Galatea is an interactive fiction video game by Emily Short featuring a modern rendition of the Greek myth of Galatea, the sculpture of a woman that gained life. It took "Best of Show" in the 2000 IF Art Show and won a XYZZY Award for Best non-player character. The game displays an unusually rich approach to non-player character dialogue and diverts from the typical puzzle-solving in interactive fiction: gameplay consists entirely of interacting with a single character in a single room. Galatea is licensed under the Creative Commons BY-NC-ND 3.0 US license. == Gameplay == Galatea alters the typical interactive fiction game mechanics by concentrating instead on the player's interactions with a single non-player character (NPC), the eponymous Galatea. Much of the interest of the piece derives from the ambiguous nature of the player–NPC dialogue: the form of the conversation and, indeed, the nature of Galatea herself shift depending on the focus the player places on certain aspects of the character's personality. Numerous endings are possible. Gameplay centers around the developing dialogue between Galatea and the player when asking about topics in the previous conversation. Two commands, "think about" and "recap", are provided to keep track of what has already been said; the former is also used to advance the storyline, as the player character draws conclusions about the story as it has unfolded to that point. The game also encourages using sensory commands ("touch", "listen to", "look at"), adding immersion to the experience. == Plot == Galatea is loosely based on the myth of Pygmalion, who carved the sculpture of a woman. In the myth, he falls in love with the statue, named Galatea or Elise in different versions, and the goddess Venus brings her to life. The story begins at the opening of an exhibition of artificial intelligences. The player, alone, discovers Galatea displayed on a pedestal with a small information placard. She is illuminated by a spotlight and wears an emerald dress. Seeing the player about to turn away, Galatea says, "They told me you were coming." From this point, the story may proceed in a number of ways depending on the player's words and actions. === Multilinear interactive fiction === Short describes this as "multilinear interactive fiction": while interactive fiction in general allows the player to find their own way through the story, this leads in most cases to a single ending (or at least a single desired 'correct' ending). With Galatea, Short presents a story with around 70 different endings and hundreds of possible ways of reaching them. The plot is thus designed to appear open-ended with the development of the story entirely dependent on what the player decides to talk or ask about or what actions they choose to perform. Thus the original author and the player share in the creation of a work of fiction. == Development == In interviews, Emily Short has explained that Galatea arose out of her efforts to develop advanced dialog coding for interactive fiction engines. Although code for simple conversational programs like ELIZA have existed since the 1960s, and limited dialog options have existed in interactive fiction since the 1970s, Short's efforts to develop chatterbot-like dialog required her to produce a simple test case scenario to test NPC interaction. Thus the single-room, single-occupant Galatea was a natural result. Development of the game progressed organically with Short engaging in test runs and drafting new dialog options for every conversational dead-end that arose. The game's multiple endings also arose in a similar fashion although Short had intended that there be multiple endings from the start. Although the nature of the game's development as well as its minimalist final form has led to questions regarding whether it is really a game and not just an experimental conversational program, Short has suggested that to her the definition of interactive fiction requires nothing more than a world model and a parser, and "anything you can cook up with those features counts as IF." Short has acknowledged the helpful influence of the close-knit IF community and the "atmosphere in which experimentation is valued" as leading to the success of her works like Galatea. == Reception == Galatea was well received, achieving critical acclaim from interactive fiction reviewers and literary scholars. The game is considered to aspire to a new level of art in interactive fiction, and thereby to have revolutionized the genre, establishing its author, Emily Short, as one of the key figures in the modern interactive fiction scene. Fellow award-winning IF author, Adam Cadre has called Galatea "the best NPC ever"—a view that was echoed by Joystiq's John Bardinelli. Cadre also describes the game as an example of an alternative kind of puzzle where "interactivity comes in deciding where to go, what to see, what to say. Rather than having to open gates along a path, you discover that they're all open at first, but stepping through one causes others to close." Galatea was described in 2007 by Indiegames.com as a "fascinating journey." In a 2009 article, Rock, Paper, Shotgun praised the depth and detail of the game, the complexities of the character design and its "masterful balance between intricacy and simplicity", and "Galatea's emotional turmoil" that is "encoded sweetly into the subtext of what's going on. By simply interacting in a logical manner, you learn more about this character than any cut-scene or info-dump could ever hope to convey." This was reiterated in a 2010 1UP.com article that listed Galatea as #2 in its "Top 5 Introductory Interactive Fiction Games" feature, describing it as intriguingly replayable, and as a "surprisingly rich game for its apparent minimalism". In 2011, PC Gamer highlighted Galatea as an example of the artistic and literary aspects of the interactive fiction genre. The titular character, Galatea, has been compared to the 2007 Portal character GLaDOS due to similarities in the personalities of the characters.
Matchbox Educable Noughts and Crosses Engine
The Matchbox Educable Noughts and Crosses Engine (sometimes called the Machine Educable Noughts and Crosses Engine or MENACE) was a mechanical computer made from 304 matchboxes designed and built by artificial intelligence researcher Donald Michie and his colleague Roger Chambers, in 1961. It was designed to play human opponents in games of noughts and crosses (tic-tac-toe) by returning a move for any given state of play and to refine its strategy through reinforcement learning. This was one of the first types of artificial intelligence. Michie and Chambers did not have immediate access to a computer; they worked around this by building the engine out of matchboxes. The matchboxes they used each represented a single possible layout of a noughts and crosses grid. When the computer first played, it would randomly choose moves based on the current layout. As it played more games, through a reinforcement loop, it disqualified strategies that led to losing games, and supplemented strategies that led to winning games. Michie held a tournament against MENACE in 1961, wherein he experimented with different openings. Following MENACE's maiden tournament against Michie, it demonstrated successful artificial intelligence in its strategy. Michie's essays on MENACE's weight initialisation and the BOXES algorithm used by MENACE became popular in the field of computer science research. Michie was honoured for his contribution to machine learning research, and was twice commissioned to program a MENACE simulation on an actual computer. == Origin == Donald Michie (1923–2007) had been on the team decrypting the German Tunny Code during World War II. Fifteen years later, he wanted to further display his mathematical and computational prowess with an early convolutional neural network. Since computer equipment was not obtainable for such uses, and Michie did not have a computer readily available, he decided to display and demonstrate artificial intelligence in a more esoteric format and constructed a functional mechanical computer out of matchboxes and beads. MENACE was constructed as the result of a bet with a computer science colleague who postulated that such a machine was impossible. Michie undertook the task of collecting and defining each matchbox as a "fun project", later turned into a demonstration tool. Michie completed his essay on MENACE in 1963, "Experiments on the mechanization of game-learning", as well as his essay on the BOXES Algorithm, written with R. A. Chambers and had built up an AI research unit in Hope Park Square, Edinburgh, Scotland. MENACE learned by playing successive matches of noughts and crosses. Each time, it would eliminate a losing strategy by the human player confiscating the beads that corresponded to each move. It reinforced winning strategies by making the moves more likely, by supplying extra beads. This was one of the earliest versions of the Reinforcement Loop, the schematic algorithm of looping the algorithm, dropping unsuccessful strategies until only the winning ones remain. This model starts as completely random, and gradually learns. == Composition == MENACE was made from 304 matchboxes glued together in an arrangement similar to a chest of drawers. Each box had a code number, which was keyed into a chart. This chart had drawings of tic-tac-toe game grids with various configurations of X, O, and empty squares, corresponding to all possible permutations a game could go through as it progressed. After removing duplicate arrangements (ones that were simply rotations or mirror images of other configurations), MENACE used 304 permutations in its chart and thus that many matchboxes. Each individual matchbox tray contained a collection of coloured beads. Each colour represented a move on a square on the game grid, and so matchboxes with arrangements where positions on the grid were already taken would not have beads for that position. Additionally, at the front of the tray were two extra pieces of card in a "V" shape, the point of the "V" pointing at the front of the matchbox. Michie and his artificial intelligence team called MENACE's algorithm "Boxes", after the apparatus used for the machine. The first stage "Boxes" operated in five phases, each setting a definition and a precedent for the rules of the algorithm in relation to the game. == Operation == MENACE played first, as O, since all matchboxes represented permutations only relevant to the "X" player. To retrieve MENACE's choice of move, the opponent or operator located the matchbox that matched the current game state, or a rotation or mirror image of it. For example, at the start of a game, this would be the matchbox for an empty grid. The tray would be removed and lightly shaken so as to move the beads around. Then, the bead that had rolled into the point of the "V" shape at the front of the tray was the move MENACE had chosen to make. Its colour was then used as the position to play on, and, after accounting for any rotations or flips needed based on the chosen matchbox configuration's relation to the current grid, the O would be placed on that square. Then the player performed their move, the new state was located, a new move selected, and so on, until the game was finished. When the game had finished, the human player observed the game's outcome. As a game was played, each matchbox that was used for MENACE's turn had its tray returned to it ajar, and the bead used kept aside, so that MENACE's choice of moves and the game states they belonged to were recorded. Michie described his reinforcement system with "reward" and "punishment". Once the game was finished, if MENACE had won, it would then receive a "reward" for its victory. The removed beads showed the sequence of the winning moves. These were returned to their respective trays, easily identifiable since they were slightly open, as well as three bonus beads of the same colour. In this way, in future games MENACE would become more likely to repeat those winning moves, reinforcing winning strategies. If it lost, the removed beads were not returned, "punishing" MENACE, and meaning that in future it would be less likely, and eventually incapable if that colour of bead became absent, to repeat the moves that cause a loss. If the game was a draw, one additional bead was added to each box. == Results in practice == === Optimal strategy === Noughts and crosses has a well-known optimal strategy. A player must place their symbol in a way that blocks the other player from achieving any rows while simultaneously making a row themself. However, if both players use this strategy, the game always ends in a draw. If the human player is familiar with the optimal strategy, and MENACE can quickly learn it, then the games will eventually only end in draws. The likelihood of the computer winning increases quickly when the computer plays against a random-playing opponent. When playing against a player using optimal strategy, the odds of a draw grow to 100%. In Donald Michie's official tournament against MENACE in 1961 he used optimal strategy, and he and the computer began to draw consistently after twenty games. Michie's tournament had the following milestones: Michie began by consistently opening with "Variant 0", the middle square. At 15 games, MENACE abandoned all non-corner openings. At just over 20, Michie switched to consistently using "Variant 1", the bottom-right square. At 60, he returned to Variant 0. As he neared 80 games, he moved to "Variant 2", the top-middle. At 110, he switched to "Variant 3", the top right. At 135, he switched to "Variant 4", middle-right. At 190, he returned to Variant 1, and at 210, he returned to Variant 0. The trend in changes of beads in the "2" boxes runs: === Correlation === Depending on the strategy employed by the human player, MENACE produces a different trend on scatter graphs of wins. Using a random turn from the human player results in an almost-perfect positive trend. Playing the optimal strategy returns a slightly slower increase. The reinforcement does not create a perfect standard of wins; the algorithm will draw random uncertain conclusions each time. After the j-th round, the correlation of near-perfect play runs: 1 − D D − D ( j + 2 ) ∑ i = 0 j D ( j i + 1 ) V i {\displaystyle {1-D \over D-D^{(j+2)}}\sum _{i=0}^{j}D^{(ji+1)}V_{i}} Where Vi is the outcome (+1 is win, 0 is draw and -1 is loss) and D is the decay factor (average of past values of wins and losses). Below, Mn is the multiplier for the n-th round of the game. == Legacy == Donald Michie's MENACE proved that a computer could learn from failure and success to become good at a task. It used what would become core principles within the field of machine learning before they had been properly theorised. For example, the combination of how MENACE starts with equal numbers of types of beads in each matchbox, and how these are then selected at random, creates a learning behaviour similar to weight initialisation
Localhost
In computer networking, localhost is a hostname that refers to the current computer used to access it. The name localhost is reserved for loopback purposes. It is used to access the network services that are running on the host via the loopback network interface. Using the loopback interface bypasses any local network interface hardware. == Loopback == The local loopback mechanism may be used to run a network service on a host without requiring a physical network interface, or without making the service accessible from the networks the computer may be connected to. For example, a locally installed website may be accessed from a Web browser by the URL http://localhost to display its home page. IPv4 network standards reserve the entire address block 127.0.0.0/8 (more than 16 million addresses) for loopback purposes. That means any packet sent to any of those addresses is looped back. The address 127.0.0.1 is the standard address for IPv4 loopback traffic; the rest are not supported by all operating systems. However, they can be used to set up multiple server applications on the host, all listening on the same port number. In the IPv6 addressing architecture there is only a single address assigned for loopback: ::1. The standard precludes the assignment of that address to any physical interface, as well as its use as the source or destination address in any packet sent to remote hosts. == Name resolution == The name localhost normally resolves to the IPv4 loopback address 127.0.0.1, and to the IPv6 loopback address ::1. This resolution is normally configured by the following lines in the operating system's hosts file: 127.0.0.1 localhost ::1 localhost The name may also be resolved by Domain Name System (DNS) servers, but there are special considerations governing the use of this name: An IPv4 or IPv6 address query for the name localhost must always resolve to the respective loopback address. Applications may resolve the name to a loopback address themselves, or pass it to the local name resolver mechanisms. When a name resolver receives an address (A or AAAA) query for localhost, it should return the appropriate loopback addresses, and negative responses for any other requested record types. Queries for localhost should not be sent to caching name servers. To avoid burdening the Domain Name System root servers with traffic, caching name servers should never request name server records for localhost, or forward resolution to authoritative name servers. When authoritative name servers receive queries for 'localhost' in spite of the provisions mentioned above, they should resolve them appropriately. In addition to the mapping of localhost to the loopback addresses (127.0.0.1 and ::1), localhost may also be mapped to other IPv4 (loopback) addresses and it is also possible to assign other, or additional, names to any loopback address. The mapping of localhost to addresses other than the designated loopback address range in the hosts file or in DNS is not guaranteed to have the desired effect, as applications may map the name internally. In the Domain Name System, the name .localhost is reserved as a top-level domain name, originally set aside to avoid confusion with the hostname localhost. Domain name registrars are precluded from delegating domain names in the top-level .localhost domain. == Historical notes == In 1981, the block 127.0.0.0/8 got a 'reserved' status, as not to assign it as a general purpose class A IP network. This block was officially assigned for loopback purposes in 1986. Its purpose as a Special Use IPv4 Address block was confirmed in 1994,, 2002, 2010,, and last in 2013. From the outset, in 1995, the single IPv6 loopback address ::1 was defined. Its purpose and definition was unchanged in 1998,, 2003,, and up to the current definition, in 2006. == Packet processing == The processing of any packet sent to a loopback address, is implemented in the link layer of the TCP/IP stack. Such packets are never passed to any network interface controller (NIC) or hardware device driver and must not appear outside of a computing system, or be routed by any router. This permits software testing and local services, even in the absence of any hardware network interfaces. Looped-back packets are distinguished from any other packets traversing the TCP/IP stack only by the special IP address they were addressed to. Thus, the services that ultimately receive them respond according to the specified destination. For example, an HTTP service could route packets addressed to 127.0.0.99:80 and 127.0.0.100:80 to different Web servers, or to a single server that returns different web pages. To simplify such testing, the hosts file may be configured to provide appropriate names for each address. Packets received on a non-loopback interface with a loopback source or destination address must be dropped. Such packets are sometimes referred to as Martian packets. As with any other bogus packets, they may be malicious and any problems they might cause can be avoided by applying bogon filtering. == Special cases == The releases of the MySQL database differentiate between the use of the hostname localhost and the use of the addresses 127.0.0.1 and ::1. When using localhost as the destination in a client connector interface of an application, the MySQL application programming interface connects to the database using a Unix domain socket, while a TCP connection via the loopback interface requires the direct use of the explicit address. One notable exception to the use of the 127.0.0.0/8 addresses is their use in Multiprotocol Label Switching (MPLS) traceroute error detection, in which their property of not being routable provides a convenient means to avoid delivery of faulty packets to end users.
Storyful
Storyful (stylized as storyful.) is a social media intelligence company headquartered in Dublin, Ireland that is a subsidiary of News Corp, offering services such as social news monitoring, video licensing, and reputation risk management tools for corporate clients. The startup was launched as the first social media newswire, a content aggregator, verifying news sources and online content in Dublin in 2010 by Mark Little, a former journalist with RTÉ News. Storyful was acquired by News Corp in 2013 for USD$25 million. == Background == Mark Little, who had worked as a television journalist for RTÉ One, founded startup Storyful in Dublin, Ireland, in 2010, as a service that "verified news sources and online content". According to Nieman Lab, Storyful had a reputation for content aggregation as a social news agency—finding, verifying, distributing, licensing, and commercializing user-generated content, social media and online content from social networking services, including videos about stories in the news, such as the Syrian Civil War, Arab Spring protests, as well as "smaller viral moments". Storyful aimed to provide authority through its verification and monitoring tools while providing authenticity through user-generated content. On 20 December 2013 News Corp purchased Storyful for US$25 million and opened a New York office in the same building as Fox News' main studios. Little left Storyful in 2015 and Gavin Sheridan, Storyful's director of innovation left in 2014. News Corp CEO Robert Thomson said that through Storyful, News Corp would "define the opportunities that the digital landscape presents, rather than simply adapt to them." After the acquisition, the company expanded its service to include "commercial and creative work". After Murdoch acquired the company, from 2014 through to February 2018, losses "swelled", requiring a series of cash injections from News Corp. During that time the company expanded aggressively globally with a staff of about 200 worldwide up from about 30 in 2014. According to The Guardian, in 2016, journalists were encouraged by Storyful to use the social media monitoring software called Verify developed by Storyful. By installing Verify's web browser extension on their computers, Verify would inform the journalists when social media content had been "verified and cleared". The Guardian revealed that through the Verify plugin, dozens of staff in four offices had access to the journalists browsing activity without them knowing. This data allowed Storyful to actively monitor its own clients' activities on social media and to "turn it into an internal feed" at Storyful that "updates in real time". In November 2018, when a video circulated by Infowars' Paul Joseph Watson appeared to prove that CNN's Jim Acosta's contact with a White House intern was a physical blow, Storyful was able to prove that the 15-second-long clip had been doctored. According to a 21 January 2019 article in CNN Business, Rob McDonagh, the editor of Storyful's U.S. news team, had proven that one of the viral videos that served as catalysts in the January 2019 Lincoln Memorial confrontation at 18 January 2019 Indigenous Peoples March, was posted by a suspicious account, under the handle @2020fight. McDonagh's team validates videos and posts before adding them to their "digest", distinguishing true stories from those that are not. Storyful attempts to validate each post or video before including it in its digest. McDonagh reviewed previous content from @2020fight's account, and found it suspicious because it had a high follower count, a "highly polarized and yet inconsistent political messaging", an "unusually high rate of tweets", and "the use of someone else's image in the profile photo." reporter Donie O'Sullivan said that the @2020fight video that had been posted on 18 January, which had 2.5 million views by 22 January, was the one that "helped frame the news cycle". Currently the website offers a service by which video can be commercially brokered. == Services == Services include a newswire service—one of their "core pillars"—and social news monitoring. By February 2018, Storyful was developing "risk and reputation monitoring" services through which they would source and verify social news, fact-checking it and contextualising it for corporate clients. They were "developing tech tools" to "explore obscure or closed networks" for their intelligence team. can use to explore obscure or closed networks. They "track deviations in social conversations around brands and organisations and catch potential risks before they blow up. Like an alerts system." The company "released a re-booted version of its Newswire platform in 2018. According to FORA, Storyful was developing new tools to combat fake news online. == Clients == When Storyful was acquired by News Corp in 2013, the company already had the Wall Street Journal, the BBC, New York Times, YouTube, ITN and Channel 4 News as clients. By 2018 their clients included CNN, ABC News and Fox News, The New York Times, the Washington Post, in the United States, the Australian Broadcasting Corporation and all of News Corp’s own publications. Most of their "reputation-conscious corporate customers" clients prefer to not be named.
Information leakage
Information leakage happens whenever a system that is designed to be closed to an eavesdropper reveals some information to unauthorized parties nonetheless. In other words: Information leakage occurs when secret information correlates with, or can be correlated with, observable information. For example, when designing an encrypted instant messaging network, a network engineer without the capacity to crack encryption codes could see when messages are transmitted, even if he could not read them. == Risk vectors == A modern example of information leakage is the leakage of secret information via data compression, by using variations in data compression ratio to reveal correlations between known (or deliberately injected) plaintext and secret data combined in a single compressed stream. Another example is the key leakage that can occur when using some public-key systems when cryptographic nonce values used in signing operations are insufficiently random. Bad randomness cannot protect proper functioning of a cryptographic system, even in a benign circumstance, it can easily produce crackable keys that cause key leakage. Information leakage can sometimes be deliberate: for example, an algorithmic converter may be shipped that intentionally leaks small amounts of information, in order to provide its creator with the ability to intercept the users' messages, while still allowing the user to maintain an illusion that the system is secure. This sort of deliberate leakage is sometimes known as a subliminal channel. Generally, only very advanced systems employ defenses against information leakage. Following are the commonly implemented countermeasures : Use steganography to hide the fact that a message is transmitted at all. Use chaffing to make it unclear to whom messages are transmitted (but this does not hide from others the fact that messages are transmitted). For busy re-transmitting proxies, such as a Mixmaster node: randomly delay and shuffle the order of outbound packets - this will assist in disguising a given message's path, especially if there are multiple, popular forwarding nodes, such as are employed with Mixmaster mail forwarding. When a data value is no longer going to be used, erase it from the memory.
Language model benchmark
A language model benchmark is a standardized test designed to evaluate the performance of language models on various natural language processing tasks. These tests are intended for comparing different models' capabilities in areas such as language understanding, generation, and reasoning. Benchmarks generally consist of a dataset and corresponding evaluation metrics. The dataset provides text samples and annotations, while the metrics measure a model's performance on tasks like answering questions, text classification, and machine translation. These benchmarks are developed and maintained by academic institutions, research organizations, and industry players to track progress in the field. In addition to accuracy, the metrics can include throughput, energy efficiency, bias, trust, and sustainability. == Overview == === Types === Benchmarks may be described by the following adjectives, not mutually exclusive: Classical: These tasks are studied in natural language processing, even before the advent of deep learning. Examples include the Penn Treebank for testing syntactic and semantic parsing, as well as bilingual translation benchmarked by BLEU scores. Question answering: These tasks have a text question and a text answer, often multiple-choice. They can be open-book or closed-book. Open-book QA resembles reading comprehension questions, with relevant passages included as annotation in the question, in which the answer appears. Closed-book QA includes no relevant passages. Closed-book QA is also called open-domain question-answering. Before the era of large language models, open-book QA was more common, and understood as testing information retrieval methods. Closed-book QA became common since GPT-2 as a method to measure knowledge stored within model parameters. Omnibus: An omnibus benchmark combines many benchmarks, often previously published. It is intended as an all-in-one benchmarking solution. Reasoning: These tasks are usually in the question-answering format, but are intended to be more difficult than standard question answering. Multimodal: These tasks require processing not only text, but also other modalities, such as images and sound. Examples include OCR and transcription. Agency: These tasks are for a language-model–based software agent that operates a computer for a user, such as editing images, browsing the web, etc. Adversarial: A benchmark is "adversarial" if the items in the benchmark are picked specifically so that certain models do badly on them. Adversarial benchmarks are often constructed after state of the art (SOTA) models have saturated (achieved 100% performance) a benchmark, to renew the benchmark. A benchmark is "adversarial" only at a certain moment in time, since what is adversarial may cease to be adversarial as newer SOTA models appear. Public/Private: A benchmark might be partly or entirely private, meaning that some or all of the questions are not publicly available. The idea is that if a question is publicly available, then it might be used for training, which would be "training on the test set" and invalidate the result of the benchmark. Usually, only the guardians of the benchmark have access to the private subsets, and to score a model on such a benchmark, one must send the model weights, or provide API access, to the guardians. The boundary between a benchmark and a dataset is not sharp. Generally, a dataset contains three "splits": training, test, and validation. Both the test and validation splits are essentially benchmarks. In general, a benchmark is distinguished from a test/validation dataset in that a benchmark is typically intended to be used to measure the performance of many different models that are not trained specifically for doing well on the benchmark, while a test/validation set is intended to be used to measure the performance of models trained specifically on the corresponding training set. In other words, a benchmark may be thought of as a test/validation set without a corresponding training set. Conversely, certain benchmarks may be used as a training set, such as the English Gigaword or the One Billion Word Benchmark, which in modern language is just the negative log-likelihood loss on a pretraining set with 1 billion words. Indeed, the distinction between benchmark and dataset in language models became sharper after the rise of the pretraining paradigm, whereby a model is first trained on massive, unlabeled datasets to learn general language patterns, syntax, and knowledge (pretraining), and the base model is then adapted to specific, downstream tasks using smaller, labeled datasets (fine-tuning). === Lifecycle === Generally, the life cycle of a benchmark consists of the following steps: Inception: A benchmark is published. It can be simply given as a demonstration of the power of a new model (implicitly) that others then picked up as a benchmark, or as a benchmark that others are encouraged to use (explicitly). Growth: More papers and models use the benchmark, and the performance on the benchmark grows. Maturity, degeneration or deprecation: A benchmark may be saturated, after which researchers move on to other benchmarks. Progress on the benchmark may also be neglected as the field moves to focus on other benchmarks. Renewal: A saturated benchmark can be upgraded to make it no longer saturated, allowing further progress. === Construction === Like datasets, benchmarks are typically constructed by several methods, individually or in combination: Web scraping: Ready-made question-answer pairs may be scraped online, such as from websites that teach mathematics and programming. Conversion: Items may be constructed programmatically from scraped web content, such as by blanking out named entities from sentences, and asking the model to fill in the blank. This was used for making the CNN/Daily Mail Reading Comprehension Task. Crowd sourcing: Items may be constructed by paying people to write them, such as on Amazon Mechanical Turk. This was used for making the MCTest. === Evaluation === Generally, benchmarks are fully automated. This limits the questions that can be asked. For example, with mathematical questions, "proving a claim" would be difficult to automatically check, while "calculate an answer with a unique integer answer" would be automatically checkable. With programming tasks, the answer can generally be checked by running unit tests, with an upper limit on runtime. The benchmark scores are of the following kinds: For multiple choice or cloze questions, common scores are accuracy (frequency of correct answer), precision, recall, F1 score, etc. pass@n: The model is given n {\displaystyle n} attempts to solve each problem. If any attempt is correct, the model earns a point. The pass@n score is the model's average score over all problems. k@n: The model makes n {\displaystyle n} attempts to solve each problem, but only k {\displaystyle k} attempts out of them are selected for submission. If any submission is correct, the model earns a point. The k@n score is the model's average score over all problems. cons@n: The model is given n {\displaystyle n} attempts to solve each problem. If the most common answer is correct, the model earns a point. The cons@n score is the model's average score over all problems. Here "cons" stands for "consensus" or "majority voting". The pass@n score can be estimated more accurately by making N > n {\displaystyle N>n} attempts, and use the unbiased estimator 1 − ( N − c n ) ( N n ) {\displaystyle 1-{\frac {\binom {N-c}{n}}{\binom {N}{n}}}} , where c {\displaystyle c} is the number of correct attempts. For less well-formed tasks, where the output can be any sentence, there are the following commonly used scores including BLEU ROUGE, METEOR, NIST, word error rate, LEPOR, CIDEr, and SPICE. === Issues === error: Some benchmark answers may be wrong. ambiguity: Some benchmark questions may be ambiguously worded. subjective: Some benchmark questions may not have an objective answer at all. This problem generally prevents creative writing benchmarks. Similarly, this prevents benchmarking writing proofs in natural language, though benchmarking proofs in a formal language is possible. open-ended: Some benchmark questions may not have a single answer of a fixed size. This problem generally prevents programming benchmarks from using more natural tasks such as "write a program for X", and instead uses tasks such as "write a function that implements specification X". inter-annotator agreement: Some benchmark questions may be not fully objective, such that even people would not agree with 100% on what the answer should be. This is common in natural language processing tasks, such as syntactic annotation. shortcut: Some benchmark questions may be easily solved by an "unintended" shortcut. For example, in the SNLI benchmark, having a negative word like "not" in the second sentence is a strong signal for the "Contradiction" category, regardless of what the se
List of network buses
List of electrical characteristics of single collision domain segment "slow speed" network buses: The number of nodes can be limited by either number of available addresses or bus capacitance. None of the above use any analog domain modulation techniques like MLT-3 encoding, PAM-5 etc. PSI5 designed with automation applications in mind is a bit unusual in that it uses Manchester code.