Speech recognition is an

interdisciplinary Interdisciplinarity or interdisciplinary studies involves the combination of multiple academic disciplines into one activity (e.g., a research project). It draws knowledge from several fields such as sociology, anthropology, psychology, economi ...

subfield of

computer science Computer science is the study of computation, information, and automation. Computer science spans Theoretical computer science, theoretical disciplines (such as algorithms, theory of computation, and information theory) to Applied science, ...

and

computational linguistics Computational linguistics is an interdisciplinary field concerned with the computational modelling of natural language, as well as the study of appropriate computational approaches to linguistic questions. In general, computational linguistics ...

that develops methodologies and technologies that enable the recognition and

translation Translation is the communication of the semantics, meaning of a #Source and target languages, source-language text by means of an Dynamic and formal equivalence, equivalent #Source and target languages, target-language text. The English la ...

of spoken language into text by computers. It is also known as automatic speech recognition (ASR), computer speech recognition or speech-to-text (STT). It incorporates knowledge and research in the

linguistics Linguistics is the scientific study of language. The areas of linguistic analysis are syntax (rules governing the structure of sentences), semantics (meaning), Morphology (linguistics), morphology (structure of words), phonetics (speech sounds ...

and

computer engineering Computer engineering (CE, CoE, or CpE) is a branch of engineering specialized in developing computer hardware and software. It integrates several fields of electrical engineering, electronics engineering and computer science. Computer engi ...

fields. The reverse process is

speech synthesis Speech synthesis is the artificial production of human speech. A computer system used for this purpose is called a speech synthesizer, and can be implemented in software or hardware products. A text-to-speech (TTS) system converts normal langua ...

. Some speech recognition systems require "training" (also called "enrollment") where an individual speaker reads text or isolated

vocabulary A vocabulary (also known as a lexicon) is a set of words, typically the set in a language or the set known to an individual. The word ''vocabulary'' originated from the Latin , meaning "a word, name". It forms an essential component of languag ...

into the system. The system analyzes the person's specific voice and uses it to fine-tune the recognition of that person's speech, resulting in increased accuracy. Systems that do not use training are called "speaker-independent" systems. Systems that use training are called "speaker dependent". Speech recognition applications include

voice user interface A voice-user interface (VUI) enables spoken human interaction with computers, using speech recognition to understand spoken commands and answer questions, and typically text to speech to play a reply. A voice command device is a device controlle ...

s such as voice dialing (e.g. "call home"), call routing (e.g. "I would like to make a collect call"), domotic appliance control, search key words (e.g. find a podcast where particular words were spoken), simple data entry (e.g., entering a credit card number), preparation of structured documents (e.g. a radiology report), determining speaker characteristics, speech-to-text processing (e.g.,

word processor A word processor (WP) is a device or computer program that provides for input, editing, formatting, and output of text, often with some additional features. Early word processors were stand-alone devices dedicated to the function, but current word ...

s or

email Electronic mail (usually shortened to email; alternatively hyphenated e-mail) is a method of transmitting and receiving Digital media, digital messages using electronics, electronic devices over a computer network. It was conceived in the ...

s), and

aircraft An aircraft ( aircraft) is a vehicle that is able to flight, fly by gaining support from the Atmosphere of Earth, air. It counters the force of gravity by using either Buoyancy, static lift or the Lift (force), dynamic lift of an airfoil, or, i ...

(usually termed direct voice input). Automatic pronunciation assessment is used in education such as for spoken language learning. The term ''voice recognition'' or '' speaker identification'' refers to identifying the speaker, rather than what they are saying. Recognizing the speaker can simplify the task of translating speech in systems that have been trained on a specific person's voice or it can be used to

authenticate Authentication (from ''authentikos'', "real, genuine", from αὐθέντης ''authentes'', "author") is the act of proving an assertion, such as the identity of a computer system user. In contrast with identification, the act of indicating ...

or verify the identity of a speaker as part of a security process. From the technology perspective, speech recognition has a long history with several waves of major innovations. Most recently, the field has benefited from advances in

deep learning Deep learning is a subset of machine learning that focuses on utilizing multilayered neural networks to perform tasks such as classification, regression, and representation learning. The field takes inspiration from biological neuroscience a ...

and

big data Big data primarily refers to data sets that are too large or complex to be dealt with by traditional data processing, data-processing application software, software. Data with many entries (rows) offer greater statistical power, while data with ...

. The advances are evidenced not only by the surge of academic papers published in the field, but more importantly by the worldwide industry adoption of a variety of deep learning methods in designing and deploying speech recognition systems.

History

The key areas of growth were: vocabulary size, speaker independence, and processing speed.

Pre-1970

* 1952 – Three Bell Labs researchers, Stephen Balashek, R. Biddulph, and K. H. Davis built a system called "Audrey" for single-speaker digit recognition. Their system located the formants in the power spectrum of each utterance. * 1960 – Gunnar Fant developed and published the source-filter model of speech production. * 1962 –

IBM International Business Machines Corporation (using the trademark IBM), nicknamed Big Blue, is an American Multinational corporation, multinational technology company headquartered in Armonk, New York, and present in over 175 countries. It is ...

demonstrated its 16-word "Shoebox" machine's speech recognition capability at the

1962 World's Fair The Century 21 Exposition (also known as the Seattle World's Fair) was a world's fair held April 21, 1962, to October 21, 1962, in Seattle, Washington, United States. * 1966 –

Linear predictive coding Linear predictive coding (LPC) is a method used mostly in audio signal processing and speech processing for representing the spectral envelope of a digital signal of speech in compressed form, using the information of a linear predictive model ...

(LPC), a

speech coding Speech coding is an application of data compression to digital audio signals containing speech. Speech coding uses speech-specific parameter estimation using audio signal processing techniques to model the speech signal, combined with generic da ...

method, was first proposed by

Fumitada Itakura is a Japanese scientist. He did pioneering work in statistical signal processing, and its application to speech analysis, synthesis and coding, including the development of the linear predictive coding (LPC) and line spectral pairs (LSP) metho ...

Nagoya University , abbreviated to or NU, is a Japanese national research university located in Chikusa-ku, Nagoya. It was established in 1939 as the last of the nine Imperial Universities in the then Empire of Japan, and is now a Designated National Universit ...

and Shuzo Saito of

Nippon Telegraph and Telephone (NTT) is a Japanese telecommunications holding company headquartered in Tokyo, Japan. Ranked 55th in ''Fortune'' Global 500, NTT is the fourth largest telecommunications company in the world in terms of revenue, as well as the third largest pu ...

(NTT), while working on speech recognition. * 1969 – Funding at

Bell Labs Nokia Bell Labs, commonly referred to as ''Bell Labs'', is an American industrial research and development company owned by Finnish technology company Nokia. With headquarters located in Murray Hill, New Jersey, Murray Hill, New Jersey, the compa ...

dried up for several years when, in 1969, the influential John Pierce wrote an open letter that was critical of and defunded speech recognition research. This defunding lasted until Pierce retired and James L. Flanagan took over. Raj Reddy was the first person to take on continuous speech recognition as a graduate student at

Stanford University Leland Stanford Junior University, commonly referred to as Stanford University, is a Private university, private research university in Stanford, California, United States. It was founded in 1885 by railroad magnate Leland Stanford (the eighth ...

in the late 1960s. Previous systems required users to pause after each word. Reddy's system issued spoken commands for playing

chess Chess is a board game for two players. It is an abstract strategy game that involves Perfect information, no hidden information and no elements of game of chance, chance. It is played on a square chessboard, board consisting of 64 squares arran ...

. Around this time Soviet researchers invented the dynamic time warping (DTW) algorithm and used it to create a recognizer capable of operating on a 200-word vocabulary. DTW processed speech by dividing it into short frames, e.g. 10ms segments, and processing each frame as a single unit. Although DTW would be superseded by later algorithms, the technique carried on. Achieving speaker independence remained unsolved at this time period.

1970–1990

* 1971 –

DARPA The Defense Advanced Research Projects Agency (DARPA) is a research and development agency of the United States Department of Defense responsible for the development of emerging technologies for use by the military. Originally known as the Adva ...

funded five years for ''Speech Understanding Research'', speech recognition research seeking a minimum vocabulary size of 1,000 words. They thought speech ''understanding'' would be key to making progress in speech ''recognition'', but this later proved untrue. BBN,

Carnegie Mellon Carnegie may refer to: People *Carnegie (surname), including a list of people with the name **Andrew Carnegie, Scottish-American industrialist and philanthropist * Clan Carnegie, a lowland Scottish clan Institutions Named for Andrew Carnegie * ...

and

Stanford Research Institute SRI International (SRI) is a nonprofit organization, nonprofit scientific research, scientific research institute and organization headquartered in Menlo Park, California, United States. It was established in 1946 by trustees of Stanford Univer ...

all participated in the program. This revived speech recognition research post John Pierce's letter. * 1972 – The IEEE Acoustics, Speech, and Signal Processing group held a conference in Newton, Massachusetts. * 1976 – The first ICASSP was held in

Philadelphia Philadelphia ( ), colloquially referred to as Philly, is the List of municipalities in Pennsylvania, most populous city in the U.S. state of Pennsylvania and the List of United States cities by population, sixth-most populous city in the Unit ...

, which since then has been a major venue for the publication of research on speech recognition. During the late 1960s Leonard Baum developed the mathematics of

Markov chain In probability theory and statistics, a Markov chain or Markov process is a stochastic process describing a sequence of possible events in which the probability of each event depends only on the state attained in the previous event. Informally ...

s at the Institute for Defense Analysis. A decade later, at CMU, Raj Reddy's students

James Baker James Addison Baker III (born April 28, 1930) is an American attorney, diplomat and statesman. A member of the Republican Party (United States), Republican Party, he served as the 10th White House chief of staff and 67th United States secretary ...

and Janet M. Baker began using the

hidden Markov model A hidden Markov model (HMM) is a Markov model in which the observations are dependent on a latent (or ''hidden'') Markov process (referred to as X). An HMM requires that there be an observable process Y whose outcomes depend on the outcomes of X ...

(HMM) for speech recognition. James Baker had learned about HMMs from a summer job at the Institute of Defense Analysis during his undergraduate education. The use of HMMs allowed researchers to combine different sources of knowledge, such as acoustics, language, and syntax, in a unified probabilistic model. * By the mid-1980s IBM's Fred Jelinek's team created a voice activated typewriter called Tangora, which could handle a 20,000-word vocabulary Jelinek's statistical approach put less emphasis on emulating the way the human brain processes and understands speech in favor of using statistical modelling techniques like HMMs. (Jelinek's group independently discovered the application of HMMs to speech.) This was controversial with linguists since HMMs are too simplistic to account for many common features of human languages. However, the HMM proved to be a highly useful way for modelling speech and replaced dynamic time warping to become the dominant speech recognition algorithm in the 1980s. * 1982 – Dragon Systems, founded by James and Janet M. Baker, was one of IBM's few competitors.

Practical speech recognition

The 1980s also saw the introduction of the n-gram language model. * 1987 – The back-off model allowed language models to use multiple length n-grams, and CSELT used HMM to recognize languages (both in software and in hardware specialized processors, e.g. RIPAC). Much of the progress in the field is owed to the rapidly increasing capabilities of computers. At the end of the DARPA program in 1976, the best computer available to researchers was the

PDP-10 Digital Equipment Corporation (DEC)'s PDP-10, later marketed as the DECsystem-10, is a mainframe computer family manufactured beginning in 1966 and discontinued in 1983. 1970s models and beyond were marketed under the DECsystem-10 name, especi ...

with 4 MB ram. It could take up to 100 minutes to decode just 30 seconds of speech. Two practical products were: * 1984 – was released the

Apricot Portable The Apricot Portable was a personal computer manufactured by ACT Ltd., and was released to the public in November 1984. It was ACT's first attempt at manufacturing a portable computer, which were gaining popularity at the time. Compared to other ...

with up to 4096 words support, of which only 64 could be held in

RAM Ram, ram, or RAM most commonly refers to: * A male sheep * Random-access memory, computer memory * Ram Trucks, US, since 2009 ** List of vehicles named Dodge Ram, trucks and vans ** Ram Pickup, produced by Ram Trucks Ram, ram, or RAM may also ref ...

at a time. *1987 – a recognizer from Kurzweil Applied Intelligence * 1990 – Dragon Dictate, a consumer product released in 1990

AT&T AT&T Inc., an abbreviation for its predecessor's former name, the American Telephone and Telegraph Company, is an American multinational telecommunications holding company headquartered at Whitacre Tower in Downtown Dallas, Texas. It is the w ...

deployed the Voice Recognition Call Processing service in 1992 to route telephone calls without the use of a human operator. The technology was developed by Lawrence Rabiner and others at Bell Labs. By this point, the vocabulary of the typical commercial speech recognition system was larger than the average human vocabulary. Raj Reddy's former student, Xuedong Huang, developed the Sphinx-II system at CMU. The Sphinx-II system was the first to do speaker-independent, large vocabulary, continuous speech recognition and it had the best performance in DARPA's 1992 evaluation. Handling continuous speech with a large vocabulary was a major milestone in the history of speech recognition. Huang went on to found the speech recognition group at Microsoft in 1993. Raj Reddy's student Kai-Fu Lee joined Apple where, in 1992, he helped develop a speech interface prototype for the Apple computer known as Casper. Lernout & Hauspie, a Belgium-based speech recognition company, acquired several other companies, including Kurzweil Applied Intelligence in 1997 and Dragon Systems in 2000. The L&H speech technology was used in the

Windows XP Windows XP is a major release of Microsoft's Windows NT operating system. It was released to manufacturing on August 24, 2001, and later to retail on October 25, 2001. It is a direct successor to Windows 2000 for high-end and business users a ...

operating system. L&H was an industry leader until an accounting scandal brought an end to the company in 2001. The speech technology from L&H was bought by ScanSoft which became Nuance in 2005.

Apple An apple is a round, edible fruit produced by an apple tree (''Malus'' spp.). Fruit trees of the orchard or domestic apple (''Malus domestica''), the most widely grown in the genus, are agriculture, cultivated worldwide. The tree originated ...

originally licensed software from Nuance to provide speech recognition capability to its digital assistant

Siri Siri ( , backronym: Speech Interpretation and Recognition Interface) is a digital assistant purchased, developed, and popularized by Apple Inc., which is included in the iOS, iPadOS, watchOS, macOS, Apple TV, audioOS, and visionOS operating sys ...

2000s

In the 2000s DARPA sponsored two speech recognition programs: Effective Affordable Reusable Speech-to-Text (EARS) in 2002 and Global Autonomous Language Exploitation (GALE). Four teams participated in the EARS program:

, a team led by BBN with LIMSI and Univ. of Pittsburgh,

Cambridge University The University of Cambridge is a Public university, public collegiate university, collegiate research university in Cambridge, England. Founded in 1209, the University of Cambridge is the List of oldest universities in continuous operation, wo ...

, and a team composed of ICSI, SRI and

University of Washington The University of Washington (UW and informally U-Dub or U Dub) is a public research university in Seattle, Washington, United States. Founded in 1861, the University of Washington is one of the oldest universities on the West Coast of the Uni ...

. EARS funded the collection of the Switchboard telephone

speech corpus A speech corpus (or spoken corpus) is a database of speech audio files and text Transcription (linguistics), transcriptions. In speech technology, speech corpora are used, among other things, to create acoustic models (which can then be used with ...

containing 260 hours of recorded conversations from over 500 speakers. The GALE program focused on

Arabic Arabic (, , or , ) is a Central Semitic languages, Central Semitic language of the Afroasiatic languages, Afroasiatic language family spoken primarily in the Arab world. The International Organization for Standardization (ISO) assigns lang ...

and

Mandarin Mandarin or The Mandarin may refer to: Language * Mandarin Chinese, branch of Chinese originally spoken in northern parts of the country ** Standard Chinese or Modern Standard Mandarin, the official language of China ** Taiwanese Mandarin, Stand ...

broadcast news speech.

Google Google LLC (, ) is an American multinational corporation and technology company focusing on online advertising, search engine technology, cloud computing, computer software, quantum computing, e-commerce, consumer electronics, and artificial ...

's first effort at speech recognition came in 2007 after hiring some researchers from Nuance. The first product was GOOG-411, a telephone based directory service. The recordings from GOOG-411 produced valuable data that helped Google improve their recognition systems. Google Voice Search is now supported in over 30 languages. In the United States, the

National Security Agency The National Security Agency (NSA) is an intelligence agency of the United States Department of Defense, under the authority of the director of national intelligence (DNI). The NSA is responsible for global monitoring, collection, and proces ...

has made use of a type of speech recognition for keyword spotting since at least 2006. This technology allows analysts to search through large volumes of recorded conversations and isolate mentions of keywords. Recordings can be indexed and analysts can run queries over the database to find conversations of interest. Some government research programs focused on intelligence applications of speech recognition, e.g. DARPA's EARS's program and IARPA's Babel program. In the early 2000s, speech recognition was still dominated by traditional approaches such as

s combined with feedforward

artificial neural networks In machine learning, a neural network (also artificial neural network or neural net, abbreviated ANN or NN) is a computational model inspired by the structure and functions of biological neural networks. A neural network consists of connected ...

.Herve Bourlard and Nelson Morgan, Connectionist Speech Recognition: A Hybrid Approach, The Kluwer International Series in Engineering and Computer Science; v. 247, Boston: Kluwer Academic Publishers, 1994. Today, however, many aspects of speech recognition have been taken over by a

method called

Long short-term memory Long short-term memory (LSTM) is a type of recurrent neural network (RNN) aimed at mitigating the vanishing gradient problem commonly encountered by traditional RNNs. Its relative insensitivity to gap length is its advantage over other RNNs, ...

(LSTM), a

recurrent neural network Recurrent neural networks (RNNs) are a class of artificial neural networks designed for processing sequential data, such as text, speech, and time series, where the order of elements is important. Unlike feedforward neural networks, which proces ...

published by Sepp Hochreiter & Jürgen Schmidhuber in 1997. LSTM RNNs avoid the vanishing gradient problem and can learn "Very Deep Learning" tasks that require memories of events that happened thousands of discrete time steps ago, which is important for speech. Around 2007, LSTM trained by Connectionist Temporal Classification (CTC)Alex Graves, Santiago Fernandez, Faustino Gomez, and Jürgen Schmidhuber (2006)
Connectionist temporal classification: Labelling unsegmented sequence data with recurrent neural nets
. Proceedings of ICML'06, pp. 369–376. started to outperform traditional speech recognition in certain applications.Santiago Fernandez, Alex Graves, and Jürgen Schmidhuber (2007)
An application of recurrent neural networks to discriminative keyword spotting
Proceedings of ICANN (2), pp. 220–229. In 2015, Google's speech recognition reportedly experienced a dramatic performance jump of 49% through CTC-trained LSTM, which is now available through

Google Voice Google Voice is a telephone service that provides a U.S. phone number to Google Account customers in the U.S. and Google Workspace (G Suite by October 2020) customers in Canada, Denmark, France, the Netherlands, Portugal, Spain, Sweden, Switz ...

to all smartphone users.Haşim Sak, Andrew Senior, Kanishka Rao, Françoise Beaufays and Johan Schalkwyk (September 2015): "."

Transformers ''Transformers'' is a media franchise produced by American toy company Hasbro and Japanese toy company Tomy, Takara Tomy. It primarily follows the heroic Autobots and the villainous Decepticons, two Extraterrestrials in fiction, alien robot fac ...

, a type of neural network based solely on "attention", have been widely adopted in computer vision and language modelling, sparking the interest of adapting such models to new domains, including speech recognition. Some recent papers reported superior performance levels using transformer models for speech recognition, but these models usually require large scale training datasets to reach high performance levels. The use of deep feedforward (non-recurrent) networks for acoustic modelling was introduced during the later part of 2009 by

Geoffrey Hinton Geoffrey Everest Hinton (born 1947) is a British-Canadian computer scientist, cognitive scientist, and cognitive psychologist known for his work on artificial neural networks, which earned him the title "the Godfather of AI". Hinton is Univer ...

and his students at the University of Toronto and by Li Deng and colleagues at Microsoft Research, initially in the collaborative work between Microsoft and the University of Toronto which was subsequently expanded to include IBM and Google (hence "The shared views of four research groups" subtitle in their 2012 review paper).NIPS Workshop: Deep Learning for Speech Recognition and Related Applications, Whistler, BC, Canada, Dec. 2009 (Organizers: Li Deng, Geoff Hinton, D. Yu). A Microsoft research executive called this innovation "the most dramatic change in accuracy since 1979". In contrast to the steady incremental improvements of the past few decades, the application of deep learning decreased word error rate by 30%. This innovation was quickly adopted across the field. Researchers have begun to use deep learning techniques for language modelling as well. In the long history of speech recognition, both shallow form and deep form (e.g. recurrent nets) of artificial neural networks had been explored for many years during 1980s, 1990s and a few years into the 2000s.Morgan, Bourlard, Renals, Cohen, Franco (1993) "Hybrid neural network/hidden Markov model systems for continuous speech recognition. ICASSP/IJPRAI" Waibel, Hanazawa, Hinton, Shikano, Lang. (1989)
Phoneme recognition using time-delay neural networks
. IEEE Transactions on Acoustics, Speech, and Signal Processing." But these methods never won over the non-uniform internal-handcrafting

Gaussian mixture model In statistics, a mixture model is a probabilistic model for representing the presence of subpopulations within an overall population, without requiring that an observed data set should identify the sub-population to which an individual observati ...

(GMM-HMM) technology based on generative models of speech trained discriminatively. A number of key difficulties had been methodologically analyzed in the 1990s, including gradient diminishing Sepp Hochreiter (1991)
Untersuchungen zu dynamischen neuronalen Netzen
, Diploma thesis. Institut f. Informatik, Technische Univ. Munich. Advisor: J. Schmidhuber. and weak temporal correlation structure in the neural predictive models. All these difficulties were in addition to the lack of big training data and big computing power in these early days. Most speech recognition researchers who understood such barriers hence subsequently moved away from neural nets to pursue generative modelling approaches until the recent resurgence of deep learning starting around 2009–2010 that had overcome all these difficulties. Hinton et al. and Deng et al. reviewed part of this recent history about how their collaboration with each other and then with colleagues across four groups (University of Toronto, Microsoft, Google, and IBM) ignited a renaissance of applications of deep feedforward neural networks for speech recognition.Keynote talk: Recent Developments in Deep Neural Networks. ICASSP, 2013 (by Geoff Hinton).Keynote talk:

," Interspeech, September 2014 (by Li Deng).

2010s

By early 2010s ''speech'' recognition, also called voice recognition was clearly differentiated from ''speaker'' recognition, and speaker independence was considered a major breakthrough. Until then, systems required a "training" period. A 1987 ad for a doll had carried the tagline "Finally, the doll that understands you." – despite the fact that it was described as "which children could train to respond to their voice". In 2017, Microsoft researchers reached a historical human parity milestone of transcribing conversational telephony speech on the widely benchmarked Switchboard task. Multiple deep learning models were used to optimize speech recognition accuracy. The speech recognition word error rate was reported to be as low as 4 professional human transcribers working together on the same benchmark, which was funded by IBM Watson speech team on the same task.

Models, methods, and algorithms

Both acoustic modelling and language modelling are important parts of modern statistically based speech recognition algorithms. Hidden Markov models (HMMs) are widely used in many systems. Language modelling is also used in many other natural language processing applications such as

document classification Document classification or document categorization is a problem in library science, information science and computer science. The task is to assign a document to one or more Class (philosophy), classes or Categorization, categories. This may be do ...

statistical machine translation Statistical machine translation (SMT) is a machine translation approach where translations are generated on the basis of statistical models whose parameters are derived from the analysis of bilingual text corpora. The statistical approach contra ...

Hidden Markov models

Modern general-purpose speech recognition systems are based on hidden Markov models. These are statistical models that output a sequence of symbols or quantities. HMMs are used in speech recognition because a speech signal can be viewed as a piecewise stationary signal or a short-time stationary signal. In a short time scale (e.g., 10 milliseconds), speech can be approximated as a

stationary process In mathematics and statistics, a stationary process (also called a strict/strictly stationary process or strong/strongly stationary process) is a stochastic process whose statistical properties, such as mean and variance, do not change over time. M ...

. Speech can be thought of as a

Markov model In probability theory, a Markov model is a stochastic model used to Mathematical model, model pseudo-randomly changing systems. It is assumed that future states depend only on the current state, not on the events that occurred before it (that is, ...

for many stochastic purposes. Another reason why HMMs are popular is that they can be trained automatically and are simple and computationally feasible to use. In speech recognition, the hidden Markov model would output a sequence of ''n''-dimensional real-valued vectors (with ''n'' being a small integer, such as 10), outputting one of these every 10 milliseconds. The vectors would consist of cepstral coefficients, which are obtained by taking a

Fourier transform In mathematics, the Fourier transform (FT) is an integral transform that takes a function as input then outputs another function that describes the extent to which various frequencies are present in the original function. The output of the tr ...

of a short time window of speech and decorrelating the spectrum using a cosine transform, then taking the first (most significant) coefficients. The hidden Markov model will tend to have in each state a statistical distribution that is a mixture of diagonal covariance Gaussians, which will give a likelihood for each observed vector. Each word, or (for more general speech recognition systems), each

phoneme A phoneme () is any set of similar Phone (phonetics), speech sounds that are perceptually regarded by the speakers of a language as a single basic sound—a smallest possible Phonetics, phonetic unit—that helps distinguish one word fr ...

, will have a different output distribution; a hidden Markov model for a sequence of words or phonemes is made by concatenating the individual trained hidden Markov models for the separate words and phonemes. Described above are the core elements of the most common, HMM-based approach to speech recognition. Modern speech recognition systems use various combinations of a number of standard techniques in order to improve results over the basic approach described above. A typical large-vocabulary system would need context dependency for the

s (so that phonemes with different left and right context would have different realizations as HMM states); it would use cepstral normalization to normalize for a different speaker and recording conditions; for further speaker normalization, it might use vocal tract length normalization (VTLN) for male-female normalization and maximum likelihood linear regression (MLLR) for more general speaker adaptation. The features would have so-called

delta Delta commonly refers to: * Delta (letter) (Δ or δ), the fourth letter of the Greek alphabet * D (NATO phonetic alphabet: "Delta"), the fourth letter in the Latin alphabet * River delta, at a river mouth * Delta Air Lines, a major US carrier ...

and delta-delta coefficients to capture speech dynamics and in addition, might use heteroscedastic linear discriminant analysis (HLDA); or might skip the delta and delta-delta coefficients and use splicing and an LDA-based projection followed perhaps by heteroscedastic linear discriminant analysis or a global semi-tied co variance transform (also known as

maximum likelihood linear transform In mathematical analysis, the maximum and minimum of a function are, respectively, the greatest and least value taken by the function. Known generically as extremum, they may be defined either within a given range (the ''local'' or ''relative ...

, or MLLT). Many systems use so-called discriminative training techniques that dispense with a purely statistical approach to HMM parameter estimation and instead optimize some classification-related measure of the training data. Examples are maximum

mutual information In probability theory and information theory, the mutual information (MI) of two random variables is a measure of the mutual Statistical dependence, dependence between the two variables. More specifically, it quantifies the "Information conten ...

(MMI), minimum classification error (MCE), and minimum phone error (MPE). Decoding of the speech (the term for what happens when the system is presented with a new utterance and must compute the most likely source sentence) would probably use the Viterbi algorithm to find the best path, and here there is a choice between dynamically creating a combination hidden Markov model, which includes both the acoustic and language model information and combining it statically beforehand (the

finite state transducer A finite-state transducer (FST) is a finite-state machine with two memory ''tapes'', following the terminology for Turing machines: an input tape and an output tape. This contrasts with an ordinary finite-state automaton, which has a single tape. ...

, or FST, approach). A possible improvement to decoding is to keep a set of good candidates instead of just keeping the best candidate, and to use a better scoring function ( re scoring) to rate these good candidates so that we may pick the best one according to this refined score. The set of candidates can be kept either as a list (the N-best list approach) or as a subset of the models (a lattice). Re scoring is usually done by trying to minimize the Bayes risk (or an approximation thereof) Instead of taking the source sentence with maximal probability, we try to take the sentence that minimizes the expectancy of a given loss function with regards to all possible transcriptions (i.e., we take the sentence that minimizes the average distance to other possible sentences weighted by their estimated probability). The loss function is usually the Levenshtein distance, though it can be different distances for specific tasks; the set of possible transcriptions is, of course, pruned to maintain tractability. Efficient algorithms have been devised to re score lattices represented as weighted finite state transducers with

edit distance In computational linguistics and computer science, edit distance is a string metric, i.e. a way of quantifying how dissimilar two String (computing), strings (e.g., words) are to one another, that is measured by counting the minimum number of opera ...

s represented themselves as a

verifying certain assumptions.

Dynamic time warping (DTW)-based speech recognition

Dynamic time warping is an approach that was historically used for speech recognition but has now largely been displaced by the more successful HMM-based approach. Dynamic time warping is an algorithm for measuring similarity between two sequences that may vary in time or speed. For instance, similarities in walking patterns would be detected, even if in one video the person was walking slowly and if in another he or she were walking more quickly, or even if there were accelerations and deceleration during the course of one observation. DTW has been applied to video, audio, and graphics – indeed, any data that can be turned into a linear representation can be analyzed with DTW. A well-known application has been automatic speech recognition, to cope with different speaking speeds. In general, it is a method that allows a computer to find an optimal match between two given sequences (e.g., time series) with certain restrictions. That is, the sequences are "warped" non-linearly to match each other. This sequence alignment method is often used in the context of hidden Markov models.

Neural networks

Neural networks emerged as an attractive acoustic modelling approach in ASR in the late 1980s. Since then, neural networks have been used in many aspects of speech recognition such as phoneme classification, phoneme classification through multi-objective evolutionary algorithms, isolated word recognition, audiovisual speech recognition, audiovisual speaker recognition and speaker adaptation.

Neural networks A neural network is a group of interconnected units called neurons that send signals to one another. Neurons can be either Cell (biology), biological cells or signal pathways. While individual neurons are simple, many of them together in a netwo ...

make fewer explicit assumptions about feature statistical properties than HMMs and have several qualities making them more attractive recognition models for speech recognition. When used to estimate the probabilities of a speech feature segment, neural networks allow discriminative training in a natural and efficient manner. However, in spite of their effectiveness in classifying short-time units such as individual phonemes and isolated words, early neural networks were rarely successful for continuous recognition tasks because of their limited ability to model temporal dependencies. One approach to this limitation was to use neural networks as a pre-processing, feature transformation or dimensionality reduction, step prior to HMM based recognition. However, more recently, LSTM and related recurrent neural networks (RNNs), Time Delay Neural Networks(TDNN's), and transformers have demonstrated improved performance in this area.

Deep feedforward and recurrent neural networks

Deep neural networks and denoising autoencoders are also under investigation. A deep feedforward neural network (DNN) is an

artificial neural network In machine learning, a neural network (also artificial neural network or neural net, abbreviated ANN or NN) is a computational model inspired by the structure and functions of biological neural networks. A neural network consists of connected ...

with multiple hidden layers of units between the input and output layers. Similar to shallow neural networks, DNNs can model complex non-linear relationships. DNN architectures generate compositional models, where extra layers enable composition of features from lower layers, giving a huge learning capacity and thus the potential of modelling complex patterns of speech data. A success of DNNs in large vocabulary speech recognition occurred in 2010 by industrial researchers, in collaboration with academic researchers, where large output layers of the DNN based on context dependent HMM states constructed by decision trees were adopted. Deng L., Li, J., Huang, J., Yao, K., Yu, D., Seide, F. et al
Recent Advances in Deep Learning for Speech Research at Microsoft
. ICASSP, 2013. See comprehensive reviews of this development and of the state of the art as of October 2014 in the recent Springer book from Microsoft Research. See also the related background of automatic speech recognition and the impact of various machine learning paradigms, notably including

, in recent overview articles. One fundamental principle of

is to do away with hand-crafted feature engineering and to use raw features. This principle was first explored successfully in the architecture of deep autoencoder on the "raw" spectrogram or linear filter-bank features,L. Deng, M. Seltzer, D. Yu, A. Acero, A. Mohamed, and G. Hinton (2010
Binary Coding of Speech Spectrograms Using a Deep Auto-encoder
Interspeech. showing its superiority over the Mel-Cepstral features which contain a few stages of fixed transformation from spectrograms. The true "raw" features of speech, waveforms, have more recently been shown to produce excellent larger-scale speech recognition results.

End-to-end automatic speech recognition

Since 2014, there has been much research interest in "end-to-end" ASR. Traditional phonetic-based (i.e., all HMM-based model) approaches required separate components and training for the pronunciation, acoustic, and

language model A language model is a model of the human brain's ability to produce natural language. Language models are useful for a variety of tasks, including speech recognition, machine translation,Andreas, Jacob, Andreas Vlachos, and Stephen Clark (2013)"S ...

. End-to-end models jointly learn all the components of the speech recognizer. This is valuable since it simplifies the training process and deployment process. For example, a n-gram language model is required for all HMM-based systems, and a typical n-gram language model often takes several gigabytes in memory making them impractical to deploy on mobile devices. Consequently, modern commercial ASR systems from

and

() are deployed on the cloud and require a network connection as opposed to the device locally. The first attempt at end-to-end ASR was with Connectionist Temporal Classification (CTC)-based systems introduced by Alex Graves of

Google DeepMind DeepMind Technologies Limited, trading as Google DeepMind or simply DeepMind, is a British–American artificial intelligence research laboratory which serves as a subsidiary of Alphabet Inc. Founded in the UK in 2010, it was acquired by Goo ...

and Navdeep Jaitly of the

University of Toronto The University of Toronto (UToronto or U of T) is a public university, public research university whose main campus is located on the grounds that surround Queen's Park (Toronto), Queen's Park in Toronto, Ontario, Canada. It was founded by ...

in 2014. The model consisted of

s and a CTC layer. Jointly, the RNN-CTC model learns the pronunciation and acoustic model together, however it is incapable of learning the language due to

conditional independence In probability theory, conditional independence describes situations wherein an observation is irrelevant or redundant when evaluating the certainty of a hypothesis. Conditional independence is usually formulated in terms of conditional probabi ...

assumptions similar to a HMM. Consequently, CTC models can directly learn to map speech acoustics to English characters, but the models make many common spelling mistakes and must rely on a separate language model to clean up the transcripts. Later,

Baidu Baidu, Inc. ( ; ) is a Chinese multinational technology company specializing in Internet services and artificial intelligence. It holds a dominant position in China's search engine market (via Baidu Search), and provides a wide variety of o ...

expanded on the work with extremely large datasets and demonstrated some commercial success in Chinese Mandarin and English. In 2016,

University of Oxford The University of Oxford is a collegiate university, collegiate research university in Oxford, England. There is evidence of teaching as early as 1096, making it the oldest university in the English-speaking world and the List of oldest un ...

presented LipNet, the first end-to-end sentence-level lipreading model, using spatiotemporal convolutions coupled with an RNN-CTC architecture, surpassing human-level performance in a restricted grammar dataset. A large-scale CNN-RNN-CTC architecture was presented in 2018 by

achieving 6 times better performance than human experts. In 2019,

Nvidia Nvidia Corporation ( ) is an American multinational corporation and technology company headquartered in Santa Clara, California, and incorporated in Delaware. Founded in 1993 by Jensen Huang (president and CEO), Chris Malachowsky, and Curti ...

launched two CNN-CTC ASR models, Jasper and QuarzNet, with an overall performance WER of 3%. Similar to other deep learning applications,

transfer learning Transfer learning (TL) is a technique in machine learning (ML) in which knowledge learned from a task is re-used in order to boost performance on a related task. For example, for image classification, knowledge gained while learning to recogniz ...

and

domain adaptation Domain adaptation is a field associated with machine learning and inductive transfer, transfer learning. It addresses the challenge of training a model on one data distribution (the source domain) and applying it to a related but different data ...

are important strategies for reusing and extending the capabilities of deep learning models, particularly due to the high costs of training models from scratch, and the small size of available corpus in many languages and/or specific domains. An alternative approach to CTC-based models are attention-based models. Attention-based ASR models were introduced simultaneously by Chan et al. of

Carnegie Mellon University Carnegie Mellon University (CMU) is a private research university in Pittsburgh, Pennsylvania, United States. The institution was established in 1900 by Andrew Carnegie as the Carnegie Technical Schools. In 1912, it became the Carnegie Institu ...

and

Google Brain Google Brain was a deep learning artificial intelligence research team that served as the sole AI branch of Google before being incorporated under the newer umbrella of Google AI, a research division at Google dedicated to artificial intelligence ...

and Bahdanau et al. of the

University of Montreal A university () is an institution of tertiary education and research which awards academic degrees in several academic disciplines. ''University'' is derived from the Latin phrase , which roughly means "community of teachers and scholars". Univ ...

in 2016. The model named "Listen, Attend and Spell" (LAS), literally "listens" to the acoustic signal, pays "attention" to different parts of the signal and "spells" out the transcript one character at a time. Unlike CTC-based models, attention-based models do not have conditional-independence assumptions and can learn all the components of a speech recognizer including the pronunciation, acoustic and language model directly. This means, during deployment, there is no need to carry around a language model making it very practical for applications with limited memory. By the end of 2016, the attention-based models have seen considerable success including outperforming the CTC models (with or without an external language model). Various extensions have been proposed since the original LAS model. Latent Sequence Decompositions (LSD) was proposed by

MIT The Massachusetts Institute of Technology (MIT) is a private research university in Cambridge, Massachusetts, United States. Established in 1861, MIT has played a significant role in the development of many areas of modern technology and sc ...

and

to directly emit sub-word units which are more natural than English characters;

and

extended LAS to "Watch, Listen, Attend and Spell" (WLAS) to handle lip reading surpassing human-level performance.

Applications

In-car systems

Typically a manual control input, for example by means of a finger control on the steering-wheel, enables the speech recognition system and this is signaled to the driver by an audio prompt. Following the audio prompt, the system has a "listening window" during which it may accept a speech input for recognition. Simple voice commands may be used to initiate phone calls, select radio stations or play music from a compatible smartphone, MP3 player or music-loaded flash drive. Voice recognition capabilities vary between car make and model. Some of the most recent car models offer natural-language speech recognition in place of a fixed set of commands, allowing the driver to use full sentences and common phrases. With such systems there is, therefore, no need for the user to memorize a set of fixed command words.

Education

Automatic

pronunciation Pronunciation is the way in which a word or a language is spoken. To This may refer to generally agreed-upon sequences of sounds used in speaking a given word or all language in a specific dialect—"correct" or "standard" pronunciation—or si ...

assessment is the use of speech recognition to verify the correctness of pronounced speech, as distinguished from manual assessment by an instructor or proctor. Also called speech verification, pronunciation evaluation, and pronunciation scoring, the main application of this technology is computer-aided pronunciation teaching (CAPT) when combined with

computer-aided instruction Educational technology (commonly abbreviated as edutech, or edtech) is the combined use of computer hardware, software, and educational theory and practice to facilitate learning and teaching. When referred to with its abbreviation, "EdTech", ...

for

computer-assisted language learning Computer-assisted language learning (CALL), known as computer-aided instruction (CAI) in British English and computer-aided language instruction (CALI) in American English, Levy (1997: p. 1) briefly defines it as "the exploration and study of co ...

(CALL), speech remediation, or accent reduction. Pronunciation assessment does not determine unknown speech (as in dictation or automatic transcription) but instead, knowing the expected word(s) in advance, it attempts to verify the correctness of the learner's pronunciation and ideally their intelligibility to listeners, sometimes along with often inconsequential prosody such as intonation, pitch,

tempo In musical terminology, tempo (Italian for 'time'; plural 'tempos', or from the Italian plural), measured in beats per minute, is the speed or pace of a given musical composition, composition, and is often also an indication of the composition ...

rhythm Rhythm (from Greek , ''rhythmos'', "any regular recurring motion, symmetry") generally means a " movement marked by the regulated succession of strong and weak elements, or of opposite or different conditions". This general meaning of regular r ...

, and stress. Pronunciation assessment is also used in reading tutoring, for example in products such as

Microsoft Teams Microsoft Teams is a team collaboration platform developed by Microsoft as part of the Microsoft 365 suite. It offers features such as workspace chat, video conferencing, file storage, and integration with both Microsoft and third-party applicat ...

and from Amira Learning. Automatic pronunciation assessment can also be used to help diagnose and treat

speech disorders Speech disorders, impairments, or impediments, are a type of communication disorder in which normal manner of articulation, speech is disrupted. This can mean fluency disorders like stuttering and cluttering. Someone who is unable to speak due to ...

such as

apraxia Apraxia is a motor disorder caused by damage to the brain (specifically the posterior parietal cortex or corpus callosum), which causes difficulty with motor planning to perform tasks or movements. The nature of the damage determines the di ...

. Assessing authentic listener intelligibility is essential for avoiding inaccuracies from accent bias, especially in high-stakes assessments; from words with multiple correct pronunciations; and from phoneme coding errors in machine-readable pronunciation dictionaries. In 2022, researchers found that some newer speech to text systems, based on end-to-end reinforcement learning to map audio signals directly into words, produce word and phrase confidence scores very closely correlated with genuine listener intelligibility. In the

Common European Framework of Reference for Languages The Common European Framework of Reference for Languages: Learning, Teaching, Assessment, abbreviated in English as CEFR, CEF, or CEFRL, is a guideline used to describe achievements of learners of foreign languages across Europe and, increasingl ...

(CEFR) assessment criteria for "overall phonological control", intelligibility outweighs formally correct pronunciation at all levels.

Health care

Medical documentation

In the

health care Health care, or healthcare, is the improvement or maintenance of health via the preventive healthcare, prevention, diagnosis, therapy, treatment, wikt:amelioration, amelioration or cure of disease, illness, injury, and other disability, physic ...

sector, speech recognition can be implemented in front-end or back-end of the medical documentation process. Front-end speech recognition is where the provider dictates into a speech-recognition engine, the recognized words are displayed as they are spoken, and the dictator is responsible for editing and signing off on the document. Back-end or deferred speech recognition is where the provider dictates into a

digital dictation A dictation machine is a sound recording device most commonly used to record Speech communication, speech for playback or to be typed into print. It includes digital voice recorders and tape recorder. The name "Dictaphone" is a trademark of the ...

system, the voice is routed through a speech-recognition machine and the recognized draft document is routed along with the original voice file to the editor, where the draft is edited and report finalized. Deferred speech recognition is widely used in the industry currently. One of the major issues relating to the use of speech recognition in healthcare is that the

American Recovery and Reinvestment Act of 2009 The American Recovery and Reinvestment Act of 2009 (ARRA) (), nicknamed the Recovery Act, was a Stimulus (economics), stimulus package enacted by the 111th U.S. Congress and signed into law by President Barack Obama in February 2009. Developed ...

( ARRA) provides for substantial financial benefits to physicians who utilize an EMR according to "Meaningful Use" standards. These standards require that a substantial amount of data be maintained by the EMR (now more commonly referred to as an

Electronic Health Record An electronic health record (EHR) is the systematized collection of electronically stored patient and population health information in a digital format. These records can be shared across different health care settings. Records are shared thro ...

or EHR). The use of speech recognition is more naturally suited to the generation of narrative text, as part of a radiology/pathology interpretation, progress note or discharge summary: the ergonomic gains of using speech recognition to enter structured discrete data (e.g., numeric values or codes from a list or a

controlled vocabulary A controlled vocabulary provides a way to organize knowledge for subsequent retrieval. Controlled vocabularies are used in subject indexing schemes, subject headings, thesauri, taxonomies and other knowledge organization systems. Controlled v ...

) are relatively minimal for people who are sighted and who can operate a keyboard and mouse. A more significant issue is that most EHRs have not been expressly tailored to take advantage of voice-recognition capabilities. A large part of the clinician's interaction with the EHR involves navigation through the user interface using menus, and tab/button clicks, and is heavily dependent on keyboard and mouse: voice-based navigation provides only modest ergonomic benefits. By contrast, many highly customized systems for radiology or pathology dictation implement voice "macros", where the use of certain phrases – e.g., "normal report", will automatically fill in a large number of default values and/or generate boilerplate, which will vary with the type of the exam – e.g., a chest X-ray vs. a gastrointestinal contrast series for a radiology system.

Therapeutic use

Prolonged use of speech recognition software in conjunction with

s has shown benefits to short-term-memory restrengthening in brain AVM patients who have been treated with resection. Further research needs to be conducted to determine cognitive benefits for individuals whose AVMs have been treated using radiologic techniques.

Military

High-performance fighter aircraft

Substantial efforts have been devoted in the last decade to the test and evaluation of speech recognition in

fighter aircraft Fighter aircraft (early on also ''pursuit aircraft'') are military aircraft designed primarily for air-to-air combat. In military conflict, the role of fighter aircraft is to establish air supremacy, air superiority of the battlespace. Domina ...

. Of particular note have been the US programme in speech recognition for the Advanced Fighter Technology Integration (AFTI)/

F-16 The General Dynamics F-16 Fighting Falcon is an American single-engine supersonic multirole fighter aircraft originally developed by General Dynamics for the United States Air Force (USAF). Designed as an air superiority day fighter, it e ...

aircraft (

F-16 VISTA The General Dynamics X-62 VISTA ("Variable Stability In-flight Simulator Test Aircraft") is an experimental aircraft, derived from the General Dynamics F-16 Fighting Falcon, F-16D Fighting Falcon, which was modified as a joint venture between Ge ...

), the programme in France for

Mirage A mirage is a naturally-occurring optical phenomenon in which light rays bend via refraction to produce a displaced image of distant objects or the sky. The word comes to English via the French ''(se) mirer'', from the Latin ''mirari'', mean ...

aircraft, and other programmes in the UK dealing with a variety of aircraft platforms. In these programmes, speech recognizers have been operated successfully in fighter aircraft, with applications including setting radio frequencies, commanding an autopilot system, setting steer-point coordinates and weapons release parameters, and controlling flight display. Working with Swedish pilots flying in the JAS-39 Gripen cockpit, Englund (2004) found recognition deteriorated with increasing g-loads. The report also concluded that adaptation greatly improved the results in all cases and that the introduction of models for breathing was shown to improve recognition scores significantly. Contrary to what might have been expected, no effects of the broken English of the speakers were found. It was evident that spontaneous speech caused problems for the recognizer, as might have been expected. A restricted vocabulary, and above all, a proper syntax, could thus be expected to improve recognition accuracy substantially. The

Eurofighter Typhoon The Eurofighter Typhoon is a European multinational twin-engine, supersonic, canard delta wing, multirole fighter. The Typhoon was designed originally as an air-superiority fighter and is manufactured by a consortium of Airbus, BAE Syste ...

, currently in service with the UK RAF, employs a speaker-dependent system, requiring each pilot to create a template. The system is not used for any safety-critical or weapon-critical tasks, such as weapon release or lowering of the undercarriage, but is used for a wide range of other cockpit functions. Voice commands are confirmed by visual and/or aural feedback. The system is seen as a major design feature in the reduction of pilot

workload The term workload can refer to several different yet related entities. An amount of labor An old definition refers to workload as the amount of work an individual has to do.Jex, S. M. (1998). Stress and job performance: Theory, research, and im ...

, and even allows the pilot to assign targets to his aircraft with two simple voice commands or to any of his wingmen with only five commands. Speaker-independent systems are also being developed and are under test for the

F-35 Lightning II The Lockheed Martin F-35 Lightning II is an American family of single-seat, single-engine, supersonic Stealth aircraft, stealth strike fighters. A multirole combat aircraft designed for both Air superiority fighter, air superiority and att ...

(JSF) and the Alenia Aermacchi M-346 Master lead-in fighter trainer. These systems have produced word accuracy scores in excess of 98%.

Helicopters

The problems of achieving high recognition accuracy under stress and noise are particularly relevant in the

helicopter A helicopter is a type of rotorcraft in which Lift (force), lift and thrust are supplied by horizontally spinning Helicopter rotor, rotors. This allows the helicopter to VTOL, take off and land vertically, to hover (helicopter), hover, and ...

environment as well as in the jet fighter environment. The acoustic noise problem is actually more severe in the helicopter environment, not only because of the high noise levels but also because the helicopter pilot, in general, does not wear a facemask, which would reduce acoustic noise in the

microphone A microphone, colloquially called a mic (), or mike, is a transducer that converts sound into an electrical signal. Microphones are used in many applications such as telephones, hearing aids, public address systems for concert halls and publi ...

. Substantial test and evaluation programmes have been carried out in the past decade in speech recognition systems applications in helicopters, notably by the U.S. Army Avionics Research and Development Activity (AVRADA) and by the Royal Aerospace Establishment ( RAE) in the UK. Work in France has included speech recognition in the Puma helicopter. There has also been much useful work in

Canada Canada is a country in North America. Its Provinces and territories of Canada, ten provinces and three territories extend from the Atlantic Ocean to the Pacific Ocean and northward into the Arctic Ocean, making it the world's List of coun ...

. Results have been encouraging, and voice applications have included: control of communication radios, setting of

navigation Navigation is a field of study that focuses on the process of monitoring and controlling the motion, movement of a craft or vehicle from one place to another.Bowditch, 2003:799. The field of navigation includes four general categories: land navig ...

systems, and control of an automated target handover system. As in fighter applications, the overriding issue for voice in helicopters is the impact on pilot effectiveness. Encouraging results are reported for the AVRADA tests, although these represent only a feasibility demonstration in a test environment. Much remains to be done both in speech recognition and in overall

speech technology Speech technology relates to the technologies designed to duplicate and respond to the human voice. They have many uses. These include aid to the voice-disabled, the hearing-disabled, and the blind, along with communication with computers without a ...

in order to consistently achieve performance improvements in operational settings.

Training air traffic controllers

Training for air traffic controllers (ATC) represents an excellent application for speech recognition systems. Many ATC training systems currently require a person to act as a "pseudo-pilot", engaging in a voice dialog with the trainee controller, which simulates the dialog that the controller would have to conduct with pilots in a real ATC situation. Speech recognition and

synthesis Synthesis or synthesize may refer to: Science Chemistry and biochemistry *Chemical synthesis, the execution of chemical reactions to form a more complex molecule from chemical precursors **Organic synthesis, the chemical synthesis of organi ...

techniques offer the potential to eliminate the need for a person to act as a pseudo-pilot, thus reducing training and support personnel. In theory, Air controller tasks are also characterized by highly structured speech as the primary output of the controller, hence reducing the difficulty of the speech recognition task should be possible. In practice, this is rarely the case. The FAA document 7110.65 details the phrases that should be used by air traffic controllers. While this document gives less than 150 examples of such phrases, the number of phrases supported by one of the simulation vendors speech recognition systems is in excess of 500,000. The USAF, USMC, US Army, US Navy, and FAA as well as a number of international ATC training organizations such as the Royal Australian Air Force and Civil Aviation Authorities in Italy, Brazil, and Canada are currently using ATC simulators with speech recognition from a number of different vendors.

Telephony and other domains

ASR is now commonplace in the field of

telephony Telephony ( ) is the field of technology involving the development, application, and deployment of telecommunications services for the purpose of electronic transmission of voice, fax, or data, between distant parties. The history of telephony is ...

and is becoming more widespread in the field of

computer gaming A video game or computer game is an electronic game that involves interaction with a user interface or input device (such as a joystick, controller, keyboard, or motion sensing device) to generate visual feedback from a display device, mos ...

and simulation. In telephony systems, ASR is now being predominantly used in contact centers by integrating it with IVR systems. Despite the high level of integration with word processing in general personal computing, in the field of document production, ASR has not seen the expected increases in use. The improvement of mobile processor speeds has made speech recognition practical in

smartphone A smartphone is a mobile phone with advanced computing capabilities. It typically has a touchscreen interface, allowing users to access a wide range of applications and services, such as web browsing, email, and social media, as well as multi ...

s. Speech is used mostly as a part of a user interface, for creating predefined or custom speech commands.

People with disabilities

People with disabilities can benefit from speech recognition programs. For individuals that are Deaf or Hard of Hearing, speech recognition software is used to automatically generate a closed-captioning of conversations such as discussions in conference rooms, classroom lectures, and/or religious services. Students who are blind (see

Blindness and education The subject of blindness and education has included evolving approaches and public perceptions of how best to address the special needs of blind students. The practice of institutionalizing the blindness, blind in asylums has a history extending b ...

) or have very low vision can benefit from using the technology to convey words and then hear the computer recite them, as well as use a computer by commanding with their voice, instead of having to look at the screen and keyboard. Students who are physically disabled have a

Repetitive strain injury A repetitive strain injury (RSI) is an injury to part of the musculoskeletal or nervous system caused by repetitive use, vibrations, compression or long periods in a fixed position. Other common names include repetitive stress injury, repetitiv ...

/other injuries to the upper extremities can be relieved from having to worry about handwriting, typing, or working with scribe on school assignments by using speech-to-text programs. They can also utilize speech recognition technology to enjoy searching the Internet or using a computer at home without having to physically operate a mouse and keyboard. Speech recognition can allow students with learning disabilities to become better writers. By saying the words aloud, they can increase the fluidity of their writing, and be alleviated of concerns regarding spelling, punctuation, and other mechanics of writing. Also, see

Learning disability Learning disability, learning disorder, or learning difficulty (British English) is a condition in the brain that causes difficulties comprehending or processing information and can be caused by several different factors. Given the "difficulty ...

. The use of voice recognition software, in conjunction with a digital audio recorder and a personal computer running word-processing software has proven to be positive for restoring damaged short-term memory capacity, in stroke and craniotomy individuals. Speech recognition is also very useful for people who have difficulty using their hands, ranging from mild repetitive stress injuries to involve disabilities that preclude using conventional computer input devices. In fact, people who used the keyboard a lot and developed RSI became an urgent early market for speech recognition. Speech recognition is used in

deaf Deafness has varying definitions in cultural and medical contexts. In medical contexts, the meaning of deafness is hearing loss that precludes a person from understanding spoken language, an audiological condition. In this context it is written ...

, such as voicemail to text, relay services, and captioned telephone. Individuals with learning disabilities who have problems with thought-to-paper communication (essentially they think of an idea but it is processed incorrectly causing it to end up differently on paper) can possibly benefit from the software but the technology is not bug proof. Also the whole idea of speak to text can be hard for intellectually disabled person's due to the fact that it is rare that anyone tries to learn the technology to teach the person with the disability. This type of technology can help those with dyslexia but other disabilities are still in question. The effectiveness of the product is the problem that is hindering it from being effective. Although a kid may be able to say a word depending on how clear they say it the technology may think they are saying another word and input the wrong one. Giving them more work to fix, causing them to have to take more time with fixing the wrong word.

Further applications

Aerospace Aerospace is a term used to collectively refer to the atmosphere and outer space. Aerospace activity is very diverse, with a multitude of commercial, industrial, and military applications. Aerospace engineering consists of aeronautics and astron ...

(e.g.

space exploration Space exploration is the process of utilizing astronomy and space technology to investigate outer space. While the exploration of space is currently carried out mainly by astronomers with telescopes, its physical exploration is conducted bo ...

spacecraft A spacecraft is a vehicle that is designed spaceflight, to fly and operate in outer space. Spacecraft are used for a variety of purposes, including Telecommunications, communications, Earth observation satellite, Earth observation, Weather s ...

, etc.) NASA's

Mars Polar Lander The Mars Polar Lander, also known as the Mars Surveyor '98 Lander, was a 290-kilogram uncrewed spacecraft lander launched by NASA on January 3, 1999, to study the soil and climate of Planum Australe, a region near the south pole on Mars. It fo ...

used speech recognition technology from Sensory, Inc. in the Mars Microphone on the Lander *Automatic

subtitling Subtitles are texts representing the contents of the audio in a film, television show, opera or other audiovisual media. Subtitles might provide a transcription or translation of spoken dialogue. Although naming conventions can vary, caption ...

with speech recognition *Automatic

emotion recognition Emotion recognition is the process of identifying human emotion. People vary widely in their accuracy at recognizing the emotions of others. Use of technology to help people with emotion recognition is a relatively nascent research area. Gener ...

*Automatic shot listing in audiovisual production *

Automatic translation Machine translation is use of computational techniques to translate text or speech from one language to another, including the contextual, idiomatic and pragmatic nuances of both languages. Early approaches were mostly rule-based or statisti ...

eDiscovery Electronic discovery (also ediscovery or e-discovery) refers to discovery (law), discovery in legal proceedings such as litigation, government investigations, or Freedom of information laws by country, Freedom of Information Act requests, where t ...

(Legal discovery) * Hands-free computing: Speech recognition computer

user interface In the industrial design field of human–computer interaction, a user interface (UI) is the space where interactions between humans and machines occur. The goal of this interaction is to allow effective operation and control of the machine fro ...

Home automation Home automation or domotics is building automation for a home. A home automation system will monitor and/or control home attributes such as lighting, climate, entertainment systems, and appliances. It may also include home security such ...

Interactive voice response Interactive voice response (IVR) is a technology that allows telephone users to interact with a computer-operated telephone system through the use of voice and DTMF tones input with a keypad. In telephony, IVR allows customers to interact with a ...

Mobile telephony Mobile telephony is the provision of wireless telephone services to mobile phones, distinguishing it from fixed-location telephony provided via landline phones. Traditionally, telephony specifically refers to voice communication, though th ...

, including mobile email * Multimodal interaction *Real Time Captioning *

Robotics Robotics is the interdisciplinary study and practice of the design, construction, operation, and use of robots. Within mechanical engineering, robotics is the design and construction of the physical structures of robots, while in computer s ...

*Security, including usage with other biometric scanners for

multi-factor authentication Multi-factor authentication (MFA; two-factor authentication, or 2FA) is an electronic authentication method in which a user is granted access to a website or application only after successfully presenting two or more distinct types of evidence ...

*Speech to text (transcription of speech into text, real time video captioning, Court reporting ) *

Telematics Telematics is an interdisciplinary field encompassing telecommunications, vehicular technologies (road transport, road safety, etc.), electrical engineering (sensors, instrumentation, wireless communications, etc.), and computer science (multimedia ...

(e.g. vehicle Navigation Systems) * Transcription (digital speech-to-text) *

Video games A video game or computer game is an electronic game that involves interaction with a user interface or input device (such as a joystick, game controller, controller, computer keyboard, keyboard, or motion sensing device) to generate visual fe ...

, with '' Tom Clancy's EndWar'' and '' Lifeline'' as working examples *

Virtual assistant A virtual assistant (VA) is a software agent that can perform a range of tasks or services for a user based on user input such as commands or questions, including verbal ones. Such technologies often incorporate chatbot capabilities to streaml ...

(e.g. Apple's Siri)

Performance

The performance of speech recognition systems is usually evaluated in terms of accuracy and speed. Accuracy is usually rated with

word error rate Word error rate (WER) is a common metric of the performance of a speech recognition or machine translation system. The WER metric typically ranges from 0 to 1, where 0 indicates that the compared pieces of text are exactly identical, and 1 (or larg ...

(WER), whereas speed is measured with the real time factor. Other measures of accuracy include Single Word Error Rate (SWER) and Command Success Rate (CSR). Speech recognition by machine is a very complex problem, however. Vocalizations vary in terms of accent, pronunciation, articulation, roughness, nasality, pitch, volume, and speed. Speech is distorted by a background noise and echoes, electrical characteristics. Accuracy of speech recognition may vary with the following: * Vocabulary size and confusability * Speaker dependence versus independence * Isolated, discontinuous or continuous speech * Task and language constraints * Read versus spontaneous speech * Adverse conditions

Accuracy

As mentioned earlier in this article, the accuracy of speech recognition may vary depending on the following factors: * Error rates increase as the vocabulary size grows: ::e.g. the 10 digits "zero" to "nine" can be recognized essentially perfectly, but vocabulary sizes of 200, 5000 or 100000 may have error rates of 3%, 7%, or 45% respectively. * Vocabulary is hard to recognize if it contains confusing letters: ::e.g. the 26 letters of the English alphabet are difficult to discriminate because they are confusing words (most notoriously, the E-set: "B, C, D, E, G, P, T, V, Z — when "Z" is pronounced "zee" rather than "zed" depending on the English region); an 8% error rate is considered good for this vocabulary. * Speaker dependence vs. independence: :: A speaker-dependent system is intended for use by a single speaker. :: A speaker-independent system is intended for use by any speaker (more difficult). * Isolated, Discontinuous or continuous speech :: With isolated speech, single words are used, therefore it becomes easier to recognize the speech. With discontinuous speech full sentences separated by silence are used, therefore it becomes easier to recognize the speech as well as with isolated speech.
With continuous speech naturally spoken sentences are used, therefore it becomes harder to recognize the speech, different from both isolated and discontinuous speech. * Task and language constraints **e.g. Querying application may dismiss the hypothesis "The apple is red." **e.g. Constraints may be semantic; rejecting "The apple is angry." **e.g. Syntactic; rejecting "Red is apple the." Constraints are often represented by grammar. * Read vs. Spontaneous Speech – When a person reads it's usually in a context that has been previously prepared, but when a person uses spontaneous speech, it is difficult to recognize the speech because of the disfluencies (like "uh" and "um", false starts, incomplete sentences, stuttering, coughing, and laughter) and limited vocabulary. * Adverse conditions – Environmental noise (e.g. Noise in a car or a factory). Acoustical distortions (e.g. echoes, room acoustics) Speech recognition is a multi-leveled pattern recognition task. * Acoustical signals are structured into a hierarchy of units, e.g.

Phoneme A phoneme () is any set of similar Phone (phonetics), speech sounds that are perceptually regarded by the speakers of a language as a single basic sound—a smallest possible Phonetics, phonetic unit—that helps distinguish one word fr ...

s, Words, Phrases, and Sentences; * Each level provides additional constraints; e.g. Known word pronunciations or legal word sequences, which can compensate for errors or uncertainties at a lower level; * This hierarchy of constraints is exploited. By combining decisions probabilistically at all lower levels, and making more deterministic decisions only at the highest level, speech recognition by a machine is a process broken into several phases. Computationally, it is a problem in which a sound pattern has to be recognized or classified into a category that represents a meaning to a human. Every acoustic signal can be broken into smaller more basic sub-signals. As the more complex sound signal is broken into the smaller sub-sounds, different levels are created, where at the top level we have complex sounds, which are made of simpler sounds on the lower level, and going to lower levels, even more, we create more basic and shorter and simpler sounds. At the lowest level, where the sounds are the most fundamental, a machine would check for simple and more probabilistic rules of what sound should represent. Once these sounds are put together into more complex sounds on upper level, a new set of more deterministic rules should predict what the new complex sound should represent. The most upper level of a deterministic rule should figure out the meaning of complex expressions. In order to expand our knowledge about speech recognition, we need to take into consideration neural networks. There are four steps of neural network approaches: * Digitize the speech that we want to recognize For telephone speech the sampling rate is 8000 samples per second; * Compute features of spectral-domain of the speech (with Fourier transform); computed every 10 ms, with one 10 ms section called a frame; Analysis of four-step neural network approaches can be explained by further information. Sound is produced by air (or some other medium) vibration, which we register by ears, but machines by receivers. Basic sound creates a wave which has two descriptions:

amplitude The amplitude of a periodic variable is a measure of its change in a single period (such as time or spatial period). The amplitude of a non-periodic signal is its magnitude compared with a reference value. There are various definitions of am ...

(how strong is it), and

frequency Frequency is the number of occurrences of a repeating event per unit of time. Frequency is an important parameter used in science and engineering to specify the rate of oscillatory and vibratory phenomena, such as mechanical vibrations, audio ...

(how often it vibrates per second). Accuracy can be computed with the help of word error rate (WER). Word error rate can be calculated by aligning the recognized word and referenced word using dynamic string alignment. The problem may occur while computing the word error rate due to the difference between the sequence lengths of the recognized word and referenced word. The formula to compute the word error rate (WER) is:

WER =

where ''s'' is the number of substitutions, ''d'' is the number of deletions, ''i'' is the number of insertions, and ''n'' is the number of word references. While computing, the word recognition rate (WRR) is used. The formula is: :

WRR = 1 - WER =  =

where ''h'' is the number of correctly recognized words: :

h = n -(s+d).

Security concerns

Speech recognition can become a means of attack, theft, or accidental operation. For example, activation words like "Alexa" spoken in an audio or video broadcast can cause devices in homes and offices to start listening for input inappropriately, or possibly take an unwanted action. Voice-controlled devices are also accessible to visitors to the building, or even those outside the building if they can be heard inside. Attackers may be able to gain access to personal information, like calendar, address book contents, private messages, and documents. They may also be able to impersonate the user to send messages or make online purchases. Two attacks have been demonstrated that use artificial sounds. One transmits ultrasound and attempt to send commands without nearby people noticing. The other adds small, inaudible distortions to other speech or music that are specially crafted to confuse the specific speech recognition system into recognizing music as speech, or to make what sounds like one command to a human sound like a different command to the system.

Further information

Conferences and journals

Popular speech recognition conferences held each year or two include SpeechTEK and SpeechTEK Europe, ICASSP, Interspeech/Eurospeech, and the IEEE ASRU. Conferences in the field of

natural language processing Natural language processing (NLP) is a subfield of computer science and especially artificial intelligence. It is primarily concerned with providing computers with the ability to process data encoded in natural language and is thus closely related ...

, such as ACL, NAACL, EMNLP, and HLT, are beginning to include papers on

speech processing Speech processing is the study of speech signals and the processing methods of signals. The signals are usually processed in a digital representation, so speech processing can be regarded as a special case of digital signal processing, applied to ...

. Important journals include the

IEEE The Institute of Electrical and Electronics Engineers (IEEE) is an American 501(c)(3) organization, 501(c)(3) public charity professional organization for electrical engineering, electronics engineering, and other related disciplines. The IEEE ...

Transactions on Speech and Audio Processing (later renamed

Transactions on Audio, Speech and Language Processing and since Sept 2014 renamed

/ACM Transactions on Audio, Speech and Language Processing—after merging with an ACM publication), Computer Speech and Language, and Speech Communication.

Books

Books like "Fundamentals of Speech Recognition" by Lawrence Rabiner can be useful to acquire basic knowledge but may not be fully up to date (1993). Another good source can be "Statistical Methods for Speech Recognition" by Frederick Jelinek and "Spoken Language Processing (2001)" by Xuedong Huang etc., "Computer Speech", by Manfred R. Schroeder, second edition published in 2004, and "Speech Processing: A Dynamic and Optimization-Oriented Approach" published in 2003 by Li Deng and Doug O'Shaughnessey. The updated textbook ''Speech and Language Processing'' (2008) by Jurafsky and Martin presents the basics and the state of the art for ASR.

Speaker recognition Speaker recognition is the identification of a person from characteristics of voices. It is used to answer the question "Who is speaking?" The term voice recognition can refer to ''speaker recognition'' or speech recognition. Speaker verification ...

also uses the same features, most of the same front-end processing, and classification techniques as is done in speech recognition. A comprehensive textbook, "Fundamentals of Speaker Recognition" is an in depth source for up to date details on the theory and practice. A good insight into the techniques used in the best modern systems can be gained by paying attention to government sponsored evaluations such as those organised by

(the largest speech recognition-related project ongoing as of 2007 is the GALE project, which involves both speech recognition and translation components). A good and accessible introduction to speech recognition technology and its history is provided by the general audience book "The Voice in the Machine. Building Computers That Understand Speech" by Roberto Pieraccini (2012). The most recent book on speech recognition is ''Automatic Speech Recognition: A Deep Learning Approach'' (Publisher: Springer) written by Microsoft researchers D. Yu and L. Deng and published near the end of 2014, with highly mathematically oriented technical detail on how deep learning methods are derived and implemented in modern speech recognition systems based on DNNs and related deep learning methods. A related book, published earlier in 2014, "Deep Learning: Methods and Applications" by L. Deng and D. Yu provides a less technical but more methodology-focused overview of DNN-based speech recognition during 2009–2014, placed within the more general context of deep learning applications including not only speech recognition but also image recognition, natural language processing, information retrieval, multimodal processing, and multitask learning.

Software

In terms of freely available resources,

Sphinx A sphinx ( ; , ; or sphinges ) is a mythical creature with the head of a human, the body of a lion, and the wings of an eagle. In Culture of Greece, Greek tradition, the sphinx is a treacherous and merciless being with the head of a woman, th ...

toolkit is one place to start to both learn about speech recognition and to start experimenting. Another resource (free but copyrighted) is the HTK book (and the accompanying HTK toolkit). For more recent and state-of-the-art techniques, Kaldi toolkit can be used. In 2017

Mozilla Mozilla is a free software community founded in 1998 by members of Netscape. The Mozilla community uses, develops, publishes and supports Mozilla products, thereby promoting free software and open standards. The community is supported institution ...

launched the open source project called Common Voice to gather big database of voices that would help build free speech recognition project DeepSpeech (available free at

GitHub GitHub () is a Proprietary software, proprietary developer platform that allows developers to create, store, manage, and share their code. It uses Git to provide distributed version control and GitHub itself provides access control, bug trackin ...

), using Google's open source platform

TensorFlow TensorFlow is a Library (computing), software library for machine learning and artificial intelligence. It can be used across a range of tasks, but is used mainly for Types of artificial neural networks#Training, training and Statistical infer ...

. When Mozilla redirected funding away from the project in 2020, it was forked by its original developers as Coqui STT using the same open-source license. Google Gboard supports speech recognition on all Android applications. It can be activated through the

icon An icon () is a religious work of art, most commonly a painting, in the cultures of the Eastern Orthodox, Oriental Orthodox, Catholic Church, Catholic, and Lutheranism, Lutheran churches. The most common subjects include Jesus, Mary, mother of ...

. Speech recognition can be activated in

Microsoft Windows Windows is a Product lining, product line of Proprietary software, proprietary graphical user interface, graphical operating systems developed and marketed by Microsoft. It is grouped into families and subfamilies that cater to particular sec ...

operating systems by pressing Windows logo key + Ctrl + S. The commercial cloud based speech recognition APIs are broadly available. For more software resources, see List of speech recognition software.