), and as an editorial board member of Ecological Informatics (2005–2009) and BMC Big Data Analytics (2015–2020). Within the IEEE Computational Intelligence Society, Fogel founded the Bioinformatics and Bioengineering Technical Committee and established the IEEE Computational Intelligence in Bioinformatics and Computational Biology conference series, chairing the first two meetings in 2004 and 2005 in San Diego. He co-founded the IEEE Conference on Artificial Intelligence in 2023. Fogel served on the IEEE Computational Intelligence Society Administrative Committee (2004–2009, 2014–2022) and served as IEEE CIS Vice President of Conferences (2010–2013, 2019). == Teaching == Gary Fogel also serves as adjunct faculty at San Diego State University in the department of aerospace engineering as well as in the Computational Science Research Center. He has authored four books and numerous articles on the history of early aviation focusing on motorless flight. He is an associate fellow of the American Institute of Aeronautics and Astronautics and serves on the AIAA History Committee. == Awards == 2023 – Outstanding Contribution to Aerospace Education Award, AIAA San Diego Section 2022 – Elected Fellow of the Asia-Pacific Artificial Intelligence Association 2019 – Top 100 AI Leaders in Drug Discovery and Advanced Healthcare by Deep Knowledge Analytics 2019 – Outstanding Contribution to Aerospace Education Award, AIAA San Diego Section 2016 – Meritorious Service Award, IEEE Computational Intelligence Society 2016 – Outstanding Contribution to the Community Award, AIAA San Diego Section 2015 – Outstanding Enhancement of the Image of the Aerospace Profession Award, AIAA San Diego Section 2012 – Medal for Significant Achievement, San Diego Chapter of Sigma Xi 2012 – Fellow of the Institute of Electrical and Electronics Engineers for contributions to computational intelligence and its application to biology, chemistry, and medicine. == Aeromodeling == Gary Fogel has established national and world records for model aircraft. He helped establish the National Model Aviation Heritage program for the Academy of Model Aeronautics. He is a leader member, contest director, and fellow of the Academy of Model Aeronautics, and was inducted into the Academy of Model Aeronautics Hall of Fame in 2025.
Dan Jurafsky
Daniel Jurafsky is a professor of linguistics and computer science at Stanford University, and also an author. With Daniel Gildea, he is known for developing the first automatic system for semantic role labeling (SRL). He is the author of The Language of Food: A Linguist Reads the Menu (2014) and a textbook on speech and language processing (2000). For the former, Jurafsky was named a finalist for the James Beard Award. Jurafsky was given a MacArthur Fellowship in 2002. == Education == Jurafsky received his B.A in linguistics (1983) and Ph.D. in computer science (1992), both at University of California, Berkeley; and then a postdoc at International Computer Science Institute, Berkeley (1992–1995). == Academic life == He is the author of The Language of Food: A Linguist Reads the Menu (W. W. Norton & Company, 2014). With James H. Martin, he wrote the textbook Speech and Language Processing: An Introduction to Natural Language Processing, Computational Linguistics, and Speech Recognition (Prentice Hall, 2000). The first automatic system for semantic role labeling (SRL, sometimes also referred to as "shallow semantic parsing") was developed by Daniel Gildea and Daniel Jurafsky to automate the FrameNet annotation process in 2002; SRL has since become one of the standard tasks in natural language processing. == Personal life == Jurafsky is Jewish. He is married. They reside in San Francisco, California. == Selected works == 2009. Speech and Language Processing: An Introduction to Natural Language Processing, Computational Linguistics, and Speech Recognition, 2nd Edition. (with James H. Martin) Prentice-Hall. ISBN 978-0131873216 2014. The Language of Food: A Linguist Reads the Menu. W. W. Norton & Company. ISBN 978-0393240832 2026. Speech and Language Processing: An Introduction to Natural Language Processing, Computational Linguistics, and Speech Recognition, 3rd Edition draft. (with James H. Martin) == Honors and awards == 1998. NSF Career Award 2002. MacArthur Fellowship 2019. LSA Fellow 2022. Atkinson Prizes in Psychological and Cognitive Sciences
Lior Ron (business executive)
Lior Ron (born March 16, 1977) is an Israeli businessman. He is the founder, chairman and former CEO of logistics technology company Uber Freight, co-founder of self-driving truck company Otto, and COO of self-driving technology company Waabi. == Early life and education == Ron grew up in Israel near Haifa. He attended the Technion – Israel Institute of Technology in Haifa, where he earned a bachelor's degree in computer science in 1997. He then joined Israeli Army Intelligence, where he served until 2004. After the Army, he earned a master's degree in computer science at Technion, incorporating artificial intelligence as he developed a biomedical device to assist patients suffering with Parkinson's disease. He then moved to California and earned an MBA from The Stanford Graduate School of Business. His undergraduate work and master's thesis were centered around AI when it was still in its early stages. == Career == === Google === In 2007, Ron joined Google as the Product Lead for Google Maps. He then worked at Motorola Mobility after it was acquired by Google, and in Google's robotics research effort. === Otto === In 2016, Ron left Google to found Otto, a company that makes self-driving kits to retrofit big rig trucks. Quoted in Wired, Ron said he left Google because he “felt an obligation to bring this technology to society sooner rather than later.” Otto launched in May 2016, and was acquired by Uber in late July of the same year. The Uber partnership allowed Ron and Otto the opportunity to develop a freight marketplace for truck drivers. === Uber Freight === On May 18, 2017, Ron and Uber launched Uber Freight, a unit of Uber initially designed as an app connecting long-haul truck drivers with companies in need of cargo shipping, with Ron as CEO. In August 2018, Uber Freight launched a new digital platform focused on shippers, to help them find the right driver for their needs. In 2021, Uber Freight acquired Transplace for $2.25 billion, expanding its services to include managed transportation, logistics software, and consulting. With Ron as CEO, Uber Freight has evolved into a full-scale logistics technology company for shippers and drivers, as Ron introduced more advanced generative AI capabilities to Uber Freight's software and Insights AI logistics platform. In September 2024, the company announced it manages nearly $20 billion in freight, and serves one in three Fortune 500 companies. In May 2025, the company launched the transportation industry's first large-scale AI-powered logistics network, with its large language model embedded directly into its transportation management system. === Waabi === On August 12, 2025, it was reported that Ron had been named chief operating officer of Waabi, a company developing autonomous driving technology using artificial intelligence. He remains as chairman of Uber Freight, with Rebecca Tinucci taking over as CEO. == Controversy == Ron co-founded Otto with Anthony Levandowski, who faces a lawsuit brought in 2017 from Google's parent company Alphabet that alleges Levandowski stole trade secrets while working for Alphabet's self-driving car division before he and Ron co-founded Otto.
Universal IR Evaluation
In computer science, Universal IR Evaluation (information retrieval evaluation) aims to develop measures of database retrieval performance that shall be comparable across all information retrieval tasks. == Measures of "relevance" == IR (information retrieval) evaluation begins whenever a user submits a query (search term) to a database. If the user is able to determine the relevance of each document in the database (relevant or not relevant), then for each query, the complete set of documents is naturally divided into four distinct (mutually exclusive) subsets: relevant documents that are retrieved, not relevant documents that are retrieved, relevant documents that are not retrieved, and not relevant documents that are not retrieved. These four subsets (of documents) are denoted by the letters a, b, c, d respectively and are called Swets variables, named after their inventor. In addition to the Swets definitions, four relevance metrics have also been defined: Recall refers to the fraction of relevant documents that are retrieved (a/(a+b)), and Precision refers to the fraction of retrieved documents that are relevant (a/(a+c)). These are the most commonly used and well-known relevance metrics found in the IR evaluation literature. Two less commonly used metrics include the Fallout, i.e., the fraction of not relevant documents that are retrieved (b/(b+d)), and the Miss, which refers to the fraction of relevant documents that are not retrieved (c/(c+d)) during any given search. == Universal IR evaluation techniques == Universal IR evaluation addresses the mathematical possibilities and relationships among the four relevance metrics Precision, Recall, Fallout and Miss, denoted by P, R, F and M, respectively. One aspect of the problem involves finding a mathematical derivation of a complete set of universal IR evaluation points. The complete set of 16 points, each one a quadruple of the form (P, R, F, M), describes all the possible universal IR outcomes. For example, many of us have had the experience of querying a database and not retrieving any documents at all. In this case, the Precision would take on the undetermined form 0/0, the Recall and Fallout would both be zero, and the Miss would be any value greater than zero and less than one (assuming a mix of relevant and not relevant documents were in the database, none of which were retrieved). This universal IR evaluation point would thus be denoted by (0/0, 0, 0, M), which represents only one of the 16 possible universal IR outcomes. The mathematics of universal IR evaluation is a fairly new subject since the relevance metrics P, R, F, M were not analyzed collectively until recently (within the past decade). A lot of the theoretical groundwork has already been formulated, but new insights in this area await discovery.
Karen Spärck Jones
Karen Ida Boalth Spärck Jones (26 August 1935 – 4 April 2007) was a self-taught programmer and a pioneering British computer and information scientist responsible for the concept of inverse document frequency (IDF), a technology that underlies most modern search engines. She was an advocate for women in computer science, her slogan being, "Computing is too important to be left to men." In 2019, The New York Times published her belated obituary in its series Overlooked, calling her "a pioneer of computer science for work combining statistics and linguistics, and an advocate for women in the field." From 2008, to recognise her achievements in the fields of information retrieval (IR) and natural language processing (NLP), the Karen Spärck Jones Award is awarded annually to a recipient for outstanding research in one or both of her fields. == Early life and education == Karen Ida Boalth Spärck Jones was born in Huddersfield, Yorkshire, England. Her parents were Alfred Owen Jones, a chemistry lecturer, and Ida Spärck, a Norwegian who worked for the Norwegian government while in exile in London during World War II. Spärck Jones was educated at a grammar school in Huddersfield and then from 1953 to 1956 at Girton College, Cambridge, studying history, with an additional final year in Moral Sciences (philosophy). While at Cambridge, Spärck Jones joined the organisation known as the Cambridge Language Research Unit (CLRU) and met the head of CLRU Margaret Masterman, who would inspire her to go into computer science. While working at the CLRU, Spärck Jones began pursuing her PhD. At the time of submission, her PhD thesis was cast aside as uninspired and lacking original thought, but was later published in its entirety as a book. She briefly became a school teacher before moving into computer science. Spärck Jones married fellow Cambridge computer scientist Roger Needham in 1958. Spärck Jones's mother, Ida Spärck, had fled Norway on one of the last boats out after the German invasion in April 1940, going on to serve the Norwegian government in exile in London throughout the war. This background of displacement and resilience shaped the household in which Spärck Jones grew up. She later kept her mother's Norwegian surname professionally after marrying, stating that "it maintains a permanent existence of your own." Spärck Jones described her entry into computing as almost accidental. She had been working as a schoolteacher when she began visiting the CLRU out of curiosity about her husband's work. It was Margaret Masterman — whom she later described as "a very strange and interesting woman" — who offered her a research position and drew her fully into the field. == Career == Spärck Jones worked at the Cambridge Language Research Unit from the late 1950s, then at Cambridge University Computer Laboratory from 1974 until her retirement in 2002. From 1999, she held the post of Professor of Computers and Information. She had been given a permanent position only in 1993, and earlier in her career had been employed on a series of short-term contracts. She continued to work in the Computer Laboratory until shortly before her death. Her publications include nine books and numerous papers. A full list of her publications is available from the Cambridge Computer Laboratory. Spärck Jones' main research interests, since the late 1950s, were natural language processing and information retrieval. In 1964, Spärck Jones published "Synonymy and Semantic Classification", which is now seen as a foundational paper in the field of natural language processing. One of her most important contributions was the concept of inverse document frequency (IDF) weighting in information retrieval, which she introduced in a 1972 paper. IDF is used in most search engines today, usually as part of the term frequency–inverse document frequency (TF–IDF) weighting scheme. In the 1980s, Spärck Jones began her work on early speech recognition systems. In 1982 she became involved in the Alvey Programme which was an initiative to motivate more computer science research across the country. == Significance of inverse document frequency == At the time Spärck Jones was working, most computer scientists were focused on making people adapt to machines — learning precise codes and commands to retrieve information. Spärck Jones was working in the opposite direction: teaching computers to understand human language as it is actually used. Her 1972 paper introduced the concept of inverse document frequency (IDF) by observing that not all words carry equal informational value. A word like "the" appears in virtually every document and tells a retrieval system almost nothing about what any specific document is about. A rare word like "photosynthesis," by contrast, is highly specific and informative. IDF assigns each word a statistical weight based on how rarely it occurs across a document collection — the rarer the word, the higher its weight. When combined with term frequency (TF), which measures how often a word appears within a single document, the resulting TF–IDF score gives every word a relevance rating that can be used to rank documents in response to a search query. By 2007, Spärck Jones noted that "pretty much every web engine uses those principles." Her colleague John Tait remarked that "a lot of the stuff she was working on until five or ten years ago seemed like mad nonsense, and now we take it for granted." The 1972 paper remains among the most cited works in information retrieval research, with over 4,500 citations recorded in Google Scholar at the time of her death. The conceptual foundation of TF–IDF — that word meaning is statistical and contextual — has also informed later developments in machine learning and natural language processing, including transformer-based language models such as BERT. == Impact on artificial intelligence == Even though Spärck Jones' views on artificial intelligence (AI) were rather pessimistic in regard to the perceived limitations of AI in information retrieval, her work in natural language processing, information retrieval, and introducing the concept of inverse document frequency (IDF) contributed to the future technological development of AI. Her statistical and ranking methods shifted the direction of the development of AI towards being more expandable and led by data. Her work had a more indirect and conceptual impact on AI, compared to the current and direct impact it has had on search engines. == Gender and advocacy == Spärck Jones spent the majority of her career at Cambridge on short-term contracts without permanent employment, a situation she attributed directly to gender. In her 2001 IEEE oral history interview she stated that Cambridge was "in many ways not user-friendly, in the sense of women-friendly." She was frequently the only woman present in professional meetings throughout her career. She channelled this experience into active advocacy. She was a founding member of the women@cl network at Cambridge's Computer Laboratory, worked on outreach programmes aimed at encouraging girls into computing, and became widely known for her slogan: "Computing is too important to be left to men." She was the first woman ever to receive the BCS Lovelace Medal. === Honours and awards === These include: Gerard Salton Award (1988) Elected a Fellow of Association for the Advancement of Artificial Intelligence (AAAI) in 1993 President of the Association for Computational Linguistics (ACL) in 1994 Honorary degree of Doctor of Science from The City University in 1997. Elected a Fellow of the British Academy (FBA), where she also served as Vice-President in 2000–2002 Fellow of European Association for Artificial Intelligence (ECCAI) Association for Information Science and Technology (ASIS&T) Award of Merit (2002) Association for Computational Linguistics (ACL) Lifetime Achievement Award (2004) ACM - AAAI Allen Newell Award (2006) BCS Lovelace Medal (2007) Association for Computing Machinery (ACM) Women's Group Athena Award (2007) == Death and legacy == Spärck Jones died on 4 April 2007, due to cancer at the age of 71. In 2008, the BCS Information Retrieval Specialist Group (BCS IRSG) in conjunction with the British Computer Society established an annual Karen Spärck Jones Award in her honour, to encourage and promote research that advances understanding of Natural Language Processing or Information Retrieval. The Karen Spärck Jones lecture sponsored by BCS recognises the contribution that women have made to computing. In August 2017, the University of Huddersfield renamed one of its campus buildings in her honour. Formerly known as Canalside West, the Spärck Jones building houses the University's School of Computing and Engineering. When Spärck Jones died in 2007, The Times did not publish an obituary for her, despite having published one for her husband Roger Needham in 2003. In 2019, The New York Times included her in its Overlooked series under the title "Ove