Competition in artificial intelligence refers to the rivalry among companies, research institutions, and governments to develop and deploy the most capable artificial intelligence (AI) systems. The competition spans multiple domains, including large language models (LLMs), autonomous vehicles, robotics, computer vision systems, natural language processing (NLP), and AI-optimized hardware. == Background == Competition in AI is driven by potential economic, strategic, and scientific advantages. Breakthroughs in AI can enhance productivity, enable new products and services, and provide geopolitical leverage. The field has experienced rapid progress since the mid-2010s, particularly in machine learning and artificial neural networks, leading to intense rivalry among leading actors. == Corporate competition == Major technology companies are among the most visible competitors in AI. In the United States, firms such as OpenAI, Google DeepMind, Meta Platforms, Microsoft, Anthropic, and Nvidia compete in building advanced LLMs, generative AI platforms, and AI-optimized graphics processing units (GPUs). In China, companies such as Baidu, Alibaba Group, Tencent, and startups such DeepSeek have become leaders in AI deployment, often with state backing. The "[war for talent]" in AI research has become a defining feature of corporate competition. Leading firms often recruit top AI researchers from rivals, sometimes offering multi-million-dollar compensation packages. == National competition == Governments see leadership in AI as a strategic priority. The United States has funded AI research for military, economic, and societal applications, while China has set a target to lead the world in AI by 2030 through its "New Generation Artificial Intelligence Development Plan". Other nations, including the UK, India, Israel, Russia, South Korea, and members of the European Union, have launched national AI strategies. In February 2026 Anthropic said Chinese companies - DeepSeek, Moonshot AI, and MiniMax - were conducting "distillation attacks" in an attempt to copy their model's capabilities, and warned that business wars were closely tied to geopolitical ones: "foreign labs that illicitly distill American models can remove safeguards, feeding model capabilities into their own military, intelligence, and surveillance systems." == Sectors of competition == === Large language models and chatbots competition === Competition to produce the most capable generative text models, with benchmarks such as MMLU and ARC used to evaluate performance has been on scale since the emergence of AI. These systems leverage deep learning, especially transformer architectures, to understand and generate human-like language. Companies and research groups globally compete to develop chatbots that are more capable, reliable, and context-aware. Among the most well-known chatbots is ChatGPT, developed by OpenAI. Since its public release in 2022, ChatGPT has rapidly gained widespread attention for its ability to engage in coherent and versatile conversations, assist with creative writing, and solve complex problems. In response, technology firms introduced competing chatbots aiming to challenge or surpass ChatGPT's capabilities. Notably, DeepSeek, a Chinese AI company, launched an advanced chatbot integrated with their R1 language model, emphasizing strong natural language understanding and multilingual support. Similarly, Grok, developed by xAI (company), integrates conversational AI into vehicles and digital assistants, combining natural language processing with real-time data for personalized user interaction. These chatbots not only compete in language tasks but also demonstrate strategic reasoning capabilities by playing complex games such as chess and Go. This form of competition is reminiscent of historic AI milestones set by programs such as Deep Blue and AlphaGo. The OpenAI’s ChatGPT has been tested in playing chess at various levels, while DeepSeek’s chatbot showcased its prowess in online chess tournaments in early 2024, winning several matches against human and AI opponents. Grok, leveraging Tesla's vast data infrastructure, has demonstrated real-time strategic decision-making in simulation environments that include chess-like games. The competition pushes rapid innovation, with firms racing to improve chatbot conversational depth, reduce biases, increase factual accuracy, and integrate multimodal inputs like images and videos. At the same time, the competition raises questions about AI safety, ethical use, and the societal impacts of increasingly human-like chatbots. === Autonomous vehicles === Companies such as Waymo, Tesla, and Baidu are racing to deploy safe and reliable self-driving car technology. === AI chips === Rivalry between Nvidia, AMD, Intel, and Huawei in designing processors optimized for AI workloads. === Military applications === Development of AI-enabled drones, surveillance systems, and decision-support tools, with associated ethical debates. == Events == In 2023, OpenAI released GPT-4, prompting competitors such as Google DeepMind to accelerate the release of their own models, including Gemini. In 2024, Chinese AI company DeepSeek launched the R1 model, leading OpenAI to release an open-source system, GPT-OSS, as a strategic countermeasure. In 2022, Tesla and Waymo both expanded autonomous taxi services in U.S. cities, competing for regulatory approval and public trust. The U.S. Department of Defense's Project Maven and China's AI-enabled surveillance programs have been cited as examples of military AI rivalry. In 2025, Microsoft hired several senior engineers from Google DeepMind, highlighting the ongoing "talent poaching" competition in the AI sector. == Risks and concerns == Critics warn that unrestrained competition in AI can undermine safety, ethics, and governance. Concerns include the proliferation of biased or unsafe models, escalation in autonomous weapons, and reduced cooperation on safety standards.
Simple interactive object extraction
Simple interactive object extraction (SIOX) is an algorithm for extracting foreground objects from color images and videos with very little user interaction. It has been implemented as "foreground selection" tool in the GIMP (since version 2.3.3), as part of the tracer tool in Inkscape (since 0.44pre3), and as function in ImageJ and Fiji (plug-in). Experimental implementations were also reported for Blender and Krita. Although the algorithm was originally designed for videos, virtually all implementations use SIOX primarily for still image segmentation. In fact, it is often said to be the current de facto standard for this task in the open-source world. Initially, a free hand selection tool is used to specify the region of interest. It must contain all foreground objects to extract and as few background as possible. The pixels outside the region of interest form the sure background while the inner region define a superset of the foreground, i.e. the unknown region. A so-called foreground brush is then used to mark representative foreground regions. The algorithm outputs a selection mask. The selection can be refined by either adding further foreground markings or by adding background markings using the background brush. Technically, the algorithm performs the following steps: Create a set of representative colors for sure foreground and sure background, the so-called color signatures. Assign all image points to foreground or background by a weighted nearest neighbor search in the color signatures. Apply some standard image processing operations like erode, dilate, and blur to remove artifacts. Find the connected foreground components that are either large enough or marked by the user. For video segmentation the sure background and sure foreground regions are learned from motion statistics. SIOX also features tools that allow sub-pixel accurate refinement of edges and high texture areas, the so-called "detail refinement brushes". As with all segmentation algorithms, there are always pictures where the algorithm does not yield perfect results. The most critical drawback of SIOX is the color dependence. Although many photos are well-separable by color, the algorithm cannot deal with camouflage. If the foreground and background share many identical shades of similar colors, the algorithm might give a result with parts missing or incorrectly classified foreground. SIOX performs about equally well on different benchmarks compared to graph-based segmentation methods, such as Grabcut. SIOX is, however, more noise robust and can therefore also be used for the segmentation of videos. Graph-based segmentation methods search for a minimum cut and therefore tend to not perform optimally with complex structures. The algorithm has initially been developed at the department of computer science at Freie Universitaet Berlin. The main developer, Gerald Friedland, is now faculty at the EECS department of the University of California at Berkeley and also a Principal Data Scientist at Lawrence Livermore National Lab. He continues to support the development through mentoring, e.g. in the Google Summer of Code.
Communication-avoiding algorithm
Communication-avoiding algorithms minimize movement of data within a memory hierarchy for improving its running-time and energy consumption. These minimize the total of two costs (in terms of time and energy): arithmetic and communication. Communication, in this context refers to moving data, either between levels of memory or between multiple processors over a network. It is much more expensive than arithmetic. == Formal theory == === Two-level memory model === A common computational model in analyzing communication-avoiding algorithms is the two-level memory model: There is one processor and two levels of memory. Level 1 memory is infinitely large. Level 0 memory ("cache") has size M {\displaystyle M} . In the beginning, input resides in level 1. In the end, the output resides in level 1. Processor can only operate on data in cache. The goal is to minimize data transfers between the two levels of memory. === Matrix multiplication === Corollary 6.2: More general results for other numerical linear algebra operations can be found in. The following proof is from. == Motivation == Consider the following running-time model: Measure of computation = Time per FLOP = γ Measure of communication = No. of words of data moved = β ⇒ Total running time = γ·(no. of FLOPs) + β·(no. of words) From the fact that β >> γ as measured in time and energy, communication cost dominates computation cost. Technological trends indicate that the relative cost of communication is increasing on a variety of platforms, from cloud computing to supercomputers to mobile devices. The report also predicts that gap between DRAM access time and FLOPs will increase 100× over coming decade to balance power usage between processors and DRAM. Energy consumption increases by orders of magnitude as we go higher in the memory hierarchy. United States president Barack Obama cited communication-avoiding algorithms in the FY 2012 Department of Energy budget request to Congress: New Algorithm Improves Performance and Accuracy on Extreme-Scale Computing Systems. On modern computer architectures, communication between processors takes longer than the performance of a floating-point arithmetic operation by a given processor. ASCR researchers have developed a new method, derived from commonly used linear algebra methods, to minimize communications between processors and the memory hierarchy, by reformulating the communication patterns specified within the algorithm. This method has been implemented in the TRILINOS framework, a highly-regarded suite of software, which provides functionality for researchers around the world to solve large scale, complex multi-physics problems. == Objectives == Communication-avoiding algorithms are designed with the following objectives: Reorganize algorithms to reduce communication across all memory hierarchies. Attain the lower-bound on communication when possible. The following simple example demonstrates how these are achieved. === Matrix multiplication example === Let A, B and C be square matrices of order n × n. The following naive algorithm implements C = C + A B: for i = 1 to n for j = 1 to n for k = 1 to n C(i,j) = C(i,j) + A(i,k) B(k,j) Arithmetic cost (time-complexity): n2(2n − 1) for sufficiently large n or O(n3). Rewriting this algorithm with communication cost labelled at each step for i = 1 to n {read row i of A into fast memory} - n2 reads for j = 1 to n {read C(i,j) into fast memory} - n2 reads {read column j of B into fast memory} - n3 reads for k = 1 to n C(i,j) = C(i,j) + A(i,k) B(k,j) {write C(i,j) back to slow memory} - n2 writes Fast memory may be defined as the local processor memory (CPU cache) of size M and slow memory may be defined as the DRAM. Communication cost (reads/writes): n3 + 3n2 or O(n3) Since total running time = γ·O(n3) + β·O(n3) and β >> γ the communication cost is dominant. The blocked (tiled) matrix multiplication algorithm reduces this dominant term: ==== Blocked (tiled) matrix multiplication ==== Consider A, B and C to be n/b-by-n/b matrices of b-by-b sub-blocks where b is called the block size; assume three b-by-b blocks fit in fast memory. for i = 1 to n/b for j = 1 to n/b {read block C(i,j) into fast memory} - b2 × (n/b)2 = n2 reads for k = 1 to n/b {read block A(i,k) into fast memory} - b2 × (n/b)3 = n3/b reads {read block B(k,j) into fast memory} - b2 × (n/b)3 = n3/b reads C(i,j) = C(i,j) + A(i,k) B(k,j) - {do a matrix multiply on blocks} {write block C(i,j) back to slow memory} - b2 × (n/b)2 = n2 writes Communication cost: 2n3/b + 2n2 reads/writes << 2n3 arithmetic cost Making b as large possible: 3b2 ≤ M we achieve the following communication lower bound: 31/2n3/M1/2 + 2n2 or Ω (no. of FLOPs / M1/2) == Previous approaches for reducing communication == Most of the approaches investigated in the past to address this problem rely on scheduling or tuning techniques that aim at overlapping communication with computation. However, this approach can lead to an improvement of at most a factor of two. Ghosting is a different technique for reducing communication, in which a processor stores and computes redundantly data from neighboring processors for future computations. Cache-oblivious algorithms represent a different approach introduced in 1999 for fast Fourier transforms, and then extended to graph algorithms, dynamic programming, etc. They were also applied to several operations in linear algebra as dense LU and QR factorizations. The design of architecture specific algorithms is another approach that can be used for reducing the communication in parallel algorithms, and there are many examples in the literature of algorithms that are adapted to a given communication topology.
Birkhoff algorithm
Birkhoff's algorithm (also called Birkhoff-von-Neumann algorithm) is an algorithm for decomposing a bistochastic matrix into a convex combination of permutation matrices. It was published by Garrett Birkhoff in 1946. It has many applications. One such application is for the problem of fair random assignment: given a randomized allocation of items, Birkhoff's algorithm can decompose it into a lottery on deterministic allocations. == Terminology == A bistochastic matrix (also called: doubly-stochastic) is a matrix in which all elements are greater than or equal to 0 and the sum of the elements in each row and column equals 1. An example is the following 3-by-3 matrix: ( 0.2 0.3 0.5 0.6 0.2 0.2 0.2 0.5 0.3 ) {\displaystyle {\begin{pmatrix}0.2&0.3&0.5\\0.6&0.2&0.2\\0.2&0.5&0.3\end{pmatrix}}} A permutation matrix is a special case of a bistochastic matrix, in which each element is either 0 or 1 (so there is exactly one "1" in each row and each column). An example is the following 3-by-3 matrix: ( 0 1 0 0 0 1 1 0 0 ) {\displaystyle {\begin{pmatrix}0&1&0\\0&0&1\\1&0&0\end{pmatrix}}} A Birkhoff decomposition (also called: Birkhoff-von-Neumann decomposition) of a bistochastic matrix is a presentation of it as a sum of permutation matrices with non-negative weights. For example, the above matrix can be presented as the following sum: 0.2 ( 0 1 0 0 0 1 1 0 0 ) + 0.2 ( 1 0 0 0 1 0 0 0 1 ) + 0.1 ( 0 1 0 1 0 0 0 0 1 ) + 0.5 ( 0 0 1 1 0 0 0 1 0 ) {\displaystyle 0.2{\begin{pmatrix}0&1&0\\0&0&1\\1&0&0\end{pmatrix}}+0.2{\begin{pmatrix}1&0&0\\0&1&0\\0&0&1\end{pmatrix}}+0.1{\begin{pmatrix}0&1&0\\1&0&0\\0&0&1\end{pmatrix}}+0.5{\begin{pmatrix}0&0&1\\1&0&0\\0&1&0\end{pmatrix}}} Birkhoff's algorithm receives as input a bistochastic matrix and returns as output a Birkhoff decomposition. == Tools == A permutation set of an n-by-n matrix X is a set of n entries of X containing exactly one entry from each row and from each column. A theorem by Dénes Kőnig says that: Every bistochastic matrix has a permutation-set in which all entries are positive.The positivity graph of an n-by-n matrix X is a bipartite graph with 2n vertices, in which the vertices on one side are n rows and the vertices on the other side are the n columns, and there is an edge between a row and a column if the entry at that row and column is positive. A permutation set with positive entries is equivalent to a perfect matching in the positivity graph. A perfect matching in a bipartite graph can be found in polynomial time, e.g. using any algorithm for maximum cardinality matching. Kőnig's theorem is equivalent to the following:The positivity graph of any bistochastic matrix admits a perfect matching.A matrix is called scaled-bistochastic if all elements are non-negative, and the sum of each row and column equals c, where c is some positive constant. In other words, it is c times a bistochastic matrix. Since the positivity graph is not affected by scaling:The positivity graph of any scaled-bistochastic matrix admits a perfect matching. == Algorithm == Birkhoff's algorithm is a greedy algorithm: it greedily finds perfect matchings and removes them from the fractional matching. It works as follows. Let i = 1. Construct the positivity graph GX of X. Find a perfect matching in GX, corresponding to a positive permutation set in X. Let z[i] > 0 be the smallest entry in the permutation set. Let P[i] be a permutation matrix with 1 in the positive permutation set. Let X := X − z[i] P[i]. If X contains nonzero elements, Let i = i + 1 and go back to step 2. Otherwise, return the sum: z[1] P[1] + ... + z[2] P[2] + ... + z[i] P[i]. The algorithm is correct because, after step 6, the sum in each row and each column drops by z[i]. Therefore, the matrix X remains scaled-bistochastic. Therefore, in step 3, a perfect matching always exists. == Run-time complexity == By the selection of z[i] in step 4, in each iteration at least one element of X becomes 0. Therefore, the algorithm must end after at most n2 steps. However, the last step must simultaneously make n elements 0, so the algorithm ends after at most n2 − n + 1 steps, which implies O ( n 2 ) {\displaystyle O(n^{2})} . In 1960, Joshnson, Dulmage and Mendelsohn showed that Birkhoff's algorithm actually ends after at most n2 − 2n + 2 steps, which is tight in general (that is, in some cases n2 − 2n + 2 permutation matrices may be required). == Application in fair division == In the fair random assignment problem, there are n objects and n people with different preferences over the objects. It is required to give an object to each person. To attain fairness, the allocation is randomized: for each (person, object) pair, a probability is calculated, such that the sum of probabilities for each person and for each object is 1. The probabilistic-serial procedure can compute the probabilities such that each agent, looking at the matrix of probabilities, prefers his row of probabilities over the rows of all other people (this property is called envy-freeness). This raises the question of how to implement this randomized allocation in practice? One cannot just randomize for each object separately, since this may result in allocations in which some people get many objects while other people get no objects. Here, Birkhoff's algorithm is useful. The matrix of probabilities, calculated by the probabilistic-serial algorithm, is bistochastic. Birkhoff's algorithm can decompose it into a convex combination of permutation matrices. Each permutation matrix represents a deterministic assignment, in which every agent receives exactly one object. The coefficient of each such matrix is interpreted as a probability; based on the calculated probabilities, it is possible to pick one assignment at random and implement it. == Extensions == The problem of computing the Birkhoff decomposition with the minimum number of terms has been shown to be NP-hard, but some heuristics for computing it are known. This theorem can be extended for the general stochastic matrix with deterministic transition matrices. Budish, Che, Kojima and Milgrom generalize Birkhoff's algorithm to non-square matrices, with some constraints on the feasible assignments. They also present a decomposition algorithm that minimizes the variance in the expected values. Vazirani generalizes Birkhoff's algorithm to non-bipartite graphs. Valls et al. showed that it is possible to obtain an ϵ {\displaystyle \epsilon } -approximate decomposition with O ( log ( 1 / ϵ 2 ) ) {\displaystyle O(\log(1/\epsilon ^{2}))} permutations.
Adaptive algorithm
An adaptive algorithm is an algorithm that changes its behavior at the time it is run, based on information available and on a priori defined reward mechanism (or criterion). Such information could be the story of recently received data, information on the available computational resources, or other run-time acquired (or a priori known) information related to the environment in which it operates. Among the most used adaptive algorithms is the Widrow-Hoff’s least mean squares (LMS), which represents a class of stochastic gradient-descent algorithms used in adaptive filtering and machine learning. In adaptive filtering the LMS is used to mimic a desired filter by finding the filter coefficients that relate to producing the least mean square of the error signal (difference between the desired and the actual signal). For example, stable partition, using no additional memory is O(n lg n) but given O(n) memory, it can be O(n) in time. As implemented by the C++ Standard Library, stable_partition is adaptive and so it acquires as much memory as it can get (up to what it would need at most) and applies the algorithm using that available memory. Another example is adaptive sort, whose behavior changes upon the presortedness of its input. An example of an adaptive algorithm in radar systems is the constant false alarm rate (CFAR) detector. In machine learning and optimization, many algorithms are adaptive or have adaptive variants, which usually means that the algorithm parameters such as learning rate are automatically adjusted according to statistics about the optimisation thus far (e.g. the rate of convergence). Examples include adaptive simulated annealing, adaptive coordinate descent, adaptive quadrature, AdaBoost, Adagrad, Adadelta, RMSprop, and Adam. In data compression, adaptive coding algorithms such as Adaptive Huffman coding or Prediction by partial matching can take a stream of data as input, and adapt their compression technique based on the symbols that they have already encountered. In signal processing, the Adaptive Transform Acoustic Coding (ATRAC) codec used in MiniDisc recorders is called "adaptive" because the window length (the size of an audio "chunk") can change according to the nature of the sound being compressed, to try to achieve the best-sounding compression strategy.
Variable data publishing
Variable-data publishing (VDP) (also known as database publishing) is a term referring to the output of a variable composition system. While these systems can produce both electronically viewable and hard-copy (print) output, the "variable-data publishing" term today often distinguishes output destined for electronic viewing, rather than that which is destined for hard-copy print (e.g. variable data printing). Essentially the same techniques are employed to perform variable-data publishing, as those utilized with variable data printing. The difference is in the interpretation for output. While variable-data printing may be interpreted to produce various print streams or page-description files (e.g. AFP/IPDS, PostScript, PCL), variable-data publishing produces electronically viewable files, most commonly seen in the forms of PDF, HTML, or XML. Variable-data composition involves the use of data to conditionally: exhibit text (static blocks and/or variable content) exhibit images select fonts select colors format page layouts & flows Variable-data may be as simple as an address block or salutation. However, it can be any or all of the document's textual content—including words, sentences, paragraphs, pages, or the entire document. In other words, it can make up as little or as much of the document as the composer desires. Variable data may also be used to exhibit various images, such as logos, products, or membership photos. Further, variable-data can be used to build rule-based design schemes, including fonts, colors, and page formats. The possibilities are vast. The variable-data tools available today, make it possible to perform variable-data composition at nearly every stage of document production. However, the level of control that can be achieved varies, based upon how far into the document production process a variable-data tool is deployed. For example, if variable-data insertion occurs just prior to output...it's not likely that the text flow or layout can be altered with nearly as much control as would be available at the time of initial document composition. Many organizations will produce multiple forms of output (aka: multi-channel output), for the same document. This ensures that the published content is available to recipients via any form of access method they might require. When multi-channel output is utilized, integrity between those output channels often becomes important. Variable-data publishing may be performed on everything from a personal computer to a mainframe system. However, the speed and practical output volumes which can be achieved are directly affected by the computer power utilized. == Origin of the concept == The term variable-data publishing was likely an offshoot of the term "variable-data printing", first introduced to the printing industry by Frank Romano, Professor Emeritus, School of Print Media, at the College of Imaging Arts and Sciences at Rochester Institute of Technology. However, the concept of merging static document elements and variable document elements predates the term and has seen various implementations ranging from simple desktop 'mail merge', to complex mainframe applications in the financial and banking industry. In the past, the term VDP has been most closely associated with digital printing machines. However, in the past 3 years the application of this technology has spread to web pages, emails, and mobile messaging.
Least-squares spectral analysis
Least-squares spectral analysis (LSSA) is a class of methods for estimating a frequency spectrum by fitting sinusoids to data using a least-squares fit. Unlike Fourier analysis, the most widely used spectral method in science, data need not be equally spaced to use LSSA. Furthermore, while Fourier analysis generally amplifies long-period noise in long or gapped records, LSSA mitigates such problems. The first strictly least-squares LSSA method was developed in 1969 and 1971, and is known as the Vaníček method or the Gauss–Vaniček method, after its inventor Petr Vaníček and Carl Friedrich Gauss, the inventor of the least-squares method for error minimization. A widely known LSSA variant is the Lomb method or the Lomb–Scargle periodogram, based on dated computational simplifications of the Vaníček method introduced in the 1970s and 1980s, first by Nicholas R. Lomb and later by Jeffrey D. Scargle. Other LSSA variants have been subsequently developed. == Historical background == The close connections between Fourier analysis, the periodogram, and the least-squares fitting of sinusoids have been known for a long time. However, most developments are restricted to complete data sets of equally spaced samples. In 1963, Freek J. M. Barning of Mathematisch Centrum, Amsterdam, handled unequally spaced data by similar techniques, including both a periodogram analysis equivalent to what nowadays is called the Lomb method and least-squares fitting of selected frequencies of sinusoids determined from such periodograms — and connected by a procedure known today as the matching pursuit with post-back fitting or the orthogonal matching pursuit. Petr Vaníček, a Canadian geophysicist and geodesist of the University of New Brunswick, proposed in 1969 also the matching-pursuit approach for equally and unequally spaced data, which he called "successive spectral analysis" and the result a "least-squares periodogram". He generalized this method to account for any systematic components beyond a simple mean, such as a "predicted linear (quadratic, exponential, ...) secular trend of unknown magnitude", and applied it to a variety of samples, in 1971. Vaníček's strictly least-squares method was then simplified in 1976 by Nicholas R. Lomb of the University of Sydney, who pointed out its close connection to periodogram analysis. Subsequently, the definition of a periodogram of unequally spaced data was modified and analyzed by Jeffrey D. Scargle of NASA Ames Research Center, who showed that, with minor changes, it becomes identical to Lomb's least-squares formula for fitting individual sinusoid frequencies. Scargle states that his paper "does not introduce a new detection technique, but instead studies the reliability and efficiency of detection with the most commonly used technique, the periodogram, in the case where the observation times are unevenly spaced," and further points out regarding least-squares fitting of sinusoids compared to periodogram analysis, that his paper "establishes, apparently for the first time, that (with the proposed modifications) these two methods are exactly equivalent." Press summarizes the development this way: A completely different method of spectral analysis for unevenly sampled data, one that mitigates these difficulties and has some other very desirable properties, was developed by Lomb, based in part on earlier work by Barning and Vanicek, and additionally elaborated by Scargle. In 1989, Michael J. Korenberg of Queen's University in Kingston, Ontario, developed the "fast orthogonal search" method of more quickly finding a near-optimal decomposition of spectra or other problems, similar to the technique that later became known as the orthogonal matching pursuit. == Development of LSSA and variants == === The Vaníček method === In the Vaníček method, a discrete data set is approximated by a weighted sum of sinusoids of progressively determined frequencies using a standard linear regression or least-squares fit. The frequencies are chosen using a method similar to Barning's, but going further in optimizing the choice of each successive new frequency by picking the frequency that minimizes the residual after least-squares fitting (equivalent to the fitting technique now known as matching pursuit with pre-backfitting). The number of sinusoids must be less than or equal to the number of data samples (counting sines and cosines of the same frequency as separate sinusoids). The relationship between the DFT and the approximation of trigonometric functions using the least-squares method is well explained in (Strutz, 2017). A data vector Φ is represented as a weighted sum of sinusoidal basis functions, tabulated in a matrix A by evaluating each function at the sample times, with weight vector x: ϕ ≈ A x , {\displaystyle \phi \approx {\textbf {A}}x,} where the weights vector x is chosen to minimize the sum of squared errors in approximating Φ. The solution for x is closed-form, using standard linear regression: x = ( A T A ) − 1 A T ϕ . {\displaystyle x=({\textbf {A}}^{\mathrm {T} }{\textbf {A}})^{-1}{\textbf {A}}^{\mathrm {T} }\phi .} Here the matrix A can be based on any set of functions mutually independent (not necessarily orthogonal) when evaluated at the sample times; functions used for spectral analysis are typically sines and cosines evenly distributed over the frequency range of interest. If we choose too many frequencies in a too-narrow frequency range, the functions will be insufficiently independent, the matrix ill-conditioned, and the resulting spectrum meaningless. When the basis functions in A are orthogonal (that is, not correlated, meaning the columns have zero pair-wise dot products), the matrix ATA is diagonal; when the columns all have the same power (sum of squares of elements), then that matrix is an identity matrix times a constant, so the inversion is trivial. The latter is the case when the sample times are equally spaced and sinusoids chosen as sines and cosines equally spaced in pairs on the frequency interval 0 to a half cycle per sample (spaced by 1/N cycles per sample, omitting the sine phases at 0 and maximum frequency where they are identically zero). This case is known as the discrete Fourier transform, slightly rewritten in terms of measurements and coefficients. x = A T ϕ {\displaystyle x={\textbf {A}}^{\mathrm {T} }\phi } — DFT case for N equally spaced samples and frequencies, within a scalar factor. === The Lomb method === Trying to lower the computational burden of the Vaníček method in 1976 (no longer an issue), Lomb proposed using the above simplification in general, except for pair-wise correlations between sine and cosine bases of the same frequency, since the correlations between pairs of sinusoids are often small, at least when they are not tightly spaced. This formulation is essentially that of the traditional periodogram but adapted for use with unevenly spaced samples. The vector x is a reasonably good estimate of an underlying spectrum, but since we ignore any correlations, Ax is no longer a good approximation to the signal, and the method is no longer a least-squares method — yet in the literature continues to be referred to as such. Rather than just taking dot products of the data with sine and cosine waveforms directly, Scargle modified the standard periodogram formula so to find a time delay τ {\displaystyle \tau } first, such that this pair of sinusoids would be mutually orthogonal at sample times t j {\displaystyle t_{j}} and also adjusted for the potentially unequal powers of these two basis functions, to obtain a better estimate of the power at a frequency. This procedure made his modified periodogram method exactly equivalent to Lomb's method. Time delay τ {\displaystyle \tau } by definition equals to tan 2 ω τ = ∑ j sin 2 ω t j ∑ j cos 2 ω t j . {\displaystyle \tan {2\omega \tau }={\frac {\sum _{j}\sin 2\omega t_{j}}{\sum _{j}\cos 2\omega t_{j}}}.} Then the periodogram at frequency ω {\displaystyle \omega } is estimated as: P x ( ω ) = 1 2 [ [ ∑ j X j cos ω ( t j − τ ) ] 2 ∑ j cos 2 ω ( t j − τ ) + [ ∑ j X j sin ω ( t j − τ ) ] 2 ∑ j sin 2 ω ( t j − τ ) ] , {\displaystyle P_{x}(\omega )={\frac {1}{2}}\left[{\frac {\left[\sum _{j}X_{j}\cos \omega (t_{j}-\tau )\right]^{2}}{\sum _{j}\cos ^{2}\omega (t_{j}-\tau )}}+{\frac {\left[\sum _{j}X_{j}\sin \omega (t_{j}-\tau )\right]^{2}}{\sum _{j}\sin ^{2}\omega (t_{j}-\tau )}}\right],} which, as Scargle reports, has the same statistical distribution as the periodogram in the evenly sampled case. At any individual frequency ω {\displaystyle \omega } , this method gives the same power as does a least-squares fit to sinusoids of that frequency and of the form: ϕ ( t ) = A sin ω t + B cos ω t . {\displaystyle \phi (t)=A\sin \omega t+B\cos \omega t.} In practice, it is always difficult to judge if a given Lomb peak is significant or not, especially when the nature of the noise is unknown, so for example a false-alarm spectr