Fact-checked by Grok 2 weeks ago

Generator

An is an electromechanical device that converts into through the process of . This conversion relies on Faraday's law, where relative motion between conductors and magnetic fields induces an , generating current in a . The underlying principle was experimentally demonstrated by in 1831 using a disk rotating between poles, marking the foundational step toward practical generation. Generators are classified primarily by output type—alternating current (AC) or (DC)—with AC generators predominant in utility-scale applications due to simpler construction and compatibility with long-distance transmission via transformers. DC generators, employing commutators to output, suit applications requiring steady , such as charging or certain . Beyond electrical classification, generators vary by (e.g., steam turbines in power plants, internal combustion engines in portables) and scale, from massive hydroelectric units producing gigawatts to compact standby models for residential backup during outages. Their deployment underpins global supply, converting diverse sources—fossil fuels, , renewables—into usable power, though efficiency and emissions depend on fuel and design, with modern advancements emphasizing reduced losses and integration with intermittent renewables.

Electric Generators

Definition and Principles

An is a device that converts into through the process of . This conversion occurs when mechanical input, such as from a or , causes relative motion between conductors and a , generating an (EMF). Generators differ from batteries, which store , by requiring continuous mechanical work to produce , with typically ranging from 80% to 98% depending on design and load. The core principle governing electric generators is Faraday's law of electromagnetic induction, formulated by in , which asserts that the magnitude of the induced in a closed circuit equals the negative rate of change of through the circuit. Mathematically, this is expressed as \mathcal{E} = -N \frac{d\Phi_B}{dt}, where \mathcal{E} is the , N is the number of turns in the coil, and \Phi_B is the . complements this by specifying the direction of the induced current, stating that it opposes the change in that produced it, conserving energy in the system. In generator operation, this opposition manifests as a mechanical resisting rotation, necessitating input power to sustain generation. Basic generator construction involves a —often with coils or magnets—and a stationary , where conductors experience a time-varying due to . For () generators, the induced varies sinusoidally as \mathcal{E} = NBA\omega \sin(\omega t), with B as strength, A as coil area, and \omega as . () generators employ a to the output, ensuring unidirectional flow, though slip rings are used in AC synchronous generators for steady without . These principles underpin scalability from small portable units outputting kilowatts to large power plant generators producing gigawatts.

Historical Development

The principle of , foundational to electric generators, was discovered by in 1831 through experiments demonstrating that a changing induces an electric current in a conductor. Faraday constructed the first electromagnetic generator, known as the Faraday disk or , consisting of a copper disk rotating between the poles of a permanent , which produced a continuous . This device, though low-output and inefficient for practical power generation, established the theoretical basis for converting into without chemical batteries. In 1832, French instrument maker developed the first , building directly on Faraday's principle by incorporating a rotating and stationary coils to generate , later rectified to via a . Pixii's machine marked the initial step toward a mechanical generator capable of sustained output, though it relied on permanent magnets and produced limited power suitable only for demonstration. Significant advancements occurred in 1866 when identified the dynamo-electric principle, enabling generators to self-excite using electromagnets powered by their own output current, eliminating the need for permanent magnets and allowing scalable voltage and power. and his brothers constructed the first functional self-exciting , which powered early electric lighting and systems, facilitating commercial viability. This innovation spurred widespread adoption in industry during the late 19th century. The late 1880s saw the rise of (AC) generators, with early forms tracing back to Pixii's AC output but gaining practicality through polyphase designs. Nikola Tesla's 1887-1888 patents for AC induction motors and polyphase systems, licensed to , enabled efficient long-distance transmission via AC generators, demonstrated at the 1893 . Concurrently, Charles S. Bradley developed the first three-phase AC generator in 1887, enhancing stability for large-scale power plants. These developments shifted dominance from dynamos to AC systems, underpinning modern electrical grids.

Types and Technologies

Electric generators are primarily classified into alternating current (AC) and direct current (DC) types according to the form of electrical output they produce. AC generators, or alternators, generate sinusoidal waveforms suitable for most modern power systems, while DC generators deliver unidirectional current, often rectified via a commutator for applications requiring steady polarity. This classification stems from the fundamental principle of electromagnetic induction, where mechanical rotation induces voltage in conductors within a magnetic field. AC generators are further divided into synchronous and asynchronous (induction) subtypes. Synchronous generators operate at a rotor speed exactly matching the synchronous speed determined by the supply frequency and number of poles, enabling precise frequency control essential for grid-connected power generation. Their rotors feature either salient poles, which project outward and suit low-speed applications like hydroelectric turbines due to higher pole counts and better stability under variable loads, or cylindrical (non-salient) rotors, which provide uniform air gaps and higher mechanical strength for high-speed steam or gas turbines. Asynchronous generators, in contrast, run slightly above synchronous speed (with positive slip) and derive excitation from residual magnetism or external capacitors, making them simpler and cheaper but less efficient for standalone operation without grid support; they are common in wind turbines for their variable-speed tolerance. Synchronous generator technologies emphasize methods to create the rotor's . Wound-field rotors use electromagnets energized by via slip rings and brushes or brushless systems employing a rotating exciter with , allowing adjustable field strength for and reactive power control. Permanent magnet synchronous generators (PMSGs) employ rare-earth s like for , eliminating field windings and brushes to reduce and losses, though they offer fixed and vulnerability to demagnetization at high temperatures; PMSGs achieve higher (up to 95%) in small-scale renewables but incur higher material costs. Brushless designs predominate in modern utility-scale units to minimize wear, with sourced from auxiliary permanent generators (PMGs) or static exciters for rapid response. DC generators, less common today due to rectification advancements in AC systems, classify by field winding configuration relative to the armature. Shunt-wound generators connect the field in parallel with the armature, providing stable voltage under varying loads via self-regulation. Series-wound types series-connect the field to the armature, yielding high starting but voltage droop with load, suited for charging. Compound-wound generators combine shunt and series fields—either cumulatively for balanced voltage or differentially for specialized arcs—subdivided into short-shunt (shunt across armature only) and long-shunt (shunt across armature plus series) for optimized no-load and full-load performance. These rely on mechanical commutation, limiting efficiency and scalability compared to AC counterparts.

Applications and Societal Impact

Electric generators form the core of large-scale power production worldwide, converting from turbines into electrical power for grid supply. In thermal power plants, turbines driven by fuels or reactions power generators that supplied the majority of global in 2024, with coal-fired facilities alone contributing significantly to baseload despite declining shares. Hydroelectric generators, harnessing flow through , accounted for 14.3% of worldwide , while wind and solar installations increasingly integrate generators or inverters for variable renewable output. In plants, generators ensure reliable output and support safety systems like reactor cooling during shutdowns. Portable and standby generators, often diesel- or gasoline-fueled, deliver critical power for residential, , and needs during failures, , or remote operations. The U.S. reports their extensive deployment at construction sites, hospitals, and military bases, preventing economic losses estimated in billions annually from outages. Automotive alternators, a compact generator variant, recharge batteries and supply electrical loads, enabling widespread personal and electrification. In renewables, generators in systems provide dispatchable power to mitigate , supporting stability as low-carbon sources reached 40% of global generation in 2024. The deployment of electric generators has profoundly shaped modern society by enabling universal electrification, which boosted productivity, extended productive hours via lighting, and facilitated advancements in healthcare, communication, and manufacturing since the late 19th century. In the U.S., the electric power industry generates an annual economic impact of $880 billion, equivalent to 5% of GDP, through direct employment of over 7 million workers and multiplier effects in supply chains. Globally, reliable generator-based power correlates with higher GDP per capita and reduced poverty, as electricity access triples the likelihood of high-skilled employment in developing regions. Conversely, generators reliant on fossil fuels—responsible for over 60% of in recent years—exacerbate , emitting pollutants that contribute to air quality issues, contamination from thermal discharges, and land disruption from . generation accounts for more than 40% of energy-related CO2 emissions worldwide, driving climate forcing through greenhouse gases, with fossil plants also releasing toxic metals into ecosystems. While low-carbon generator technologies like nuclear and hydro yield fewer injuries (0.06 per TWh) and fatalities (0.001 per TWh) than fossil alternatives, the sector's overall externalities impose health costs from and , disproportionately affecting low-income communities near facilities. Transitioning to diverse generator fleets, including advanced gas turbines for peaking, underscores ongoing trade-offs between reliability, cost, and ecological sustainability. Recent advancements in technology have focused on enhancing and reducing dependencies, particularly through superconducting materials and compositions. High-temperature superconducting (HTS) generators have demonstrated potential for significant improvements, with GE successfully trialing HTS windings in 2025 that operate at elevated temperatures, minimizing cryogenic cooling requirements compared to traditional low-temperature superconductors. These developments enable lighter, more compact designs suitable for turbines, where the U.S. Department of Energy funded projects yielding superconducting generators with stronger magnetic fields and reduced weight. Concurrently, efforts to eliminate rare-earth elements in permanent synchronous generators (PMSGs) have advanced, with Ames Laboratory developing a rare-earth-free bonded in 2025 that maintains at temperatures up to 100°C, doubling performance under heat stress. Companies like GreenSpur have commercialized rare-earth-free generators, addressing supply chain vulnerabilities amid geopolitical tensions over rare-earth mining. Asynchronous generators have seen reliability enhancements through technological refinements, contributing to market growth projected at a 9.8% CAGR from 2026 to 2033, driven by applications in variable-speed renewable systems. Synchronous generator markets, valued at USD 5.58 billion in 2025, reflect adoption of modular designs for easier and in power generation. Future trends emphasize integration with sources and further gains to support decarbonization goals, such as tripling global renewable capacity by 2030. Superconducting technologies are poised for broader deployment in and applications, offering efficiencies up to 99.5% in electric motors adaptable to generators, alongside reduced losses via advanced cryogenic systems. Rare-earth-free permanent magnets, including those from Niron Magnetics scaling production in 2024, will likely dominate in sustainable designs, mitigating environmental impacts of rare-earth extraction while maintaining high performance in PMSGs for offshore and electric vehicles. systems combining generators with and , augmented by for , are expected to prevail in backup power and microgrids, aligning with IRENA's doubling targets by 2030. These innovations prioritize causal factors like material physics and electromagnetic optimization over unsubstantiated claims, ensuring verifiable reductions in energy losses and operational costs.

Mathematical Generators

In Abstract Algebra and Group Theory

In group theory, a generating set of a group G is a S \subseteq G such that every of G can be expressed as a finite product of elements from S union their inverses, meaning the generated by S, denoted \langle S \rangle, equals G. A group is finitely generated if it admits a finite generating set; otherwise, it may require infinitely many generators, as in the additive group of real numbers (\mathbb{R}, +), which lacks even a countable generating set due to its uncountable over \mathbb{Q}. The full set G itself always serves as a generating set, though typically redundant, while minimal generating sets (those with no proper generating G) capture essential structure, and their is the group's when finite. A special case arises with cyclic groups, generated by a single element g \in G where \langle g \rangle = G and the order of g equals |G| if finite. For the cyclic group \mathbb{Z}/n\mathbb{Z}, generators are integers k with \gcd(k, n) = 1, such as for n=12, the generators 1, 5, 7, and 11, since each produces all residue classes via multiples modulo 12. Not every element generates the group; for instance, in \mathbb{Z}/6\mathbb{Z}, 2 has order 3 and generates only the subgroup {0, 2, 4}. Infinite cyclic groups like \mathbb{Z} are generated by 1 or -1 alone. Groups are often presented via generators and relations, specifying a set S and equations that the generators satisfy, defining G as the quotient of the free group on S by the normal closure of those relations; for example, the dihedral group of order $2n has presentation \langle r, s \mid r^n = s^2 = 1, srs^{-1} = r^{-1} \rangle. Free groups on n generators consist of reduced words over S \cup S^{-1} with no further relations, serving as universal objects for n-generated groups. In broader , analogous concepts apply to and , where generators span ideals or the module via ring operations, but group-theoretic generation emphasizes products and inverses without additional structure.

Generating Functions and Series

Generating functions provide a formal method to encode sequences of numbers as coefficients in a power series, enabling algebraic manipulation to derive properties of the sequence. For an infinite sequence \{a_n\}_{n=0}^\infty, the ordinary generating function is defined as G(x) = \sum_{n=0}^\infty a_n x^n, where the coefficients a_n capture combinatorial counts or recurrence solutions. This representation transforms sequence problems into operations on rational functions or polynomials, leveraging series expansions for closed-form expressions. Two primary variants address distinct combinatorial structures: ordinary generating functions suit unlabeled objects, such as partitions or unrestricted sequences, while exponential generating functions, E(x) = \sum_{n=0}^\infty a_n \frac{x^n}{n!}, account for labeled permutations by incorporating factorials to normalize for symmetries in ordered sets. For instance, the exponential generating function for permutations of n elements yields E(x) = e^x, reflecting n! as the coefficient of x^n / n!, derived from the exponential series expansion. Operations like multiplication correspond to convolution of sequences—product of generating functions G(x)H(x) encodes sums \sum_k a_k b_{n-k}—facilitating enumeration of composite structures, as in the binomial theorem where (1 - x)^{-k} generates combinations with repetition. In solving linear recurrences, generating functions convert difference equations into algebraic equations by summing the relation multiplied by x^n. For the defined by f_0 = 0, f_1 = 1, f_n = f_{n-1} + f_{n-2} for n \geq 2, the F(x) = \sum f_n x^n satisfies F(x) = \frac{x}{1 - x - x^2} after substituting the recurrence and solving, yielding coefficients via partial fractions or series reversion. This method extends to higher-order recurrences with constant coefficients, where the denominator polynomial mirrors the , providing explicit solutions without matrix methods. Applications in include deriving partition functions, such as the \prod_{k=1}^\infty (1 - x^k)^{-1} for integer partitions, whose coefficients count unrestricted sums equaling n. Generating series generalize these to Laurent or Puiseux forms for , but core utility lies in exact and proof via bijective correspondences between series coefficients and combinatorial objects. Limitations arise in non-rational cases, requiring or singularity analysis for growth rates, as formalized in Flajolet and Sedgewick's analytic . Empirical validation through computational verification confirms these derivations, underscoring generating functions' role in bridging algebraic manipulation with discrete counting principles.

Other Mathematical Contexts

In linear algebra, a generating set (or spanning set) for a V over a F is a subset S \subseteq V such that every vector in V can be expressed as a finite linear combination of elements from S with coefficients in F. The span of S, denoted \operatorname{span}(S), is the smallest subspace containing S, and S generates V if \operatorname{span}(S) = V. A basis for V is a minimal generating set that is also linearly independent, with the dimension of V equal to the cardinality of any basis. In , a , also known as a , is an integer g coprime to n such that the powers g^k \mod n for k = 1, 2, \dots, \phi(n) produce all integers coprime to n, where \phi is ; thus, g generates the (\mathbb{Z}/n\mathbb{Z})^*. exist precisely when n = 1, 2, 4, p^k, or $2p^k for odd prime p and positive integer k. For prime p, the of g p must be p-1, ensuring it cycles through all nonzero residues p. In category theory, a generator of a category \mathcal{C} is an object G such that the representable functor \operatorname{Hom}_{\mathcal{C}}(G, -) is faithful, meaning distinct morphisms f, g: X \to Y satisfy \operatorname{Hom}_{\mathcal{C}}(G, f) \neq \operatorname{Hom}_{\mathcal{C}}(G, g), thereby distinguishing arrows via maps from G. Compact projective generators play a key role in characterizations like Gabriel's theorem, identifying module categories as abelian categories with such a generator.

Computing and Programming Generators

Iterator and Yield-Based Generators

Iterator-based generators, often implemented via the yield keyword in languages like Python, enable the creation of iterators that produce values on demand, suspending and resuming execution to maintain state without storing the entire sequence in memory. An iterator is an object adhering to the iterator protocol, supporting methods for initialization and sequential value retrieval, typically via __iter__ and __next__ in Python. Generators simplify iterator implementation by transforming ordinary functions into stateful iterators through yield, which pauses execution after producing a value, preserving local variables and control flow for resumption on subsequent calls. This contrasts with traditional iterators, which require explicit class-based implementation of iteration methods, whereas generators leverage function syntax for conciseness. Introduced in 2.2 on December 16, 2001, via PEP 255, the yield statement addressed the need for lightweight, coroutine-like iteration without full systems, drawing inspiration from languages like for efficient sequence generation. Upon invocation, a generator returns an object without immediate execution; the first next() call runs the function until the initial yield, outputting a value and suspending, with subsequent calls resuming from that point until exhaustion or another yield. This mechanism supports bidirectional communication in advanced forms, as later enhancements like PEP 342 in Python 2.5 allowed send() for passing values back to the generator, evolving toward capabilities. (Note: While PEP 342 details are from Python 2.5 release notes, confirming evolution from basic yield.) The primary advantages include memory efficiency for large or infinite sequences, as computation occurs lazily during iteration rather than upfront allocation, reducing overhead in scenarios like processing streams or Fibonacci sequences. For instance, a generator for even numbers can yield indefinitely without memory exhaustion:
python
def evens():
    n = 0
    while True:
        yield n
        n += 2
Iterating via for num in evens(): produces values sequentially, halting on explicit StopIteration. Similar yield-based constructs appear in JavaScript (via function* since ECMAScript 2015) and C# (yield return since .NET 2.0 in 2005), enabling iterable objects with pausing semantics, though implementation details vary—JavaScript generators support delegation via yield*. These features enhance readability and performance over manual iterator classes, particularly for one-pass data processing, but require careful handling of exceptions and closure to avoid resource leaks. Empirical benchmarks show generators using significantly less RAM than lists for million-element sequences, confirming their utility in resource-constrained environments.

Code and Data Generators

Code generators in are automated systems that produce executable or source code from higher-level specifications, models, or intermediate representations, thereby reducing manual coding effort and minimizing errors in repetitive tasks. These tools are integral to compiler backends, where they translate intermediate code into target machine instructions, as seen in simple code generators that optimize and address computation for efficiency. Parser generators, such as or , exemplify this by deriving lexical analyzers and syntax parsers from definitions, enabling rapid prototyping of language processors. Data generators complement code generation by synthesizing input datasets for software validation, particularly in testing scenarios where real data is scarce, sensitive, or insufficient for coverage. They employ techniques like random sampling, rule-based fabrication, or model-driven synthesis to create realistic yet controlled data, supporting unit tests, integration checks, and performance simulations. Tools such as Mockaroo facilitate this by producing customizable , , or SQL datasets mimicking production schemas, which aids in reproducing edge cases without risks. In practice, code and data generators often integrate within development pipelines; for instance, generators might output boilerplate with embedded generation routines for demo purposes, as in scripting languages like where procedural scripts populate fixtures from definitions. This synergy enhances scalability, with empirical studies showing automated test generation via program execution and minimization reducing manual effort by targeting path coverage effectively. However, challenges persist, including ensuring generated 's maintainability and 's fidelity to real-world distributions, necessitating validation against domain constraints.

Random Number and Pseudorandom Generators

Random number generators (RNGs) in produce sequences of numbers intended to simulate for applications such as simulations, , and statistical sampling. They are categorized into true random number generators (TRNGs), which derive from unpredictable physical processes, and pseudorandom number generators (PRNGs), which employ deterministic algorithms to generate sequences that statistically mimic but are reproducible given the same initial . TRNGs rely on hardware mechanisms to capture inherent unpredictability, such as thermal noise in resistors, quantum fluctuations in photodiodes, or radioactive decay timings, ensuring outputs are non-deterministic and resistant to prediction even with full knowledge of prior outputs. These methods provide high-entropy bits but often at lower generation rates compared to software-based alternatives, with bit rates typically ranging from kilobits to megabits per second depending on the implementation. PRNGs, by contrast, start from a seed value—often sourced from a TRNG or system pool—and apply mathematical transformations to produce long sequences efficiently. Common algorithms include the (LCG), defined by the recurrence X_{n+1} = (a X_n + c) \mod m, where parameters a, c, and m determine and quality; though simple and fast, LCGs can exhibit detectable patterns if poorly parameterized. The , developed in 1997 by Makoto Matsumoto and Takuji Nishimura, offers a of $2^{19937} - 1 and passes extensive statistical tests, making it suitable for non-cryptographic simulations despite vulnerabilities to in contexts. Quality assessment of both TRNGs and PRNGs involves statistical test suites to verify uniformity, independence, and absence of correlations. The NIST Statistical Test Suite, revised in 2010, includes 15 tests such as frequency, runs, and , applied to sequences from generators; passing requires at least 97.5% of test instances to succeed at a 1% significance level. The Diehard battery, created by George Marsaglia in 1995 and extended in Dieharder, evaluates properties like overlapping permutations and parking lot simulations on large datasets, often exceeding 10 GB, to detect subtle flaws. In practice, cryptographically secure PRNGs (CSPRNGs) like those based on hash functions or block ciphers are preferred for applications, periodically reseeded by TRNG entropy to mitigate predictability risks from compromise. Historical development traces to mid-20th-century , with the first PRNG implemented on the in 1951 using electronic noise, evolving from manual tables like RAND Corporation's 1947 million-digit compilation to modern hybrid systems balancing speed and .

Signal and Specialized Generators

In Electronics and Circuit Theory

Signal generators in produce electrical waveforms with precisely controlled attributes, including , , phase, and shape, serving as stimuli for testing and analyzing . These devices enable engineers to inject known signals into systems to measure responses, such as , , or transient behavior, fundamental to validation and . In theory, they operationalize ideal source models—voltage or current generators with negligible —allowing empirical verification of theoretical predictions like Bode plots or Nyquist criteria. Function generators, a primary type, synthesize standard periodic waveforms including sine, square, , and ramp forms, typically spanning frequencies from sub-Hz to tens of MHz with amplitudes up to 20 V peak-to-peak. Their core relies on relaxation oscillators or phase-shift networks, such as the for low-distortion sine waves, followed by shaping circuits like comparators for square waves and integrators for triangles. Arbitrary waveform generators (AWGs) extend this by storing digitized custom s in memory and replaying them via digital-to-analog conversion, supporting complex signals for mixed-signal testing with sampling rates exceeding 1 GS/s in modern units. Pulse generators deliver short, high-fidelity for evaluation, characterizing rise/fall times, , and delays in logic gates or interconnects, often with widths adjustable from nanoseconds to seconds. RF and microwave signal generators handle sinusoidal carriers up to GHz ranges, incorporating capabilities like AM, , or to simulate communication signals, crucial for receiver sensitivity tests under standards such as IEEE 802.11. Vector signal generators add for complex schemes, enabling impairment emulation in or systems. In applications, these generators facilitate impedance via reflection coefficients, analysis using Y-factor methods, and simulations by providing excitation for nonlinear characterization. For instance, a might drive an under test at 1 kHz to quantify flatness, with output monitored via analyzers for products. Precision in generation—often with below -100 dBc/Hz at 10 kHz offset—ensures accuracy, though real devices include attenuators and filters to match 50-ohm impedances and suppress harmonics. Limitations arise from challenges in multi-channel setups and the need for to maintain to standards like those from NIST.

In Audio and Music Synthesis

In audio and music synthesis, generators encompass modules that produce the foundational signals for sound creation, including oscillators for generating periodic waveforms and envelope generators for temporal shaping of those signals. Oscillators, often voltage-controlled (VCOs), output basic waveforms like sine, square, triangle, or sawtooth at audio frequencies (typically 20 Hz to 20 kHz), serving as the primary tone sources in subtractive, additive, and other synthesis methods. These waveforms provide the raw harmonic content, with sine waves representing pure tones without harmonics, while square and sawtooth waves introduce richer overtones due to their discontinuous nature, enabling timbral variety when filtered or modified. Envelope generators (EGs) create time-varying control voltages to modulate parameters such as , , or filter cutoff, typically following the ADSR model: (rise time from zero to peak), (fall to sustain level), sustain (held level during note duration), and release (fall to zero after note-off). This model, implemented in analog synthesizers since the , allows precise control over sound evolution, mimicking natural instrument transients like a piano's or a violin's gradual . Digital implementations, common in software synthesizers, often extend ADSR with multi-stage curves or looping for evolving textures. Low-frequency oscillators (LFOs), operating below audio range (under 20 Hz), function as specialized generators for , applying cyclic variations to (vibrato), (tremolo), or via waveforms like triangles or random . Function generators in modular systems generalize this, producing arbitrary voltage contours triggered by gates, bridging envelope and oscillator roles for complex patching. generators, producing aperiodic signals like white or , add stochastic elements for percussion, wind effects, or as modulation sources, with shaping (e.g., rolling off high frequencies at -3 /) enhancing realism. These generators interact causally in synthesis chains: oscillators provide signals, envelopes and shape them, and modulators like LFOs introduce dynamics, with analog instability (e.g., oscillator drift from component tolerances) contributing organic imperfections absent in perfect digital replicas. Empirical testing shows analog generators yield warmer tones due to harmonic distortion from nonlinearities, though digital equivalents achieve parity via and modeling, as verified in waveform analysis.

Generative Systems in Artificial Intelligence

Core Concepts and Architectures

Generative models constitute a class of frameworks designed to approximate the p(\mathbf{x}) from which observed data samples are drawn, enabling the synthesis of novel data instances that statistically mimic the training set. These models operate by learning latent representations or transformations that reverse-engineer the data process, contrasting with discriminative models that estimate conditional distributions p(\mathbf{y}|\mathbf{x}) for or tasks. Empirical success in domains such as image synthesis and text stems from scalable neural architectures trained on vast datasets, though challenges like mode collapse—where models fail to capture data diversity—persist due to optimization instabilities in high-dimensional spaces. A foundational architecture is the Generative Adversarial Network (GAN), introduced in , comprising a generator network G that maps random noise vectors \mathbf{z} to G(\mathbf{z}), pitted against a discriminator D trained to distinguish real data from fakes via a objective: \min_G \max_D \mathbb{E}_{\mathbf{x} \sim p_{\text{data}}}[\log D(\mathbf{x})] + \mathbb{E}_{\mathbf{z} \sim p_{\mathbf{z}}}[\log(1 - D(G(\mathbf{z})))]. This adversarial training fosters implicit without explicit likelihood computation, yielding high-fidelity outputs in applications like photorealistic image generation, as evidenced by variants achieving FID scores below 5 on datasets such as by 2017. However, training instability arises from non-convergent equilibria, often mitigated by techniques like spectral normalization. Variational Autoencoders (VAEs), proposed in , extend principles with probabilistic latent variables, employing an encoder to infer approximate posterior q(\mathbf{z}|\mathbf{x}) and a for likelihood p(\mathbf{x}|\mathbf{z}), optimized via the (ELBO): \mathcal{L} = \mathbb{E}_{q(\mathbf{z}|\mathbf{x})}[\log p(\mathbf{x}|\mathbf{z})] - D_{\text{KL}}(q(\mathbf{z}|\mathbf{x}) || p(\mathbf{z})). This variational inference framework enables amortized inference and smooth latent traversals, facilitating disentangled representations in tasks like facial attribute , with reconstruction errors on MNIST dropping to near-zero levels in early implementations. VAEs excel in structured output but often produce blurrier samples than GANs due to pixel-independent assumptions in the decoder. Contemporary advancements include diffusion models, formalized in 2020 through denoising diffusion probabilistic models (DDPMs), which iteratively corrupt data via a forward adding over T timesteps, then train a to reverse this process by predicting noise \epsilon at each step: \mathbf{x}_{t-1} = \frac{1}{\sqrt{\alpha_t}} \left( \mathbf{x}_t - \frac{1 - \alpha_t}{\sqrt{1 - \bar{\alpha}_t}} \epsilon_\theta(\mathbf{x}_t, t) \right) + \sigma_t \mathbf{z}. Sampling from pure noise yields diverse, high-quality outputs, surpassing GANs in Inception scores on subsets (e.g., 200+ vs. prior 100+), driven by stable training via variance-preserving noise schedules. Diffusion's causal chain of perturbations aligns with thermodynamic analogies, enabling conditional generation in systems like , released in 2022. Autoregressive models, particularly transformer-based variants, decompose joint distributions into sequential conditionals p(\mathbf{x}) = \prod_{i=1}^n p(x_i | \mathbf{x}_{<i}), leveraging self-attention mechanisms to capture long-range dependencies without recurrence. In generative contexts, decoder-only transformers like architectures process tokenized sequences autoregressively, achieving reductions from 20+ to under 10 on benchmarks such as WikiText-103 through scaling laws correlating loss with parameter count (e.g., L \approx A N^{-0.095} for N parameters). This enables coherent long-form text generation, though outputs remain token-conditioned and prone to repetition without external safeguards.

Applications and Achievements

Generative AI systems have found extensive applications in , enabling tasks such as automated content creation, , and dialogue systems. Large language models like those powering chatbots and writing assistants process vast datasets to produce coherent text for summarization, , and automation, with adoption doubling from 2023 to 2024 among surveyed firms. In , these models assist in , , and generating functional code snippets, enhancing developer productivity by automating repetitive tasks. In image and video synthesis, generative models such as diffusion-based architectures produce photorealistic visuals from textual prompts, supporting applications in , , and film . Milestones include the development of in 2022 for open-source image generation and subsequent advancements like Meta's Emu Video in 2023 for text-to-video creation, which automate scene transitions and editing. These tools extend to in , where synthetic images improve training datasets for and analysis. Scientific domains benefit from generative AI in and , where models design novel molecular structures by optimizing chemical properties against empirical constraints. A key achievement occurred in 2020 when Exscientia advanced the first AI-generated candidate, DSP-1181, into clinical trials for obsessive-compulsive , demonstrating reduced design timelines from years to months. Further, generative approaches generate synthetic to accelerate target identification, with studies showing potential cost reductions in R&D pipelines. Applications also include variants and simulation of chemical reactions, aiding generation grounded in physical laws. In and creative sectors, generative drives gains, with empirical studies average worker output increases of 14% in tasks like writing and , particularly among lower-skilled performers. detection leverages anomaly-pattern recognition in , while uses scenario simulations for . Achievements include global private in generative reaching $33.9 billion in 2024, an 18.7% rise from 2023, reflecting scaled deployments in industries from to . These systems have also enabled breakthroughs in music and personalized , though real-world depends on validation against causal benchmarks rather than isolated metrics.

Criticisms, Risks, and Empirical Realities

Generative AI models, particularly large language models and diffusion-based systems, exhibit a propensity for hallucinations, where they generate plausible but factually incorrect information. Empirical assessments of AI-assisted legal research tools revealed hallucination rates ranging from 17% to 33%, undermining assertions of citation accuracy by vendors such as LexisNexis and Thomson Reuters. These errors stem from probabilistic generation mechanisms that prioritize pattern matching over verifiable truth, as opposed to deliberate deception, with studies classifying hallucinations into categories like factual distortions and confabulations derived from training data gaps. In scientific and informational contexts, such outputs have propagated misinformation, as evidenced by fabricated references in academic-style generations. Biases in generative outputs often amplify present in training corpora, leading to discriminatory content. Evaluations of large language models demonstrated consistent and racial biases in AI-generated articles, with outputs favoring certain demographics over others in portrayals. Experimental research confirmed that without explicit safeguards, these systems exacerbate existing societal biases, though interventions can mitigate or reverse them in controlled settings. Such patterns reflect causal dependencies on non-representative or historically skewed datasets, rather than inherent model neutrality, with peer-reviewed analyses highlighting under-representation and selection effects in domains like legal texts. Legal challenges have mounted over the use of copyrighted materials in , with plaintiffs arguing infringement through unauthorized ingestion of protected works. Over two dozen U.S. lawsuits by 2024 targeted firms including , Stability AI, and , alleging that scraping books, images, and articles for model development violates doctrines. A 2025 federal ruling in marked the first substantive victory for plaintiffs, denying motions to dismiss claims against an AI developer for on licensed datasets without permission. Courts have grappled with whether outputs constitute derivative works, with ongoing cases testing transformative use under precedents like those for search engines, though generative replication risks remain unresolved. Misuse of generative tools for deepfakes and introduces risks of deception and societal disruption. Realistic audio, video, and text forgeries have facilitated campaigns, schemes, and election interference, with documented instances eroding trust in visual . Systematic reviews identify political deepfakes as a vector for , distinct from traditional fakes due to AI's scalability and indistinguishability. Empirical tracking shows adversaries leveraging these for and , amplifying threats in low-verification environments like . Economic and environmental externalities compound these concerns. Labor market data indicate displacement in AI-exposed sectors: freelancers in writing and fields experienced a 2% drop in contracts and 5% earnings decline post-generative tool adoption, correlating with task . High-exposure occupations showed unemployment rises aligned with AI integration intensity, per occupational analyses from 2023-2025. Environmentally, training and inference demands strain resources; U.S. data centers consumed 4.4% of national electricity in 2023, projected to triple by 2028 amid AI growth, alongside substantial water usage for cooling—equivalent to thousands of households per model run. E-waste from specialized and non-recyclable components in GPUs further burdens , though some models optimize for lower-impact . These realities underscore that while generative AI yields capabilities, its unchecked scaling incurs verifiable costs without proportional safeguards.

References

  1. [1]
    8.7: The Electric Generator - Engineering LibreTexts
    Sep 12, 2022 · A generator is a device that transforms mechanical energy into electrical energy, typically by electromagnetic induction via Faraday's Law.
  2. [2]
    How electricity is generated - U.S. Energy Information Administration ...
    An electric generator is a device that converts a form of energy into electricity. There are many different types of electricity generators.
  3. [3]
    Michael Faraday's generator | Royal Institution
    Faraday created the first transformer in August 1831. A few months later he designed and made this simple piece of apparatus based on his ring, developing the ...
  4. [4]
    History of Electric Generators - BYJU'S
    Sep 23, 2019 · Electric generators, also known as dynamos is an electric machine that converts mechanical energy into electrical energy. The electric ...
  5. [5]
    Generators - Engineering and Technology History Wiki
    Feb 19, 2020 · One of the most important inventors of generators was German Werner von Siemens, who designed improved DC generators and called them dynamos. An ...
  6. [6]
    Electric Generators | Physics - Lumen Learning
    Electric generators induce an emf by rotating a coil in a magnetic field, as briefly discussed in Induced Emf and Magnetic Flux.
  7. [7]
    23.5 Electric Generators – College Physics
    electric generator: a device for converting mechanical work into electric energy; it induces an emf by rotating a coil in a magnetic field · emf induced in a ...
  8. [8]
    13.6 Electric Generators and Back Emf – University Physics Volume 2
    An electric generator rotates a coil in a magnetic field, inducing an emf given as a function of time by ϵ = N B A ω sin ( ω t ) where A is the area of an N- ...
  9. [9]
    Electromagnetic Induction and Faradays Law - Electronics Tutorials
    Electromagnetic induction uses the relationship between electricity and magnetism whereby an electric current flowing through a single wire will produce a ...
  10. [10]
    Electric Generators – ISP209: The Mystery of the Physical World
    A generator with a single rectangular coil rotated at constant angular velocity in a uniform magnetic field produces an emf that varies sinusoidally in time.
  11. [11]
    [PDF] 0420 - E111 - Chapter 09 - Generator Exciter and Voltage Regulator.
    Electromagnetic induction, the basic principle of generator operation, involves the movement of an electrical conductor through a magnetic field. Figure 9-1 ...
  12. [12]
    Faraday (1791) - Energy Kids - EIA
    In 1831, using his "induction ring", Faraday made one of his greatest discoveries - electromagnetic induction: the "induction" or generation of electricity in a ...
  13. [13]
    Pixii Machine - Magnet Academy - National MagLab
    French instrument maker Hippolyte Pixii harnessed these ideas in 1832 with the Pixii Machine. It was the first practical mechanical generator of electrical ...
  14. [14]
    Electricity timeline - Energy Kids - EIA
    Sir Charles Algernon Parsons (England) invented a steam turbine generator, capable of generating huge amounts of electricity. 1886. William Stanley, Jr. (United ...
  15. [15]
    Dynamo machine - Siemens Global
    Werner von Siemens laid the foundations for a new age of electricity. Faster than others, he recognized the potential of the dynamoelectric principle.Missing: improvements | Show results with:improvements
  16. [16]
    Werner von Siemens - Magnet Academy - National MagLab
    In 1866, the research of Werner von Siemens would lead to his discovery of the dynamo electric principle that paved the way for the large-scale generation ...
  17. [17]
    Tesla (1856) - Energy Kids - EIA
    Westinghouse used Tesla's alternating current system to light the World's Columbian Exposition of 1893 in Chicago. Then in 1896, Tesla's system was used at ...
  18. [18]
    AC Power History - Edison Tech Center
    1887 - C.S. Bradley builds the first AC 3 phase generator. Up until this time Siemens and Westinghouse had been producing single phase AC generators. The 3 ...<|separator|>
  19. [19]
    What is an Induction Generator or Asynchronous Generator?
    The generators are classified into AC and DC generators. Where the AC generator is further classified into alternator or synchronous generator and asynchronous ...
  20. [20]
    Part 3 – DC Generators – PEG-3722 Electrotechnology
    There are three main types of DC generators named according to the field and armature winding connections: Series-wound; Shunt-wound; Compound-wound. Series- ...
  21. [21]
    Synchronous Generator - an overview | ScienceDirect Topics
    In general, synchronous generators can be either nonpermanent (regular) or permanent magnetic types and both generate reactive power. The permanent magnetic ...
  22. [22]
    Construction and Working Principle of Synchronous Generator
    For the alternator, there are two types of rotor constructions are used viz. the salient-pole type and the cylindrical rotor type. Salient Pole Rotor. The ...
  23. [23]
    Generator Excitation 101 - Emerson Automation Experts
    Jul 25, 2017 · Exciter systems can be rotating or static. Rotating includes brushless and brushed types and static includes compound sources and potential ...
  24. [24]
    [PDF] Brushless and Permanent Magnet Free Wound Field Synchronous ...
    Jun 9, 2015 · Barriers. – Magnet cost (about $200) is about 75% of the 2020 motor cost target; eliminating PMs reduces motor cost by 30%.
  25. [25]
    Types of DC Generators - Tutorials Point
    A long-shunt compound DC generator is one in which shunt field winding is in parallel with both series field winding and armature winding as shown in Figure-5.
  26. [26]
    Types of DC Generators - Working, Diagram & Construction
    Jul 2, 2021 · A compound generator is a combination of a series and a shunt generator. In this type of generator, both the shunt and series field windings are ...<|separator|>
  27. [27]
    Global Electricity Review 2025 - Ember
    Apr 8, 2025 · Renewable power sources added a record 858 TWh of generation in 2024, 49% more than the previous record of 577 TWh set in 2022. The record ...
  28. [28]
    Executive summary – Electricity 2024 – Analysis - IEA
    Global electricity demand is expected to rise at a faster rate over the next three years, growing by an average of 3.4% annually through 2026.
  29. [29]
    Power Generation Applications - mtu Solutions
    They ensure the power supply to important components such as the reactor cooling system - to ensure a controlled shutdown of the reactor.
  30. [30]
    How Generators in Power Stations Work - AZoM
    Jun 18, 2019 · Generators typically use an electromagnet, which is created by electricity and a rapidly spinning turbine to produce massive amounts of current.
  31. [31]
    [PDF] Powering America - Edison Electric Institute
    In total, the electric power industry's economic impact is $880 billion annually—approximately 5 percent of the nation's total GDP. Electric companies provide ...
  32. [32]
    How does energy impact economic growth? An overview of the ...
    Mar 7, 2023 · Electricity shortages were found to reduce the likelihood of an individual being employed in a high skilled job by 35-41% and of being self- ...
  33. [33]
    Carbon Dioxide Emissions From Electricity
    Sep 3, 2024 · Over 40% of energy-related carbon dioxide (CO2) emissions are due to the burning of fossil fuels for electricity generation.
  34. [34]
    Human Health & Environmental Impacts of the Electric Power Sector
    Feb 6, 2025 · Electric power generation is a significant source of toxic metals and other pollutants discharged into water bodies, as well as land pollution ...
  35. [35]
    An Environmental and Societal Analysis of the US Electrical Energy ...
    Electric power generation contributes 0.06 injuries–illnesses/TWh and 0.001 fatalities/TWh, of which fossil fuels contributed 70% and 15%, respectively. Fossil ...
  36. [36]
    GE Successfully Trials Breakthrough High Temperature ...
    Mar 24, 2025 · But new “high temperature superconductors” (HTS) exhibit the phenomenon at much higher temperatures. Such machines will need less complex ...
  37. [37]
    Department of Energy Selects Projects to Develop High-Efficiency ...
    Two projects developed superconducting generators, which make a much stronger magnetic field using superconducting windings. This results in a significant ...
  38. [38]
    Scientists develop rare earth free magnet for use in industrial motors
    Apr 30, 2025 · They developed a rare-earth-free bonded magnet that not only retains its magnetism at high temperatures, coercivity nearly doubles with a 100°C (212°F) ...
  39. [39]
    GreenSpur's Rare Earth-free Generator Technology
    Dec 22, 2023 · The Rare Earth-free generator, designed to generate power without relying on traditional rare earth materials, stands as a resilient solution amid geopolitical ...
  40. [40]
    Asynchronous Generator Market Projections 2025 - LinkedIn
    May 22, 2025 · Asynchronous Generator Market size was valued at USD 2.1 Billion in 2024 and is forecasted to grow at a CAGR of 9.8% from 2026 to 2033, ...
  41. [41]
    Synchronous Generator Market Size, Share, 2025-2030 Outlook
    Sep 15, 2025 · The Synchronous Generator Market is expected to reach USD 5.58 billion in 2025 and grow at a CAGR of 5.51% to reach USD 7.30 billion by 2030 ...Missing: advancements asynchronous
  42. [42]
    Permanent Magnet Synchronous Generator Analysis 2025 and ...
    Rating 4.8 (1,980) May 31, 2025 · Focus on modular designs: Modular generators offer greater flexibility and ease of maintenance.Missing: electromagnetic asynchronous
  43. [43]
    Tripling renewable power and doubling energy efficiency by 2030
    TOTAL GLOBAL RENEWABLE POWER GENERATION CAPACITY WILL NEED TO TRIPLE BY 2030 · ENERGY EFFICIENCY IMPROVEMENTS MUST DOUBLE BY 2030 TO REMAIN ON A 1.5°C PATHWAY.
  44. [44]
    Electric Aircraft Motor Gets Superconducting Upgrade - IEEE Spectrum
    Jun 26, 2025 · With superconductors just in the rotor coils, the motor will achieve efficiencies in the range of 98 to 99.5 percent, which is about four or ...
  45. [45]
    Rare-earth-free permanent-magnet maker emphasizes sustainability ...
    Oct 23, 2024 · Niron Magnetics is launching a manufacturing facility for rare-earth-free permanent magnets in Minnesota.
  46. [46]
    2.4: Generating Sets - Mathematics LibreTexts
    Apr 16, 2022 · Definition: Generating Set ... If G is a group and S is a subset of G such that G = ⟨ S ⟩ , then S is called a generating set of G .
  47. [47]
    What are the generators of (R,+)? - Math Stack Exchange
    Apr 8, 2013 · There is no finite set, nor minimal set of generators for (R,+), hence, it is an infinitely generated (uncountable) group.
  48. [48]
    3.1: Generating Sets - Mathematics LibreTexts
    Mar 5, 2022 · This tells us that for any group we can find a generating set. Usually, we try to find a generating set as small as possible. Sometimes ...
  49. [49]
    [PDF] cyclic-groups.pdf
    Example. (Finding the generators of a cyclic group) List the generators of: (a) Z12. (b) Zp, where p is prime. (a) The generators of Z12 are 1, 5, 7, and 11.
  50. [50]
    4.1: Cyclic Subgroups - Mathematics LibreTexts
    Jun 4, 2022 · Example 4 . 1 . 1 ... hence, Z 6 is a cyclic group. Not every element in a cyclic group is necessarily a generator of the group. The order of 2 ∈ ...
  51. [51]
    [PDF] GENERATING SETS 1. Introduction In Rn, every vector can be ...
    A group has a one-element generating set exactly when it is a cyclic group. For instance, Z has the one-element generating sets {1} and {−1}. Example 1.5 ...
  52. [52]
    Free Group -- from Wolfram MathWorld
    A group is called a free group if no relation exists between its group generators other than the relationship between an element and its inverse.
  53. [53]
    generators of ring - abstract algebra - Math Stack Exchange
    Sep 23, 2014 · As in group theory we have the set of generators as those elements on which if we repeatedly apply binary operation of the group we get all ...
  54. [54]
    7.1: What is a Generating Function? - Mathematics LibreTexts
    Jul 7, 2021 · A generating function is a formal structure that is closely related to a numerical sequence, but allows us to manipulate the sequence as a single entity.
  55. [55]
    [PDF] Generating Functions
    Mar 1, 2015 · We are going to discuss enumeration problems, and how to solve them using a powerful tool: generating functions.
  56. [56]
    3. Generating Functions
    Generating functions are a central concept in algorithm analysis and combinatorics, linking algorithms and analytic methods. They represent sequences of ...
  57. [57]
    [PDF] Notes on exponential generating functions and structures
    Thus we have found a combinatorial interpretation of the coefficients of the power series for tanx—it is the exponential generating function for alternating ...
  58. [58]
    Generating Functions - Discrete Mathematics
    A generating function is a tool in discrete math that uses a power series to encode a sequence, displaying its terms as coefficients.
  59. [59]
    8.3: Using Generating Functions to Solve Recursively-Defined ...
    Jul 7, 2021 · 1) Rearrange the recurrence relation into the form · 2) Define the generating function · 3) Find a linear combination of the generating function ...
  60. [60]
    [PDF] Notes on Ordinary Generating Functions - Berkeley Math
    Generating functions solve enumeration problems, counting objects of a given size or weight, like length of a word or size of a subset.
  61. [61]
    [PDF] Lecture 6 Generating Sets and Bases - UTK Math
    Definition. 1. The set of vectors v1,v2, ..., vn is linearly independent if the set is not linearly dependent( ...
  62. [62]
    Number Theory - Generators
    A unit g ∈ Z n ∗ is called a generator or primitive root of Z n ∗ if for every a ∈ Z n ∗ we have g k = a for some integer k . In other words, if we start ...
  63. [63]
    [PDF] The Primitive Root Theorem - witno.COM
    The primitive root theorem identifies all the positive integers n modulo which primi- tive roots exist. We give detailed proof of this theorem using elementary ...
  64. [64]
    Primitive Root - Algorithms for Competitive Programming
    Jun 6, 2022 · In modular arithmetic, a number is called a primitive root modulo n if every number coprime to is congruent to a power of modulo.
  65. [65]
    separator in nLab
    Aug 11, 2025 · Definitions​​ Definition 3.1. An object S ∈ 𝒞 of a category 𝒞 is called a separator or a generator or a separating object or a generating object, ...
  66. [66]
    Generators | Annoying Precision - WordPress.com
    May 17, 2015 · We proved a theorem due to Gabriel characterizing categories of modules as cocomplete abelian categories with a compact projective generator.
  67. [67]
    Glossary — Python 3.14.0 documentation
    An object created by a generator function. Each yield temporarily suspends processing, remembering the execution state (including local variables and pending ...PEP 492 · Buffer Protocol · Method Resolution Order (MRO)
  68. [68]
    6. Expressions — Python 3.14.0 documentation
    A generator expression yields a new generator object. Its syntax is the same as for comprehensions, except that it is enclosed in parentheses instead of ...
  69. [69]
    What is the difference between an Iterator and a Generator?
    Jun 20, 2009 · An iterator traverses a collection one at a time. A generator generates a sequence, one item at a time. You might for example, iterate over the result of a ...python - Iterators vs Generators vs Iterables - Stack OverflowUnderstanding generators in Python - Stack OverflowMore results from stackoverflow.com
  70. [70]
    PEP 255 – Simple Generators - Python Enhancement Proposals
    Jan 31, 2025 · This PEP introduces the concept of generators to Python, as well as a new statement used in conjunction with them, the yield statement.
  71. [71]
    What's New in Python 2.2 — Python 3.14.0 documentation
    The big difference between yield and a return statement is that on reaching a yield the generator's state of execution is suspended and local variables are ...
  72. [72]
    7. Simple statements — Python 3.14.0 documentation
    Yield expressions and statements are only used when defining a generator function, and are only used in the body of the generator function. Using yield in a ...
  73. [73]
    How to Use Generators and yield in Python
    Generators are lazy iterators that don't store data in memory. The `yield` statement controls flow, suspending execution and saving state.
  74. [74]
    Python Iterators and Generators Tutorial - DataCamp
    Nov 4, 2022 · Iterators are objects that can be iterated upon, while generators are special functions using `yield` to return iterators, and are a simpler ...
  75. [75]
    Iterators and generators - JavaScript - MDN Web Docs
    Jul 29, 2025 · Iterators are objects defining sequences, while generators are created by function* syntax, returning a special iterator. Both customize for...of ...
  76. [76]
    3. Generators and Iterators | Advanced - Python-course.eu
    Mar 20, 2024 · Iterators return data one element at a time. Generators are functions that use `yield` to produce iterators, returning a sequence of results.
  77. [77]
  78. [78]
    Simple Code Generator - GeeksforGeeks
    Jul 26, 2025 · A code generator is a compiler that translates the intermediate representation of the source program into the target program. In other words, a ...
  79. [79]
    What is a code generator? - Codebots
    Jul 28, 2022 · Code generators are ubiquitous in software engineering and when they are understood, they can lead to all sorts of benefits.
  80. [80]
    Approaches for Test Data Generation in Software Testing
    Jul 23, 2025 · The Test Data Generation is the process of collecting and managing a large amount of data from various resources just to implement the test ...
  81. [81]
    How to generate test data: A complete guide - Blog - Synthesized.io
    Mar 30, 2023 · In this post, we will cover the main points of data generation from scratch and based on existing values.
  82. [82]
    Mockaroo - Random Data Generator and API Mocking Tool | JSON ...
    A free test data generator and API mocking tool - Mockaroo lets you create custom CSV, JSON, SQL, and Excel datasets to test and demo your software.
  83. [83]
    Demo Data as Code - Communications of the ACM
    Oct 1, 2019 · The project described here happened many years ago when the available tools were Perl, awk, and sed. Modern tools make this much easier. Python ...
  84. [84]
    [PDF] Automated software test data generation - Purdue Computer Science
    In this paper we present an alternative approach of test data generation which is based on actual execution of the program under test, function minimization ...
  85. [85]
    What is Test Data Generation? Advantages Vs Challenges - Qodo
    What is Test Data Generation? Generating test data involves the creation of a dataset that new or updated software applications will utilize for testing.
  86. [86]
    True Random vs. Pseudorandom Number Generation - wolfSSL
    Jul 13, 2021 · Pseudorandom numbers are software-generated using algorithms, while true random numbers use unpredictable physical variables like radioactive ...
  87. [87]
    Difference between Pseudorandom Number Generators and True ...
    Jun 13, 2024 · PRNGs use deterministic algorithms with a seed, while TRNGs use unpredictable physical phenomena for true randomness. PRNGs are predictable ...
  88. [88]
    The Mechanics of True Random Number Generation - EntropiQ
    Jul 1, 2025 · There are two primary methods for generating these numbers: Pseudo-Random Number Generators (PRNGs) and True Random Number Generators (TRNGs). – ...
  89. [89]
    How Do Computers Generate Random Numbers? - freeCodeCamp
    Oct 26, 2020 · Let's consider two principal methods used to generate random numbers. The first method is based on a physical process, and harvests the source of randomness.
  90. [90]
    Pseudo-Random Number Generators: From the Origins to Modern ...
    Nov 17, 2024 · The Mersenne Twister, introduced by Makoto Matsumoto and Takuji Nishimura in 1997, became a popular PRNG due to its long period (2^19937 - 1) ...Modern Prng Algorithms · Xorshift And... · Comparison Of Prngs
  91. [91]
    Mersenne Twister - an overview | ScienceDirect Topics
    The Mersenne Twister (MT19937) is a pseudorandom number generator used in simulations, known for its long period and good statistical properties.
  92. [92]
    [PDF] A Statistical Test Suite for Random and Pseudorandom Number ...
    The NIST Statistical Test Suite supplies the user with nine pseudo-random number generators. A brief description of each pseudo-random number generator follows.
  93. [93]
    Robert G. Brown's General Tools Page - Duke Physics
    Dieharder is a random number generator (rng) testing suite. It is intended to test generators, not files of possibly random numbers as the latter is a ...
  94. [94]
    A Brief History of Random Numbers - Carl Tashian
    Mar 10, 2017 · In 1951, a random number generator was first added to a general-purpose computer, the Ferranti Mark 1. The Mark ...
  95. [95]
  96. [96]
    What is a Signal Generator? | テクトロニクス - Tektronix
    Introduction to Signal Generators. The signal generator is exactly what its name implies: a generator of signals used as a stimulus for electronic measurements.
  97. [97]
    8 Types of Signal Generators: An Essential Guide for Engineers
    At its core, a signal generator serves as the maestro of electronic testing, directing a symphony of signals that breathe life into circuits and systems for ...
  98. [98]
    Signal Generators — Part 1 - | Nuts & Volts Magazine
    Signal generators and signal sources are instruments that generate controlled signals for use in testing and measurement.
  99. [99]
    Function Generator Circuit Concepts, Part 1 - EDN Network
    Nov 9, 2015 · Function generators (FGs) are measurement instruments in the category of waveform sources. They grew out of both a need for a versatile source ...
  100. [100]
    Types of signal generators | Tektronix
    Apr 28, 2021 · Types of signal generators · Waveform or function generators · Arbitrary function generator · Arbitrary waveform generator · RF signal generator · RF ...
  101. [101]
    Types Of Signal Generators & Applications, Block Diagram
    Feb 12, 2022 · A signal generator is a device that generates signals in a laboratory environment. It can troubleshoot or test electronic devices to create a repeating or non- ...
  102. [102]
    Different Types of Signal Generators and Their Applications
    A quick look at some of the most common signal generators used to produce waveforms, frequencies, radio frequencies (RF), digital patterns, etc.
  103. [103]
    Electrical Waveforms and Signals - Electronics Tutorials
    A Function Generator or sometimes called a Waveform Generator is a device or circuit that produces a variety of different waveforms at a desired frequency. It ...
  104. [104]
    Signal Generators: Working Principles, Types, and Applications
    Jun 3, 2025 · Signal generators are designed to produce electrical signals with highly precise characteristics, such as amplitude, frequency, phase, and waveform.Signal Generators: Working... · Function Generator · Vector Signal Generator<|separator|>
  105. [105]
    What is a Signal Generator? | Tektronix
    Arbitrary waveform generators (AWG) and function generators are aimed primarily at analog and mixed-signal applications. These instruments use sampling ...
  106. [106]
    Digital audio synthesis part 1: The oscillator - Norwegian Creations
    Mar 18, 2022 · The oscillator is a signal generator that creates a waveform with different properties (such as shape, frequency and amplitude).
  107. [107]
  108. [108]
    A Simple Guide to Modulation: Envelope Generators - InSync
    Feb 11, 2022 · Envelope generators provide a means for shaping the waveforms to produce a more musical character. Most commonly, envelope generators are used ...
  109. [109]
    ADSR explained: How to control synth envelopes in your music
    Jun 28, 2023 · An ADSR envelope is a tool for controlling how a sound changes over time. Envelopes help give shape to a sound by changing certain parameters of the sound from ...
  110. [110]
    Synthesis Chapter Four: Envelopes - Introduction to Computer Music
    An envelope generator is a module that outputs a single stream of voltage when it is gated or triggered, normally by a keyboard. It outputs a unipolar DC ...
  111. [111]
    Envelope Generators - BasicSynth
    Synthesis systems typically include envelope generators that produce curved segments in place of straight lines. Curved segments often have a more natural ...<|separator|>
  112. [112]
    [1406.2661] Generative Adversarial Networks - arXiv
    Jun 10, 2014 · We propose a new framework for estimating generative models via an adversarial process, in which we simultaneously train two models.
  113. [113]
    What is Generative AI? | IBM
    Generative AI model architectures and how they have evolved. Variational autoencoders (VAEs), which drove breakthroughs in image recognition, natural language ...What is generative AI? · How generative AI works
  114. [114]
    [PDF] Generative Adversarial Nets - NIPS papers
    We propose a new framework for estimating generative models via an adversar- ial process, in which we simultaneously train two models: a generative model G.
  115. [115]
    [1312.6114] Auto-Encoding Variational Bayes - arXiv
    Dec 20, 2013 · Auto-Encoding Variational Bayes introduces a stochastic variational inference algorithm for efficient learning in probabilistic models with ...
  116. [116]
    [PDF] Denoising Diffusion Probabilistic Models
    We show that diffusion models actually are capable of generating high quality samples, sometimes better than the published results on other types of generative ...
  117. [117]
    Complete Guide to Five Generative AI Models - Coveo
    Apr 4, 2025 · In this guide, we break down five foundational types of generative models—GANs, VAEs, autoregressive models, flow-based models, and transformers ...
  118. [118]
    The state of AI in 2023: Generative AI's breakout year | McKinsey
    Aug 1, 2023 · The latest annual McKinsey Global Survey on the current state of AI confirms the explosive growth of generative AI (gen AI) tools.
  119. [119]
    60+ Generative AI Statistics You Need to Know in 2025 - AmplifAI
    Jun 2, 2025 · In just one year (2023-2024), Gen AI adoption doubled to 65%. Companies that moved early saw clear returns with each dollar invested in Gen AI ...<|separator|>
  120. [120]
    Top 12 Generative AI Models to Explore in 2025 - Analytics Vidhya
    Dec 9, 2024 · Applications: Software development, code generation, code review, debugging support, and enhancing coding productivity. Innovations: PanGu- ...
  121. [121]
    Emu Video and Emu Edit: Our latest generative AI research milestones
    Nov 16, 2023 · We're announcing new research into controlled image editing based solely on text instructions and a method for text-to-video generation based on diffusion ...Missing: synthesis | Show results with:synthesis
  122. [122]
    Generative AI in Computer Vision | ImageVision.ai
    Jul 9, 2024 · Generative AI is significantly advancing Computer Vision, providing robust capabilities for image creation, enhancement, and analysis.
  123. [123]
    How Artificial Intelligence is Revolutionizing Drug Discovery
    Mar 20, 2023 · Milestones in AI-Enabled Drug Discovery · In early 2020, Exscientia announced the first-ever AI-designed drug molecule to enter human clinical ...
  124. [124]
    Unleashing the power of generative AI in drug discovery
    Artificial intelligence (AI) is revolutionizing drug discovery by enhancing precision, reducing timelines and costs, and enabling AI-driven computer-aided drug ...
  125. [125]
    The Economics of Generative AI | NBER
    Apr 24, 2024 · On average, worker productivity increased by 14 percent. These gains were concentrated among the lowest quintile of workers, whether measured by ...
  126. [126]
    Top Generative AI Use Cases and Applications in 2025 - AlmaBetter
    Aug 25, 2025 · Fraud Detection: Generative AI models can analyze transaction data and identify anomalies and patterns associated with fraudulent activities.
  127. [127]
    The 2025 AI Index Report | Stanford HAI
    Generative AI saw particularly strong momentum, attracting $33.9 billion globally in private investment—an 18.7% increase from 2023.
  128. [128]
    Top 122 Generative AI Applications & Real-Life Examples
    Beyond generation, generative AI can help with editing, storyboarding, and animation. These tools automate camera motion, lip-syncing, and scene transitions.
  129. [129]
    [PDF] Free? Assessing the Reliability of Leading AI Legal Research Tools
    AI legal research tools from LexisNexis and Thomson Reuters hallucinate 17-33% of the time, despite claims of "hallucination-free" citations.
  130. [130]
    AI hallucination: towards a comprehensive classification of distorted ...
    Sep 27, 2024 · This study aims to systematically categorize distorted information within AIGC, delve into its internal characteristics, and provide theoretical guidance for ...
  131. [131]
    A Conceptual Framework for Studying AI Hallucinations in (Science ...
    Apr 18, 2025 · This paper proposes a conceptual framework for understanding AI hallucinations as a distinct form of misinformation.
  132. [132]
    Bias of AI-generated content: an examination of news produced by ...
    Mar 4, 2024 · Our study reveals that the AIGC produced by each examined LLM demonstrates substantial gender and racial biases.
  133. [133]
    Stereotypical bias amplification and reversal in an experimental ...
    Apr 9, 2025 · Our findings strongly imply that without robust safeguards, generative AI will amplify existing bias. But with safeguards, existing bias can be eliminated and ...
  134. [134]
    [PDF] Bias in Legal Data for Generative AI
    Aug 13, 2024 · Legal data bias includes under-representation, non-representative data, selection bias towards marginal cases, historical bias, and gender bias.
  135. [135]
    Every AI Copyright Lawsuit in the US, Visualized | WIRED
    Dec 19, 2024 · Over the past two years, dozens of other copyright lawsuits against AI companies have been filed at a rapid clip. The plaintiffs include ...
  136. [136]
    Federal Court Sides with Plaintiff in the First Major AI Copyright ...
    Feb 17, 2025 · On February 11, 2025, a Delaware federal court issued the first major decision concerning the use of copyrighted material to train AI.
  137. [137]
    A Tale of Three Cases: How Fair Use Is Playing Out in AI Copyright ...
    Jul 7, 2025 · In this article, we review the court's reasoning in each of these three cases to elucidate how the battle for a fair use finding is playing out for AI ...
  138. [138]
    The Dark Side Of AI: How Deepfakes And Disinformation ... - Forbes
    Nov 6, 2024 · As AI-generated content becomes increasingly sophisticated, businesses face unprecedented threats from deepfakes and disinformation.
  139. [139]
    Risks and benefits of artificial intelligence deepfakes: Systematic ...
    Merging AI incidents research with political misinformation research: Introducing the political deepfakes incidents database. Proceedings of the AAAI ...
  140. [140]
    [PDF] Increasing Threat of DeepFake Identities - Homeland Security
    This may no longer always be the case and may allow any adversary interested in sowing misinformation or disinformation to leverage far more realistic image, ...
  141. [141]
    Is generative AI a job killer? Evidence from the freelance market
    Jul 8, 2025 · We found that freelancers in occupations more exposed to generative AI have experienced a 2% decline in the number of contracts and a 5% drop in earnings.How might AI affect... · AI in online labor markets · The role of worker experience
  142. [142]
    Is AI Contributing to Rising Unemployment? | St. Louis Fed
    Aug 26, 2025 · Occupations that embraced generative AI most intensively showed the largest unemployment gains, with a correlation coefficient of 0.57. Again, ...
  143. [143]
    Why AI uses so much energy—and what we can do about it
    Apr 8, 2025 · In 2023, data centers consumed 4.4% of U.S. electricity—a number that could triple by 2028. AI's rapid expansion also drives higher water usage, ...
  144. [144]
    A social-environmental impact perspective of generative artificial ...
    Another environmental impact of GenAI pertains to e-waste generation and the fact that some of the materials used in GPUs or data centers are non-recyclable, ...