Meta AI is a research division of Meta (formerly Facebook) that develops artificial intelligence and augmented reality technologies. == History == Meta AI was founded in 2013 as Facebook Artificial Intelligence Research (FAIR). It has workspaces in Menlo Park, London, New York City, Paris, Seattle, Pittsburgh, Tel Aviv, and Montreal as of 2025. In 2016, FAIR partnered with Google, Amazon, IBM, and Microsoft in creating the Partnership on Artificial Intelligence to Benefit People and Society. Meta AI was directed by Yann LeCun until 2018, when Jérôme Pesenti succeeded the role. Pesenti is formerly the CTO of IBM's big data group. FAIR's research includes self-supervised learning, generative adversarial networks, document classification and translation, and computer vision. FAIR released Torch deep-learning modules as well as PyTorch in 2017, an open-source machine learning framework, which was subsequently used in several deep learning technologies, such as Tesla's autopilot and Uber's Pyro. That same year, a pair of chatbots were falsely rumored to be discontinued for developing a language that was unintelligible to humans. FAIR clarified that the research had been shut down because they had accomplished their initial goal to understand how languages are generated by their models, rather than out of fear. FAIR was renamed Meta AI following the rebranding that changed Facebook, Inc. to Meta Platforms Inc. On October 1, 2025, Facebook announced "We will soon use your interactions with AI at Meta to personalize the content and ads you see". == Virtual assistant == Meta AI is also the name of the virtual assistant developed by the team, now integrated as a chatbot into Meta's social networking products. It is also available as a subscription-based stand-alone app. The virtual assistant was pre-installed on the second generation of Ray-Ban Meta smartglasses, and can incorporate inputs from the glasses' cameras after an update. It is also available on Quest 2 and newer HMDs. Since May 2024, the chatbot has summarized news from various outlets without linking directly to original articles, including in Canada, where news links are banned on its platforms. This use of news content without compensation and attribution has raised ethical and legal concerns, especially as Meta continues to reduce news visibility on its platforms. == Current research == === Natural language processing and chatbot === Natural language processing is the ability for machines to understand and generate natural language. The team is also researching unsupervised machine translation and multilingual chatbots. ==== Galactica ==== Galactica is a large language model (LLM) designed for generating scientific text. It was available for three days from 15 November 2022, before being withdrawn for generating racist and inaccurate content. ==== Llama ==== Llama is an LLM released in February 2023. As of January 2026, the most recent release is the Llama 4. === Hardware === Meta used CPUs and in-house custom chips before 2022; they switched to Nvidia GPUs since then. MTIA v1, one of their early chips, is designed for the company's content recommendation algorithms. It was fabricated on TSMC's 7 nm process technology and consumed 25W, capable of 51.2 TFlops FP16. == Controversy == The French media outlet Mediapart reports that in 2022, Facebook's parent company illegally used works accumulated by the pirate site LibGen to train its artificial intelligence.
Ware report
Security Controls for Computer Systems, commonly called the Ware report, is a 1970 text by Willis Ware that was foundational in the field of computer security. == Development == A defense contractor in St. Louis, Missouri, had bought an IBM mainframe computer, which it was using for classified work on a fighter aircraft. To provide additional income, the contractor asked the Department of Defense (DoD) for permission to sell computer time on the mainframe to local businesses via remote terminals, while the classified work continued. At the time, the DoD did not have a policy to cover this. The DoD's Advanced Research Projects Agency (DARPA) asked Ware - a RAND employee - to chair a committee to examine and report on the feasibility of security controls for computer systems. The committee's report was a classified document given in January 1970 to the Defense Science Board (DSB), which had taken over the project from ARPA. After declassification, the report was published by RAND in October 1979. == Influence == The IEEE Computer Society said the report was widely circulated, and the IEEE Annals of the History of Computing said that it, together with Ware's 1967 Spring Joint Computer Conference session, marked the start of the field of computer security. The report influenced security certification standards and processes, especially in the banking and defense industries, where the report was instrumental in creating the Orange Book.
New media
New media are communication technologies that enable or enhance interaction between users, as well as interaction between users and content. In the middle of the 1990s, the phrase "new media" became widely used as part of a sales pitch for the influx of interactive CD-ROMs for entertainment and education. The new media technologies, sometimes known as Web 2.0, include a wide range of web-related communication tools such as blogs, wikis, online social networking, virtual worlds, and other social media platforms. The phrase "new media" refers to computational media that share material online and through computers. New media inspire new ways of thinking about older media. Media do not replace one another in a clear, linear succession, instead evolving in a more complicated network of interconnected feedback loops . What is different about new media is how they specifically refashion traditional media and how older media refashion themselves to meet the challenges of new media. Unless they contain technologies that enable digital generative or interactive processes, broadcast television programs, non-interactive news websites, feature films, magazines, and books are not considered to be new media. The term "new media" stands in contrast to old media, which dominated the media landscape as a form of mass media for many years. == History == In the 1950s, connections between computing and radical art began to grow stronger. It was not until the 1980s that Alan Kay and his co-workers at Xerox PARC began to give the computability of a personal computer to the individual, rather than have a big organization be in charge of this. In the late 1980s and early 1990s, however, we seem to witness a different kind of parallel relationship between social changes and computer design. Although causally unrelated, conceptually, it makes sense that the Cold War and the design of the Web took place at exactly the same time. Writers and philosophers such as Marshall McLuhan were instrumental in the development of media theory during this period which is now famous declaration in Understanding Media: The Extensions of Man, that "the medium is the message" drew attention to the too often ignored influence media and technology themselves, rather than their "content," have on humans' experience of the world and on society broadly. Until the 1980s, media relied primarily upon print and analog broadcast models such as television and radio. The last twenty-five years have seen the rapid transformation into media which are predicated upon the use of digital technologies such as the Internet and video games. However, these examples are only a small representation of new media. The use of digital computers has transformed the remaining 'old' media, as suggested by the advent of digital television and online publications. Even traditional media forms such as the printing press have been transformed through the application of technologies by using of image manipulation software like Adobe Photoshop and desktop publishing tools. Andrew L. Shapiro argues that the "emergence of new, digital technologies signals a potentially radical shift of who is in control of information, experience and resources". W. Russell Neuman suggests that whilst the "new media" have technical capabilities to pull in one direction, economic and social forces pull back in the opposite direction. According to Neuman, "We are witnessing the evolution of a universal interconnected network of audio, video, and electronic text communications that will blur the distinction between interpersonal and mass communication; and between public and private communication". Neuman argues that new media will: Alter the meaning of geographic distance. Allow for a huge increase in the volume of communication. Provide the possibility of increasing the speed of communication. Provide opportunities for interactive communication. Allow forms of communication that were previously separate to overlap and interconnect. Consequently, it has been the contention of scholars such as Douglas Kellner and James Bohman that new media and particularly the Internet will provide the potential for a democratic postmodern public sphere, in which citizens can participate in well informed, non-hierarchical debate pertaining to their social structures. Contradicting these positive appraisals of the potential social impacts of new media are scholars such as Edward S. Herman and Robert McChesney who have suggested that the transition to new media has seen a handful of powerful transnational telecommunications corporations who achieve a level of global influence which was hitherto unimaginable. Scholars have highlighted both the positive and negative potential and actual implications of new media technologies, suggesting that some of the early work in new media studies was guilty of technologicaldeterminism – whereby the effects of media were determined by the technologies themselves, rather than by tracing the complex social networks that governed the development, funding, implementation, and future evolution of any technology. Based on the argument that people have a limited amount of time to spend on the consumption of different media, displacement theory argue that the viewership or readership of one particular outlet leads to the reduction in the amount of time spent by the individual on another. The introduction of new media, such as the internet, therefore reduces the amount of time individuals would spend on existing "old" media, which could ultimately lead to the end of such traditional media. == Definition == Although, there are several ways that new media may be described, Lev Manovich, in an introduction to The New Media Reader, defines new media by using eight propositions: New media versus cyberculture – Cyberculture is the various social phenomena that are associated with the Internet and network communications (blogs, online multi-player gaming), whereas new media is concerned more with cultural objects and paradigms (digital to analog television, smartphones). New media as computer technology used as a distribution platform – New media are the cultural objects which use digital computer technology for distribution and exhibition. e.g. (at least for now) Internet, Web sites, computer multimedia, Blu-ray disks etc. The problem with this is that the definition must be revised every few years. The term "new media" will not be "new" anymore, as most forms of culture will be distributed through computers. New media as digital data controlled by software – The language of new media is based on the assumption that, in fact, all cultural objects that rely on digital representation and computer-based delivery do share a number of common qualities. New media is reduced to digital data that can be manipulated by software as any other data. Now media operations can create several versions of the same object. An example is an image stored as matrix data which can be manipulated and altered according to the additional algorithms implemented, such as color inversion, gray-scaling, sharpening, rasterizing, etc. New media as the mix between existing cultural conventions and the conventions of software – New media today can be understood as the mix between older cultural conventions for data representation, access, and manipulation and newer conventions of data representation, access, and manipulation. The "old" data are representations of visual reality and human experience, and the "new" data is numerical data. The computer is kept out of the key "creative" decisions, and is delegated to the position of a technician. e.g. In film, software is used in some areas of production, in others are created using computer animation. New media as the aesthetics that accompanies the early stage of every new modern media and communication technology – While ideological tropes indeed seem to be reappearing rather regularly, many aesthetic strategies may reappear two or three times ... In order for this approach to be truly useful it would be insufficient to simply name the strategies and tropes and to record the moments of their appearance; instead, we would have to develop a much more comprehensive analysis which would correlate the history of technology with social, political, and economical histories or the modern period. New media as faster execution of algorithms previously executed manually or through other technologies – Computers are a huge speed-up of what were previously manual techniques. e.g. calculators. Dramatically speeding up the execution makes possible previously non-existent representational technique. This also makes possible of many new forms of media art such as interactive multimedia and video games. On one level, a modern digital computer is just a faster calculator, we should not ignore its other identity: that of a cybernetic control device. New media as the encoding of modernist avant-garde; new media as metamedia – Manovi
Robert Abel and Associates
Robert Abel and Associates (RA&A) was an American pioneering animation production company specializing in television commercials made with computer graphics. Founded by Robert Abel and Con Pederson in 1971, RA&A was especially known for their art direction and won many Clio Awards. Abel and his team created some of the most advanced and impressive computer-animated works of their time, including full ray-traced renders and fluid character animation at a time when such things were largely unknown. A variety of high-profile television advertisements, graphics sequences for motion pictures (including The Andromeda Strain and Tron), and work on laserdisc video games such as Cube Quest, put Abel and his team on the map in the early 1980s. The company was also originally commissioned to create the visual effects for Star Trek: The Motion Picture, but were subsequently taken off the project for mishandling funds. The company was also notable on its work for The Jacksons' 1981 music video "Can You Feel It." RA&A was on the southwest corner of Highland Avenue and Romaine in the heart of Hollywood, California. RA&A closed in 1987 following an ill-fated merger with now-defunct Omnibus Computer Graphics, Inc., a company which had been based in Toronto. Many people who worked at RA&A went on to other ground-breaking projects, including the founding of Wavefront Technologies, Rhythm & Hues and other studios. Many RA&A people went on to win Academy Awards.
Friending and following
Friending is the act of adding someone to a list of "friends" on a social networking service. The notion does not necessarily involve the concept of friendship. It is also distinct from the idea of a "fan"—as employed on the WWW sites of businesses, bands, artists, and others—since it is more than a one-way relationship. A "fan" only receives things. A "friend" can communicate back to the person friending. The act of "friending" someone usually grants that person special privileges (on the service) with respect to oneself. On Facebook, for example, one's "friends" have the privilege of viewing and posting to one's "timeline". Following is a similar concept on other social network services, such as Twitter and Instagram, where a person (follower) chooses to add content from a person or page to their newsfeed. Unlike friending, following is not necessarily mutual, and a person can unfollow (stop following) or block another user at any time without affecting that user's following status. The first scholarly definition and examination of friending and defriending (the act of removing someone from one's friend list, also called unfriending) was David Fono and Kate Raynes-Goldie's "Hyperfriendship and beyond: Friends and Social Norms on LiveJournal" from 2005, which identified the use of the term as both a noun and a verb by users of early social network site and blogging platform LiveJournal, which was originally launched in 1999. == Friend/follower count, friend collecting, and multiple accounts == The addition of people to a friend list without regard to whether one actually is their friend is sometimes known as friend whoring. Matt Jones of Dopplr went so far as to coin the expression "friending considered harmful" to describe the problem of focusing upon the friending of more and more people at the expense of actually making any use of a social network. Friend collecting is the adding of hundreds or thousands of friends/followers, a not uncommon order of magnitude on some social sites. As a result, many teen users feel pressured to heavily curate their posts, posting only carefully posed and edited photographs with well-thought-out captions. Some Instagram users will create a second account, known as a Finsta (short for "Fake Instagram"). A Finsta is typically private, and the owner only allows close friends to follow it. Since the follower count is kept down, the posts can be more candid and silly in nature. Users may also create multiple accounts based on their interests. Someone with a personal social media account might be a photographer and maintain a separate account for that. There is risk associated with following large numbers of people: scholars say that social anxiety could be an effect of managing a large social media network, as users can feel jealous and have a "fear of missing out". == Unfriending and unfollowing == Unfriending is the act of removing someone from a friends list. On Facebook, this means the action is unilateral, meaning, the friendship is terminated on both sides. The act of unfriending is often used when one user was flirting and made the other uncomfortable. Unfollowing is a little different. When a user unfollows someone on Instagram or Twitter, it continues a one-sided relationship. Often, the unfollowed user doesn't realize they were unfollowed, so they continue the following. == Social network friending and friendship == There are distinct groups of "friends" that one can friend on a social networking service. The notion of a social network friend does not necessarily embody the concept of friendship. Although terminology has not yet evolved to distinguish the different types of social networking friends, they can be broken into the following three categories. friends who are actually known These are people that may be one's friends or family in real life, with whom one has regular interaction either on-line or off-line. organizational friends These are companies and other organizations who maintain a "friending" relationship as a contacts list. complete strangers These are social networking "friends" with whom one has no relationship at all. Within these categories "friends" can be made up of strong ties, weak existing ties, weak latent ties, and parasocial ties. Strong ties can be made up of close family members and friends where self-disclosure, intimacy and frequent content occur. Weak existing ties can be made up of acquaintances, co-workers and distance relatives with whom the user has inconsistent contact. Weak latent ties can be made up of people within a similar geographical location or profession that can be used as a potential future bridge to other connections. Parasocial ties can be made up of celebrities, public figures and media personas. Human nature is to reciprocate a friending, marking someone as a friend who has marked oneself as a friend. This is a social norm for social networking services. However, this leads to mixing up who is an actual friend, and who is a contact. Tagging someone as a "contact" who has marked one as a "friend" can be perceived as impolite. Other concerns about this issue are treated in Sherry Turkle's Alone Together which analyses many behavioral dynamics in social media friendships. Turkle defines herself as "cautiously optimistic", but expresses concern that distance communications may undermine genuine face-to-face spoken discourses, lessening people's expectations of one another. One social networking service, FriendFeed, allows one to friend someone as a "fake" friend. The person "fake" friended receives the usual notifications for friending, but that person's updates are not received. Gavin Bell, author of Building Social Web Applications, describes this mechanism as "ludicrous". Results from a 2007 survey the Center for the Digital Future stated that only 23% of internet users have at least one virtual friend whom they have only met online. Ideally the number of virtual friends is directly proportional to the use of the Internet, but the same survey showed 20% of heavy-users (more than 3 hours/day) who claimed an average of 8.7% online friends, reported at least one relationship that started virtually and migrated to in-person contact. This results and other concerning issues are included in the book Networked: The New Social Operating System co-written by Lee Rainie and Barry Wellman in 2012. == Ethical considerations == The act of "friending" someone on a social networking service has particular ethical implications for judges in the United States. Judicial codes of conducts in the various states generally incorporate some form of provision that judges should avoid even the appearance of impropriety. Whether this regulates and even prohibits judges "friending" attorneys that appear before them, and law enforcement personnel, has been the subject of some analysis by the judicial ethics panels of the various states. They haven't all agreed on the guidance that they have given to judges: The New York state Judicial Ethics committee in 2009 simply advised judges to employ caution, noting that the issue of "friending" someone on a social networking service is a publicly observable act that has little difference from other public behavior concerns judges already face. The Florida Judicial Ethics Advisory committee in 2009 noted that, judges being normal human beings, it was unavoidable for judges to form friendships without the responsibilities of their job. It prohibited judges from friending any attorneys that appeared before them, whilst allowing friending of those who do not, on the grounds that it may give the appearance to the general public (even if the substance is otherwise) that those attorneys who are friended hold special sway with the judge. A minority opinion of the committee asserted that there is a substantive difference between "friending" on a social networking service and actual friendship, and that the general public, being aware of the norms of social networking services, was capable of drawing this distinction and would not reasonably conclude either a special degree of influence or a violation of the code of judicial conduct. This minority opinion was outnumbered twice in 2009, both in the Judicial Ethics Advisory and in the Florida Supreme Court Judicial Ethics Advisory committee. The South Carolina judicial conduct committee in 2009 permitted judges to friend attorneys and law enforcement personnel, with the proviso that no judicial business should be conducted upon nor discussed via the social networking service. "... a judge should not become isolated from the community in which the judge lives.", the committee stated. The Kentucky Judicial Ethics committee in 2010 took the same position as the minority opinion in Florida. It urged judges to exercise caution, but recognized that the act of friending "does not, in and of itself, indicate the degree or intensity of a judge's relationship with the person who is the 'friend'
Zero-shot learning
Zero-shot learning (ZSL) is a problem setup in deep learning where, at test time, a learner observes samples from classes which were not observed during training, and needs to predict the class that they belong to. The name is a play on words based on the earlier concept of one-shot learning, in which classification can be learned from only one, or a few, examples. Zero-shot methods generally work by associating observed and non-observed classes through some form of auxiliary information, which encodes observable distinguishing properties of objects. For example, given a set of images of animals to be classified, along with auxiliary textual descriptions of what animals look like, an artificial intelligence model which has been trained to recognize horses, but has never been given a zebra, can still recognize a zebra when it also knows that zebras look like striped horses. This problem is widely studied in computer vision, natural language processing, and machine perception. == Background and history == The first paper on zero-shot learning in natural language processing appeared in a 2008 paper by Chang, Ratinov, Roth, and Srikumar, at the AAAI'08, but the name given to the learning paradigm there was dataless classification. The first paper on zero-shot learning in computer vision appeared at the same conference, under the name zero-data learning. The term zero-shot learning itself first appeared in the literature in a 2009 paper from Palatucci, Hinton, Pomerleau, and Mitchell at NIPS'09. This terminology was repeated later in another computer vision paper and the term zero-shot learning caught on, as a take-off on one-shot learning that was introduced in computer vision years earlier. In computer vision, zero-shot learning models learned parameters for seen classes along with their class representations and rely on representational similarity among class labels so that, during inference, instances can be classified into new classes. In natural language processing, the key technical direction developed builds on the ability to "understand the labels"—represent the labels in the same semantic space as that of the documents to be classified. This supports the classification of a single example without observing any annotated data, the purest form of zero-shot classification. The original paper made use of the Explicit Semantic Analysis (ESA) representation but later papers made use of other representations, including dense representations. This approach was also extended to multilingual domains, fine entity typing and other problems. Moreover, beyond relying solely on representations, the computational approach has been extended to depend on transfer from other tasks, such as textual entailment and question answering. The original paper also points out that, beyond the ability to classify a single example, when a collection of examples is given, with the assumption that they come from the same distribution, it is possible to bootstrap the performance in a semi-supervised like manner (or transductive learning). Unlike standard generalization in machine learning, where classifiers are expected to correctly classify new samples to classes they have already observed during training, in ZSL, no samples from the classes have been given during training the classifier. It can therefore be viewed as an extreme case of domain adaptation. == Prerequisite information for zero-shot classes == Naturally, some form of auxiliary information has to be given about these zero-shot classes, and this type of information can be of several types. Learning with attributes: classes are accompanied by pre-defined structured description. For example, for bird descriptions, this could include "red head", "long beak". These attributes are often organized in a structured compositional way, and taking that structure into account improves learning. While this approach was used mostly in computer vision, there are some examples for it also in natural language processing. Learning from textual description. As pointed out above, this has been the key direction pursued in natural language processing. Here class labels are taken to have a meaning and are often augmented with definitions or free-text natural-language description. This could include for example a wikipedia description of the class. Class-class similarity. Here, classes are embedded in a continuous space. A zero-shot classifier can predict that a sample corresponds to some position in that space, and the nearest embedded class is used as a predicted class, even if no such samples were observed during training. == Generalized zero-shot learning == The above ZSL setup assumes that at test time, only zero-shot samples are given, namely, samples from new unseen classes. In generalized zero-shot learning, samples from both new and known classes, may appear at test time. This poses new challenges for classifiers at test time, because it is very challenging to estimate if a given sample is new or known. Some approaches to handle this include: a gating module, which is first trained to decide if a given sample comes from a new class or from an old one, and then, at inference time, outputs either a hard decision, or a soft probabilistic decision a generative module, which is trained to generate feature representation of the unseen classes—a standard classifier can then be trained on samples from all classes, seen and unseen. == Domains of application == Zero shot learning has been applied to the following fields: image classification semantic segmentation image generation object detection natural language processing computational biology abstract reasoning
Style sheet (web development)
A web style sheet is a form of separation of content and presentation for web design in which the markup (i.e., HTML or XHTML) of a webpage contains the page's semantic content and structure, but does not define its visual layout (style). Instead, the style is defined in an external style sheet file using a style sheet language such as CSS or XSLT. This design approach is identified as a "separation" because it largely supersedes the antecedent methodology in which a page's markup defined both style and structure. The philosophy underlying this methodology is a specific case of separation of concerns. == Benefits == Separation of style and content has advantages, but has only become practical after improvements in popular web browsers' CSS implementations. === Speed === Overall, users experience of a site utilising style sheets will generally be quicker than sites that do not use the technology. ‘Overall’ as the first page will probably load more slowly – because the style sheet AND the content will need to be transferred. Subsequent pages will load faster because no style information will need to be downloaded – the CSS file will already be in the browser’s cache. === Maintainability === Holding all the presentation styles in one file can reduce the maintenance time and reduces the chance of error, thereby improving presentation consistency. For example, the font color associated with a type of text element may be specified — and therefore easily modified — throughout an entire website simply by changing one short string of characters in a single file. The alternative approach, using styles embedded in each individual page, would require a cumbersome, time consuming, and error-prone edit of every file. === Accessibility === Sites that use CSS with either XHTML or HTML are easier to tweak so that they appear similar in different browsers (Chrome, Internet Explorer, Mozilla Firefox, Opera, Safari, etc.). Sites using CSS "degrade gracefully" in browsers unable to display graphical content, such as Lynx, or those so very old that they cannot use CSS. Browsers ignore CSS that they do not understand, such as CSS 3 statements. This enables a wide variety of user agents to be able to access the content of a site even if they cannot render the style sheet or are not designed with graphical capability in mind. For example, a browser using a refreshable braille display for output could disregard layout information entirely, and the user would still have access to all page content. === Customization === If a page's layout information is stored externally, a user can decide to disable the layout information entirely, leaving the site's bare content still in a readable form. Site authors may also offer multiple style sheets, which can be used to completely change the appearance of the site without altering any of its content. Most modern web browsers also allow the user to define their own style sheet, which can include rules that override the author's layout rules. This allows users, for example, to bold every hyperlink on every page they visit. Browser extensions like Stylish and Stylus have been created to facilitate management of such user style sheets. === Consistency === Because the semantic file contains only the meanings an author intends to convey, the styling of the various elements of the document's content is very consistent. For example, headings, emphasized text, lists and mathematical expressions all receive consistently applied style properties from the external style sheet. Authors need not concern themselves with the style properties at the time of composition. These presentational details can be deferred until the moment of presentation. === Portability === The deferment of presentational details until the time of presentation means that a document can be easily re-purposed for an entirely different presentation medium with merely the application of a new style sheet already prepared for the new medium and consistent with elemental or structural vocabulary of the semantic document. A carefully authored document for a web page can easily be printed to a hard-bound volume complete with headers and footers, page numbers and a generated table of contents simply by applying a new style sheet.