Artificial intelligence in Indonesia refers to development, use and governance of artificial intelligence in Indonesia. Indonesia has treated AI as a national policy area through the Strategi Nasional Kecerdasan Artifisial or National Artificial Intelligence Strategy for 2020–2045. Public discussion has focused on the role of AI in sectors such as health, agriculture, education, mobile technology and e-commerce. Recent developments include AI ethics guidance issued by the communications ministry. Proposals for a national AI roadmap and sovereign AI fund, investment in cloud and AI infrastructure, and local-language AI initiatives for Bahasa Indonesia and regional Indonesian languages. == National strategy == Indonesia's National Artificial Intelligence Strategy is known in Indonesian as Strategi Nasional Kecerdasan Artifisial or Stranas KA. The strategy was published as a long-term framework for the development and use of AI between 2020 and 2045. It is intended to guide ministries, government agencies, regional governments and other stakeholders. The strategy identifies five priority sectors: health services, bureaucratic reform, education and research, food security, and mobility and smart cities. OECD lists the Ministry of Research and Technology and the National Research and Innovation Agency as organisations associated with the strategy. The strategy was developed through consultation with public and private stakeholders. == Institutions == The Indonesian Artificial Intelligence Industry Research and Innovation Collaboration, known as KORIKA is the nodal agency for the national AI strategy. KORIKA describes its vision as creating a collaborative ecosystem to accelerate implementation of the national AI strategy towards Vision Indonesia 2045. The Ministry of Communication and Digital Affairs has also been involved in AI governance, digital policy and public communication. In 2025, Reuters reported that the ministry was preparing a national AI roadmap to give investors and developers a clearer view of Indonesia's market, infrastructure and computing capacity. == AI Governance == Indonesia has introduced policy guidance on the ethical use of artificial intelligence. The policy sets out ethical values for the development and use of AI. These include humanity, security, transparency, credibility and accountability, personal data protection, sustainable development and intellectual property protection. A UNESCO country profile on Indonesia noted that Indonesia had adopted a national AI strategy and had policy frameworks. It also identified gaps in internet access, gender inclusion, language datasets, digital talent and cybersecurity. UNESCO recommended that Indonesia update its AI standards, invest in ethical AI, strengthen research coordination and consider establishing a national agency for artificial intelligence. In May 2026, Antara News reported comments by Deputy Minister of Communication and Digital Affairs Nezar Patria. Who said that AI safety requires partnerships, shared standards and continuing dialogue. == Sectors == AI policy discussions in Indonesia have identified health, agriculture, education, government services, mobility and smart cities as areas where AI could be applied. Mobile technology and e-commerce have been discussed as important areas of AI adoption in Indonesia. Research on AI adoption in Indonesia by Siddhartha Paul Tiwari and Adi Fahrudin has also examined mobile and e-commerce sectors. UNESCO has also noted that Indonesia's large digital economy and startup ecosystem have supported AI adoption, while also pointing to challenges in talent, research capacity and cybersecurity. Indonesia is one of the developing-country markets attracting AI infrastructure investment, including data centres. == Challenges == Indonesia faces several challenges in developing and governing AI. These include gaps in computing infrastructure, uneven connectivity outside major cities, shortages of skilled workers, limited research funding, cybersecurity risks, misinformation, data leaks and the underrepresentation of Indonesian and indigenous languages in AI datasets. UNESCO noted that Bahasa is spoken by around 200 million people but remains underrepresented in AI. It also noted that Indonesia has more than 700 indigenous languages, many of which face the risk of extinction. UNESCO recommended stronger coordination in AI research and a more unified strategy for using AI in language preservation.
Groover
Groover is an online platform, record label and distributor, connecting artists and musicians with music professionals and media outlets. The service was founded in 2018 in France and operates from offices in Paris and New York. The platform has over 3,000 active contacts, including SPIN Magazine and Sofar Sounds. Groover uses a micro-payment model. Among the platform's over 500,000 regular users are record labels such as Ninja Tune, Ba Da Bing Records, Dance To The Radio, Roche Musique, Wagram Music, Secret City Records, and artists including Bonobo, Michael Bolton, Aloe Blacc, Haddaway, Passenger, La Femme and Chinese Man. == History == Groover was launched at the MaMA Music Convention in October 2018. It was co-founded by Dorian Perron, Romain Palmieri, and Rafaël Cohen while they were students at UC Berkeley. Initially growing in France, the company has expanded to the United States, Canada, the United Kingdom, Brazil, Italy, and elsewhere in Europe. In March 2019, Groover was part of the Business France delegation at the South by Southwest (SXSW) festival. In June 2019, Groover raised €1.3 million from various angel investors. In April 2021, Groover acquired the platform Soonvibes, which had 70,000 users at the time, in order to strengthen its community in the electronic music space. In November 2021, Groover announced a €6 million funding round from Bpifrance Creative Industries and Partech. Between 2023 and 2025, Groover entered strategic partnerships with major artist service providers, including CD Baby, TuneCore, SoundCloud, UnitedMasters, Symphonic Distribution, Audiomack and SACEM. In February 2024, Groover announced a Series A funding round of $8 million from OneRagTime, Trind, Techmind, and Mozza Angels. == Function == Using a micro-payment system, professionals listen to tracks and provide written feedback. These professionals retain full editorial independence and are under no obligation to share the track or contact the artist. == Awards == 2nd Prize for Music Innovation 2023 from the Centre national de la musique (France) "Future Creator" Award at the Petit Poucet Competition 2019 Jury's Special Mention at the MaMA Invent 2019 competition 1st Prize for Digital Initiative in Culture, Communication & Media 2019 awarded by Audiens "Start-up of the Year" at the Social Music Awards 2020 French American Entrepreneurship Award 2022 at the French Consulate in New York
Ordered weighted averaging
In applied mathematics, specifically in fuzzy logic, the ordered weighted averaging (OWA) operators provide a parameterized class of mean type aggregation operators. They were introduced by Ronald R. Yager. Many notable mean operators such as the max, arithmetic average, median and min, are members of this class. They have been widely used in computational intelligence because of their ability to model linguistically expressed aggregation instructions. == Definition == An OWA operator of dimension n {\displaystyle \ n} is a mapping F : R n → R {\displaystyle F:\mathbb {R} ^{n}\rightarrow \mathbb {R} } that has an associated collection of weights W = [ w 1 , … , w n ] {\displaystyle \ W=[w_{1},\ldots ,w_{n}]} lying in the unit interval and summing to one and with F ( a 1 , … , a n ) = ∑ j = 1 n w j b j {\displaystyle F(a_{1},\ldots ,a_{n})=\sum _{j=1}^{n}w_{j}b_{j}} where b j {\displaystyle b_{j}} is the jth largest of the a i {\displaystyle a_{i}} . By choosing different W one can implement different aggregation operators. The OWA operator is a non-linear operator as a result of the process of determining the bj. == Notable OWA operators == F ( a 1 , … , a n ) = max ( a 1 , … , a n ) {\displaystyle \ F(a_{1},\ldots ,a_{n})=\max(a_{1},\ldots ,a_{n})} if w 1 = 1 {\displaystyle \ w_{1}=1} and w j = 0 {\displaystyle \ w_{j}=0} for j ≠ 1 {\displaystyle j\neq 1} F ( a 1 , … , a n ) = min ( a 1 , … , a n ) {\displaystyle \ F(a_{1},\ldots ,a_{n})=\min(a_{1},\ldots ,a_{n})} if w n = 1 {\displaystyle \ w_{n}=1} and w j = 0 {\displaystyle \ w_{j}=0} for j ≠ n {\displaystyle j\neq n} F ( a 1 , … , a n ) = a v e r a g e ( a 1 , … , a n ) {\displaystyle \ F(a_{1},\ldots ,a_{n})=\mathrm {average} (a_{1},\ldots ,a_{n})} if w j = 1 n {\displaystyle \ w_{j}={\frac {1}{n}}} for all j ∈ [ 1 , n ] {\displaystyle j\in [1,n]} == Properties == The OWA operator is a mean operator. It is bounded, monotonic, symmetric, and idempotent, as defined below. == Characterizing features == Two features have been used to characterize the OWA operators. The first is the attitudinal character, also called orness. This is defined as A − C ( W ) = 1 n − 1 ∑ j = 1 n ( n − j ) w j . {\displaystyle A-C(W)={\frac {1}{n-1}}\sum _{j=1}^{n}(n-j)w_{j}.} It is known that A − C ( W ) ∈ [ 0 , 1 ] {\displaystyle A-C(W)\in [0,1]} . In addition A − C(max) = 1, A − C(ave) = A − C(med) = 0.5 and A − C(min) = 0. Thus the A − C goes from 1 to 0 as we go from Max to Min aggregation. The attitudinal character characterizes the similarity of aggregation to OR operation(OR is defined as the Max). The second feature is the dispersion. This defined as H ( W ) = − ∑ j = 1 n w j ln ( w j ) . {\displaystyle H(W)=-\sum _{j=1}^{n}w_{j}\ln(w_{j}).} An alternative definition is E ( W ) = ∑ j = 1 n w j 2 . {\displaystyle E(W)=\sum _{j=1}^{n}w_{j}^{2}.} The dispersion characterizes how uniformly the arguments are being used. == Type-1 OWA aggregation operators == The above Yager's OWA operators are used to aggregate the crisp values. Can we aggregate fuzzy sets in the OWA mechanism? The Type-1 OWA operators have been proposed for this purpose. So the type-1 OWA operators provides us with a new technique for directly aggregating uncertain information with uncertain weights via OWA mechanism in soft decision making and data mining, where these uncertain objects are modelled by fuzzy sets. The type-1 OWA operator is defined according to the alpha-cuts of fuzzy sets as follows: Given the n linguistic weights { W i } i = 1 n {\displaystyle \left\{{W^{i}}\right\}_{i=1}^{n}} in the form of fuzzy sets defined on the domain of discourse U = [ 0 , 1 ] {\displaystyle U=[0,\;\;1]} , then for each α ∈ [ 0 , 1 ] {\displaystyle \alpha \in [0,\;1]} , an α {\displaystyle \alpha } -level type-1 OWA operator with α {\displaystyle \alpha } -level sets { W α i } i = 1 n {\displaystyle \left\{{W_{\alpha }^{i}}\right\}_{i=1}^{n}} to aggregate the α {\displaystyle \alpha } -cuts of fuzzy sets { A i } i = 1 n {\displaystyle \left\{{A^{i}}\right\}_{i=1}^{n}} is given as Φ α ( A α 1 , … , A α n ) = { ∑ i = 1 n w i a σ ( i ) ∑ i = 1 n w i | w i ∈ W α i , a i ∈ A α i , i = 1 , … , n } {\displaystyle \Phi _{\alpha }\left({A_{\alpha }^{1},\ldots ,A_{\alpha }^{n}}\right)=\left\{{{\frac {\sum \limits _{i=1}^{n}{w_{i}a_{\sigma (i)}}}{\sum \limits _{i=1}^{n}{w_{i}}}}\left|{w_{i}\in W_{\alpha }^{i},\;a_{i}}\right.\in A_{\alpha }^{i},\;i=1,\ldots ,n}\right\}} where W α i = { w | μ W i ( w ) ≥ α } , A α i = { x | μ A i ( x ) ≥ α } {\displaystyle W_{\alpha }^{i}=\{w|\mu _{W_{i}}(w)\geq \alpha \},A_{\alpha }^{i}=\{x|\mu _{A_{i}}(x)\geq \alpha \}} , and σ : { 1 , … , n } → { 1 , … , n } {\displaystyle \sigma :\{\;1,\ldots ,n\;\}\to \{\;1,\ldots ,n\;\}} is a permutation function such that a σ ( i ) ≥ a σ ( i + 1 ) , ∀ i = 1 , … , n − 1 {\displaystyle a_{\sigma (i)}\geq a_{\sigma (i+1)},\;\forall \;i=1,\ldots ,n-1} , i.e., a σ ( i ) {\displaystyle a_{\sigma (i)}} is the i {\displaystyle i} th largest element in the set { a 1 , … , a n } {\displaystyle \left\{{a_{1},\ldots ,a_{n}}\right\}} . The computation of the type-1 OWA output is implemented by computing the left end-points and right end-points of the intervals Φ α ( A α 1 , … , A α n ) {\displaystyle \Phi _{\alpha }\left({A_{\alpha }^{1},\ldots ,A_{\alpha }^{n}}\right)} : Φ α ( A α 1 , … , A α n ) − {\displaystyle \Phi _{\alpha }\left({A_{\alpha }^{1},\ldots ,A_{\alpha }^{n}}\right)_{-}} and Φ α ( A α 1 , … , A α n ) + , {\displaystyle \Phi _{\alpha }\left({A_{\alpha }^{1},\ldots ,A_{\alpha }^{n}}\right)_{+},} where A α i = [ A α − i , A α + i ] , W α i = [ W α − i , W α + i ] {\displaystyle A_{\alpha }^{i}=[A_{\alpha -}^{i},A_{\alpha +}^{i}],W_{\alpha }^{i}=[W_{\alpha -}^{i},W_{\alpha +}^{i}]} . Then membership function of resulting aggregation fuzzy set is: μ G ( x ) = ∨ α : x ∈ Φ α ( A α 1 , ⋯ , A α n ) α α {\displaystyle \mu _{G}(x)=\mathop {\vee } _{\alpha :x\in \Phi _{\alpha }\left({A_{\alpha }^{1},\cdots ,A_{\alpha }^{n}}\right)_{\alpha }}\alpha } For the left end-points, we need to solve the following programming problem: Φ α ( A α 1 , ⋯ , A α n ) − = min W α − i ≤ w i ≤ W α + i A α − i ≤ a i ≤ A α + i ∑ i = 1 n w i a σ ( i ) / ∑ i = 1 n w i {\displaystyle \Phi _{\alpha }\left({A_{\alpha }^{1},\cdots ,A_{\alpha }^{n}}\right)_{-}=\min \limits _{\begin{array}{l}W_{\alpha -}^{i}\leq w_{i}\leq W_{\alpha +}^{i}A_{\alpha -}^{i}\leq a_{i}\leq A_{\alpha +}^{i}\end{array}}\sum \limits _{i=1}^{n}{w_{i}a_{\sigma (i)}/\sum \limits _{i=1}^{n}{w_{i}}}} while for the right end-points, we need to solve the following programming problem: Φ α ( A α 1 , ⋯ , A α n ) + = max W α − i ≤ w i ≤ W α + i A α − i ≤ a i ≤ A α + i ∑ i = 1 n w i a σ ( i ) / ∑ i = 1 n w i {\displaystyle \Phi _{\alpha }\left({A_{\alpha }^{1},\cdots ,A_{\alpha }^{n}}\right)_{+}=\max \limits _{\begin{array}{l}W_{\alpha -}^{i}\leq w_{i}\leq W_{\alpha +}^{i}A_{\alpha -}^{i}\leq a_{i}\leq A_{\alpha +}^{i}\end{array}}\sum \limits _{i=1}^{n}{w_{i}a_{\sigma (i)}/\sum \limits _{i=1}^{n}{w_{i}}}} Zhou et al. presented a fast method to solve two programming problem so that the type-1 OWA aggregation operation can be performed efficiently. == OWA for committee voting == Amanatidis, Barrot, Lang, Markakis and Ries present voting rules for multi-issue voting, based on OWA and the Hamming distance. Barrot, Lang and Yokoo study the manipulability of these rules.
TCEC Season 14
The 14th season of the Top Chess Engine Championship took place between 17 November 2018 and 24 February 2019. Stockfish was the defending champion, having defeated Komodo in the previous season's superfinal. The season is notable for two things: the emergence of two strong, new engines, the Komodo variant Komodo Monte Carlo tree search (MCTS) and the neural network engine Leela Chess Zero, and the dramatic superfinal. Komodo MCTS and Leela fought their way from Division 4 and Division 3 respectively to the Premier Division, with Leela further qualifying for the superfinal against Stockfish. The superfinal was a topsy-turvy affair with the lead changing hands several times. It finished as the closest superfinal TCEC has ever seen, with Stockfish winning by a single game, 50.5–49.5 (+10 =81 -9). == Overview == === Structure === The season comprised five divisions: from the lowest Division 4 to the Premier Division. The top two engines of each division promote to the division above, while the bottom two engines relegate. The top two engines of the Premier Division contest a 100-game superfinal. The lengths of the opening books used increases as the divisions progress. The superfinal itself used a custom opening book designed by Jeroen Noomen. === Rules === The TCEC draw and win rules were slightly modified for Season 14. The game is now adjudicated as drawn if, after move 30, both engines have evals ±0.08 for five consecutive moves, and there are neither pawn moves nor a capture. Win adjudication now occurs if both engines have an eval of ±10 for five consecutive moves. Following the controversy over DeusX's participation last season, the uniqueness rule for neural networks was modified such that at least two of the following three hallmarks must be unique: The code for training the neural network The neural network (and weights file) itself The engine that executes this network This change meant DeusX did not meet the uniqueness criteria and therefore did not participate. Aside from this change, the season used the standard rules of the TCEC. == Results == === Division 4 === New entrant Komodo MCTS dominated Division 4, winning by a clear four points, although it did lose a game to second-place finisher rofChade. Fellow new entrant Scorpio NN performed badly and finished last, drawing only one game and losing the rest. === Division 3 === The neural network engine Leela Chess Zero had just missed promotion to Division 2 in the previous season. Since its relatively weak performance last season was partly due to hardware problems, and since it had shown a lot of improvement in strength, it was the hot favourite in this division. Leela lived up to its billing by comprehensively defeating everyone else. In a portent of future divisions however, Leela surprisingly dropped a game to third-place Arasan. Komodo MCTS was also improving quickly, and an updated version finished second behind Leela. The gap between second and third was 6.5 points, illustrating the gulf in class. === Division 2 === Although Division 2 engines are significantly stronger than Division 3, Leela and Komodo MCTS continued to dominate the competition, and again finished first and second. Komodo MCTS only lost one game to Leela, while Leela's tendency to occasionally lose to weaker engines saw her losing a game to 4th-placed Booot. Third place finisher Xiphos gave Leela and Komodo MCTS a run for their money, and was in the running up until the final rounds when it lost a crucial game to Leela. This loss left it one point behind Komodo MCTS in the final standings. === Division 1 === Leela and Komodo MCTS's rampage through the lower divisions continued, and they again finished first and second. In a demonstration of how much it had improved, Leela scored 20/28 in this division, the same score it had achieved in Division 2. This was also a TCEC points record for this division. However, Leela dropped a game against fourth-place finisher Chiron. Komodo MCTS, which had yet to lose a game in the lower divisions except to Leela, also conceded its first loss to third-place Fizbo. At the other end of the table, former champions Jonny and Fritz, which had not been updated, found themselves outclassed and finished second-last and last respectively; however with fellow competitor Ginkgo crashing five times (and therefore being disqualified), Jonny managed to stay in the division. The penultimate game for this division set a new TCEC moves record for a decisive game: 308 moves before Leela defeated Fritz. === Premier division === This was the strongest premier division ever, with multiple-time champions Stockfish, Komodo, and Houdini in the mix. Right from the start it became clear that Stockfish was in a league of its own, and it dominated the division, scoring wins against every other engine without losing a game. Second place however was a hotly-contested affair, with Leela, Komodo and Houdini neck-and-neck for most of the division. Houdini took the early lead, but Komodo gained second after winning two games by forfeit when its sibling Komodo MCTS crashed. This led to murmurs of a "Konspiracy". However, when both Komodo and Houdini failed to score more wins against the lower half of the field, Leela was able to take the lead. Halfway through the division the race was upended again when Leela went through a bad streak, losing three games in a row to Stockfish, Komodo, and Fire. This led to Komodo regaining second place, only for Komodo MCTS to crash yet again. By TCEC rules this meant Komodo MCTS was disqualified and all its scores were zeroed out, which put Leela back in second place. With three games left, Leela missed a win against Andscacs, which would've more or less secured her a place in the superfinal. Meanwhile, Komodo kept the division interesting by winning two of its last three games. Because Komodo had superior tiebreakers to Leela, this meant Komodo would qualify for the superfinal unless Leela managed to hold Stockfish to a draw with Black in the last game of the division. In a tense final game, Stockfish came close to winning, but missed the winning line. Leela managed to draw and qualified for the superfinal. At the other end of the table, it was quickly apparent that Ethereal and Andscacs were the weakest engines and would likely relegate. However, when Komodo MCTS was disqualified (and therefore relegated), it threw both engines a lifeline, since they could now stay in the division by beating the other. Andscacs was able to score a head-to-head win against Ethereal, but was crushed by Stockfish (+0 =2 -4) and Leela (+0 =3 -3). Ethereal didn't manage to score a win in the entire division, but did manage to score more draws than Andscacs, condemning Andscacs to relegation. === Superfinal === Going into the superfinal expectations were high for Leela: she had received a new network and had just won her first major competition when she defeated Houdini in the second TCEC cup. However, she had won the tournament without having played Stockfish (who had been surprisingly eliminated by Houdini in the semifinals). That, plus the fact that Stockfish dominated Premier Division and had never lost a match to Leela, left it unclear which engine was superior, although most spectators favored Stockfish. The superfinal turned out to be a roller-coaster. It began with Stockfish drawing first blood in game 7, and then scoring another win in game 10. Leela hit back with wins in game 11 and 13, but then lost games 20, 21, and 22. This gave Stockfish a 3-point lead. However, in the next 30 games, Leela was the only one to score wins: it first equalized by winning games 25, 27, and 29, and then took the lead by winning games 49 and 53. Stockfish won game 56, but Leela won game 63, maintaining her lead. There followed two dramatic games. In game 65, Leela built up a winning position. Stockfish showed a +153 evaluation, indicating that it had found a forced line leading to an endgame tablebase win; indeed analysis with 7-piece tablebases showed that Leela's position was winning. Under previous seasons' rules, the game would have been adjudicated as a win because Leela's evaluation was above 6.5. However under the new rules, Leela's +8.92 evaluation was not enough to adjudicate. It turned out that Leela could not see the winning line, and shuffled her pieces aimlessly, leading to a 50-move draw. In game 66, Stockfish was given a substantial advantage by the opening, but failed to make the most of it. The evaluations were leveling out to zero when the internet connection to the GPU servers was cut off. By tournament rules, this meant the game was replayed from scratch. After a further internet disconnection and restart, Stockfish handled the opening better and won, leaving Leela with a 1-point lead. In the last third of the superfinal, there followed more drama as Leela often built up strong advantages, but Stockfish showed great resourcefulness in defending inferior positions. Meanwh
Federal Virtual World Challenge
The Federal Virtual Challenge, formerly The Federal Virtual Worlds Challenge is a competition led by the Simulation and Training Technology Center (United States Army Research, Development and Engineering Command). The event is conducted in order to reach a global development community that will create innovative and interactive training and analysis services in virtual worlds. The inaugural event began in 2009 with the awards being conducted during March 2010 GameTech conference in Orlando, Florida. == Description == The focus of the challenge is training or analysis capability conducted wholly in a virtual environment. The training and analysis audience includes all United States Federal Agencies including, Department of Defense, Department of Homeland Security, Department of Transportation, and Department of Health and Human Services, NASA, DOT, and many more.
Wetware (brain)
Wetware is a term drawn from the computer-related idea of hardware or software, but applied to biological life forms. == Usage == The prefix "wet" is a reference to the water found in living creatures. Wetware is used to describe the elements equivalent to hardware and software found in a person, especially the central nervous system (CNS) and the human mind. The term wetware finds use in works of fiction, in scholarly publications and in popularizations. The "hardware" component of wetware concerns the bioelectric and biochemical properties of the CNS, specifically the brain. If the sequence of impulses traveling across the various neurons are thought of symbolically as software, then the physical neurons would be the hardware. The amalgamated interaction of this software and hardware is manifested through continuously changing physical connections, and chemical and electrical influences that spread across the body. The process by which the mind and brain interact to produce the collection of experiences that we define as self-awareness is in question. == History == Although the exact definition has shifted over time, the term Wetware and its fundamental reference to "the physical mind" has been around at least since the mid-1950s. Mostly used in relatively obscure articles and papers, it was not until the heyday of cyberpunk, however, that the term found broad adoption. Among the first uses of the term in popular culture was the Bruce Sterling novel Schismatrix (1985) and the Michael Swanwick novel Vacuum Flowers (1987). Rudy Rucker references the term in a number of books, including one entitled Wetware (1988): ... all sparks and tastes and tangles, all its stimulus/response patterns – the whole bio-cybernetic software of mind. Rucker did not use the word to simply mean a brain, nor in the human-resources sense of employees. He used wetware to stand for the data found in any biological system, analogous perhaps to the firmware that is found in a ROM chip. In Rucker's sense, a seed, a plant graft, an embryo, or a biological virus are all wetware. DNA, the immune system, and the evolved neural architecture of the brain are further examples of wetware in this sense. Rucker describes his conception in a 1992 compendium The Mondo 2000 User's Guide to the New Edge, which he quotes in a 2007 blog entry. Early cyber-guru Arthur Kroker used the term in his blog. With the term getting traction in trendsetting publications, it became a buzzword in the early 1990s. In 1991, Dutch media theorist Geert Lovink organized the Wetware Convention in Amsterdam, which was supposed to be an antidote to the "out-of-body" experiments conducted in high-tech laboratories, such as experiments in virtual reality. Timothy Leary, in an appendix to Info-Psychology originally written in 1975–76 and published in 1989, used the term wetware, writing that "psychedelic neuro-transmitters were the hot new technology for booting-up the 'wetware' of the brain". Another common reference is: "Wetware has 7 plus or minus 2 temporary registers." The numerical allusion is to a classic 1957 article by George A. Miller, The magical number 7 plus or minus two: some limits in our capacity for processing information, which later gave way to Miller's law.
Model collapse
Model collapse, also known by other names such as "AI inbreeding", "AI cannibalism", "Habsburg AI", and "model autophagy disorder" or "MAD" is a phenomenon noted in artificial intelligence studies, where machine learning models gradually degrade due to errors coming from uncurated synthetic data, or due to training on the outputs of another model such as prior versions of itself. It is unclear to what extent the phenomenon threatens the long-term development of such models, and some techniques have been proposed to mitigate the effect. == Characteristics == Shumailov et al. coined the term to describe two specific stages to the degradation of machine learning models: early model collapse and late model collapse: In early model collapse, the model begins losing information about the tails of the distribution – mostly affecting minority data. Later work highlighted that early model collapse is hard to notice, since overall performance may appear to improve, while the model loses performance on minority data. In late model collapse, the model loses a significant proportion of its performance, confusing concepts and losing most of its variance. == Mechanism == Using synthetic data as training data can lead to issues with the quality and reliability of the trained model. Model collapse occurs for three main reasons: functional approximation errors sampling errors learning errors Importantly, it happens in even the simplest of models, where not all of the error sources are present. In more complex models the errors often compound, leading to faster collapse. == Disagreement over real-world impact == Some researchers and commentators on model collapse warn that the phenomenon could fundamentally threaten future generative AI development: As AI-generated data is shared on the Internet, it will inevitably end up in future training datasets, which are often crawled from the Internet. If training on "slop" (large quantities of unlabeled synthetic data) inevitably leads to model collapse, this could therefore pose a difficult problem. However, recently, other researchers have disagreed with this argument, showing that if synthetic data accumulates alongside human-generated data, model collapse is avoided. The researchers argue that data accumulating over time is a more realistic description of reality than deleting all existing data every year, and that the real-world impact of model collapse may not be as catastrophic as feared. An alternative branch of the literature investigates the use of machine learning detectors and watermarking to identify model generated data and filter it out. == Mathematical models of the phenomenon == === 1D Gaussian model === In 2024, a first attempt has been made at illustrating collapse for the simplest possible model — a single dimensional normal distribution fit using unbiased estimators of mean and variance, computed on samples from the previous generation. To make this more precise, we say that original data follows a normal distribution X 0 ∼ N ( μ , σ 2 ) {\displaystyle X^{0}\sim {\mathcal {N}}(\mu ,\sigma ^{2})} , and we possess M 0 {\displaystyle M_{0}} samples X j 0 {\displaystyle X_{j}^{0}} for j ∈ { 1 , … , M 0 } {\displaystyle j\in {\{\,1,\dots ,M_{0}\,{}\}}} . Denoting a general sample X j i {\displaystyle X_{j}^{i}} as sample j ∈ { 1 , … , M i } {\displaystyle j\in {\{\,1,\dots ,M_{i}\,{}\}}} at generation i {\displaystyle i} , then the next generation model is estimated using the sample mean and variance: μ i + 1 = 1 M i ∑ j X j i ; σ i + 1 2 = 1 M i − 1 ∑ j ( X j i − μ i + 1 ) 2 . {\displaystyle \mu _{i+1}={\frac {1}{M_{i}}}\sum _{j}X_{j}^{i};\quad \sigma _{i+1}^{2}={\frac {1}{M_{i}-1}}\sum _{j}(X_{j}^{i}-\mu _{i+1})^{2}.} Leading to a conditionally normal next generation model X j i + 1 | μ i + 1 , σ i + 1 ∼ N ( μ i + 1 , σ i + 1 2 ) {\displaystyle X_{j}^{i+1}|\mu _{i+1},\;\sigma _{i+1}\sim {\mathcal {N}}(\mu _{i+1},\sigma _{i+1}^{2})} . In theory, this is enough to calculate the full distribution of X j i {\displaystyle X_{j}^{i}} . However, even after the first generation, the full distribution is no longer normal: It follows a variance-gamma distribution. To continue the analysis, instead of writing the probability density function at each generation, it is possible to explicitly construct them in terms of independent random variables using Cochran's theorem. To be precise, μ 1 {\displaystyle \mu _{1}} and σ 1 {\displaystyle \sigma _{1}} are independent, with μ 1 ∼ N ( μ , σ 2 M 0 ) {\displaystyle \mu _{1}\sim {\mathcal {N}}\left(\mu ,{\frac {\sigma ^{2}}{M_{0}}}\right)} and ( M 0 − 1 ) σ 1 2 ∼ σ 2 Γ ( M 0 − 1 2 , 1 2 ) {\displaystyle (M_{0}-1)\,\sigma _{1}^{2}\sim \sigma ^{2}\,\Gamma \left({\frac {M_{0}-1}{2}},{\frac {1}{2}}\right)} , following a Gamma distribution. Denoting with Z {\displaystyle Z} Gaussian random variables distributed according to N ( 0 , 1 ) {\displaystyle {\mathcal {N}}(0,1)} and with S i {\displaystyle S^{i}} random variables distributed with 1 M i − 1 − 1 Γ ( M i − 1 − 1 2 , 1 2 ) {\displaystyle {\frac {1}{M_{i-1}-1}}\Gamma \left({\frac {M_{i-1}-1}{2}},{\frac {1}{2}}\right)} , it turns out to be possible to write samples at each generation as X j 0 = μ + σ Z j 0 , {\textstyle X_{j}^{0}=\mu +\sigma Z_{j}^{0},} X j 1 = μ + σ M 0 Z 1 + σ S 1 Z j 1 , {\textstyle X_{j}^{1}=\mu +{\frac {\sigma }{\sqrt {M_{0}}}}Z^{1}+\sigma {\sqrt {S^{1}}}Z_{j}^{1},} and more generally X j n = μ + σ M 0 Z 1 + σ M 1 S 1 Z 2 + ⋯ + σ M n − 1 S 1 × ⋯ × S n − 1 Z n + σ S 1 × ⋯ × S n Z j n . {\displaystyle X_{j}^{n}=\mu +{\frac {\sigma }{\sqrt {M_{0}}}}Z^{1}+{\frac {\sigma }{\sqrt {M_{1}}}}{\sqrt {S^{1}}}Z^{2}+\dots +{\frac {\sigma }{\sqrt {M_{n-1}}}}{\sqrt {S^{1}\times \dots \times S^{n-1}}}Z^{n}+\sigma {\sqrt {S^{1}\times \dots \times S^{n}}}Z_{j}^{n}.} Note, that these are not joint distributions, as Z n {\displaystyle Z^{n}} and S n {\displaystyle S^{n}} depend directly on Z j n − 1 {\displaystyle Z_{j}^{n-1}} , but when considering X j n {\displaystyle X_{j}^{n}} on its own the formula above provides all the information about the full distribution. To analyse the model collapse, we can first calculate variance and mean of samples at generation n {\displaystyle n} . This would tell us what kind of distributions we expect to arrive at after n {\displaystyle n} generations. It is possible to find its exact value in closed form, but the mean and variance of the square root of gamma distribution are expressed in terms of gamma functions, making the result quite clunky. Following, it is possible to expand all results to second order in each of 1 / M i {\displaystyle 1/M_{i}} , assuming each sample size to be large. It is then possible to show that 1 σ 2 Var ( X j n ) = 1 M 0 + 1 M 1 + ⋯ + 1 M n − 1 + 1 + O ( M i − 2 ) . {\displaystyle {\frac {1}{\sigma ^{2}}}\operatorname {Var} (X_{j}^{n})={\frac {1}{M_{0}}}+{\frac {1}{M_{1}}}+\dots +{\frac {1}{M_{n-1}}}+1+{\mathcal {O}}\left(M_{i}^{-2}\right).} And if all sample sizes M i = M {\displaystyle M_{i}=M} are constant, this diverges linearly as n → ∞ {\displaystyle n\to \infty } : Var ( X j n ) = σ 2 ( 1 + n M ) ; E ( X j n ) = μ . {\displaystyle \operatorname {Var} (X_{j}^{n})=\sigma ^{2}\left(1+{\frac {n}{M}}\right);\quad \mathbb {E} (X_{j}^{n})=\mu .} This is the same scaling as for a single dimensional Gaussian random walk. However, divergence of the variance of X j n {\displaystyle X_{j}^{n}} does not directly provide any information about the corresponding estimates of μ n + 1 {\displaystyle \mu _{n+1}} and σ n + 1 {\displaystyle \sigma _{n+1}} , particularly how different they are from the original μ {\displaystyle \mu } and σ {\displaystyle \sigma } . It turns out to be possible to calculate the distance between the true distribution and the approximated distribution at step n + 1 {\displaystyle n+1} , using the Wasserstein-2 distance (which is also sometimes referred to as risk): E [ W 2 2 ( N ( μ , σ 2 ) , N ( μ n + 1 , σ n + 1 2 ) ) ] = 3 2 σ 2 ( 1 M 0 + 1 M 1 + ⋯ + 1 M n ) + O ( M i − 2 ) , {\displaystyle \mathbb {E} \left[\mathbb {W} _{2}^{2}\left({\mathcal {N}}(\mu ,\sigma ^{2}),{\mathcal {N}}(\mu _{n+1},\sigma _{n+1}^{2})\right)\right]={\frac {3}{2}}\sigma ^{2}\left({\frac {1}{M_{0}}}+{\frac {1}{M_{1}}}+\dots +{\frac {1}{M_{n}}}\right)+{\mathcal {O}}\left(M_{i}^{-2}\right),} Var [ W 2 2 ( N ( μ , σ 2 ) , N ( μ n + 1 , σ n + 1 2 ) ) ] = 1 2 σ 4 ( 3 M 0 2 + 3 M 1 2 + ⋯ + 3 M n 2 + ∑ i ≠ j 4 M i M j ) + O ( M i − 3 ) . {\displaystyle \operatorname {Var} \left[\mathbb {W} _{2}^{2}\left({\mathcal {N}}(\mu ,\sigma ^{2}),{\mathcal {N}}(\mu _{n+1},\sigma _{n+1}^{2})\right)\right]={\frac {1}{2}}\sigma ^{4}\left({\frac {3}{M_{0}^{2}}}+{\frac {3}{M_{1}^{2}}}+\dots +{\frac {3}{M_{n}^{2}}}+\sum _{i\neq j}{\frac {4}{M_{i}M_{j}}}\right)+{\mathcal {O}}\left(M_{i}^{-3}\right).} This directly shows why model collapse occurs in this simple model. Due to errors from re-sampling the approximated distribution, each generation ends up corresponding to a