Misplaced Pages

Dictaphone

Article snapshot taken from Wikipedia with creative commons attribution-sharealike license. Give it a read and then ask your questions in the chat. We can research this topic together.

Dictaphone was an American company founded by Alexander Graham Bell that produced dictation machines . It is now a division of Nuance Communications , based in Burlington, Massachusetts .

#484515

119-642: Although the name "Dictaphone" is a trademark , it has become genericized as a means to refer to any dictation machine . The Volta Laboratory was established by Alexander Graham Bell in Washington, D.C. in 1881. When the Laboratory's sound-recording inventions were sufficiently developed with the assistance of Charles Sumner Tainter and others, Bell and his associates set up the Volta Graphophone Company , which later merged with

238-580: A deep learning method called Long short-term memory (LSTM), a recurrent neural network 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) started to outperform traditional speech recognition in certain applications. In 2015, Google's speech recognition reportedly experienced

357-430: A finite state transducer verifying certain assumptions. 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

476-592: A "first-to-file" system, which grants rights to the first entity to register the mark. However, well-known trademarks are an exception, as they may receive protection even without registration. In contrast, a few countries, like the United States, Canada, and Australia, follow a "first-to-use" or hybrid system, where using the mark in commerce can establish certain rights, even without registration. However, registration in these countries still provides stronger legal protection and enforcement. For example, in

595-519: A chest X-ray vs. a gastrointestinal contrast series for a radiology system. Prolonged use of speech recognition software in conjunction with word processors 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. Substantial efforts have been devoted in

714-486: A circle of events, as Dictaphone had been sold to Lernout & Hauspie prior to L&H's bankruptcy which resulted in Dictaphone becoming an independent company. In March 2007, Nuance acquired Focus Informatics and, with the intention of further expansion in its healthcare-transcription business, linked it with its Dictaphone division. Trademark A trademark (also written trade mark or trade-mark )

833-434: 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 processors or emails ), and aircraft (usually termed direct voice input ). Automatic pronunciation assessment

952-407: A combination hidden Markov model, which includes both the acoustic and language model information and combining it statically beforehand (the finite state transducer , 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

1071-444: A company or product. A trademark, by contrast, offers legal protection for a brand with enforceable rights over the brand's identity and distinguishing elements. Trademark law is designed to fulfill the public policy objective of consumer protection , by preventing the public from being misled as to the origin or quality of a product or service. By identifying the commercial source of products and services, trademarks facilitate

1190-467: A competitor uses the same or a confusingly similar trademark for the same or similar products in a jurisdiction where the trademark is protected. This concept is recognized in many jurisdictions, including the United States, the European Union, and other countries, though specific legal standards may vary. To establish trademark infringement in court, the plaintiff generally must show: Trademark

1309-577: A dictating machine using an erasable belt made of magnetic tape which enabled the user to correct dictation errors rather than marking errors on a paper tab. Dictaphone in turn added magnetic recording models while still selling the models recording on the Lexan belts. Machines based on magnetic tape recording were introduced in the late seventies, initially using the standard compact (or "C") cassette , but soon, in dictation machines, using mini-cassettes or microcassettes instead. Using smaller cassette sizes

SECTION 10

#1732787082485

1428-446: 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 and delta-delta coefficients to capture speech dynamics and in addition, might use heteroscedastic linear discriminant analysis (HLDA); or might skip

1547-630: A distinctive label or ticket'. In the United States , Congress first attempted to establish a federal trademark regime in 1870. This statute purported to be an exercise of Congress' Copyright Clause powers. However, the Supreme Court struck down the 1870 statute in the Trade-Mark Cases later on in the decade. In 1881, Congress passed a new trademark act, this time according to its Commerce Clause powers. Congress revised

1666-648: A dramatic performance jump of 49% through CTC-trained LSTM, which is now available through Google Voice to all smartphone users. Transformers , a type of neural network based solely on "attention", have been widely adopted in computer vision and language modeling, 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 modeling

1785-461: A few years into the 2000s. But these methods never won over the non-uniform internal-handcrafting Gaussian mixture model / hidden Markov model (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 and weak temporal correlation structure in the neural predictive models. All these difficulties were in addition to

1904-496: 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

2023-437: A generic product or service name. They should stand out from the surrounding text using capital letters, bold type, italics, color, underlining, quotation marks, or a unique stylized format. For example, say “LEGO® toy blocks” instead of “Lego’s.” A trademark may be designated by the following symbols: While ™ and ℠ apply to unregistered marks (™ for goods and ℠ for services), the ® symbol indicates official registration with

2142-452: A list or a controlled vocabulary ) 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

2261-609: A loss of rights in the trademark. It is still possible to make significant changes to the underlying goods or services during a sale without jeopardizing the trademark, but companies will often contract with the sellers to help transition the mark and goods or services to the new owners to ensure continuity of the trademark. Trademarks are often confused with patents and copyrights . Although all three laws protect forms of intangible property, collectively known as intellectual property (IP), they each have different purposes and objectives: Among these types of IP, only trademark law offers

2380-430: A renaissance of applications of deep feedforward neural networks for speech recognition. 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

2499-568: 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 and big data . 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. The key areas of growth were: vocabulary size, speaker independence, and processing speed. Raj Reddy

SECTION 20

#1732787082485

2618-509: 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 . Speech can be thought of as a Markov model 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,

2737-534: A single unit. Although DTW would be superseded by later algorithms, the technique carried on. Achieving speaker independence remained unsolved at this time period. During the late 1960s Leonard Baum developed the mathematics of Markov chains at the Institute for Defense Analysis . A decade later, at CMU, Raj Reddy's students James Baker and Janet M. Baker began using the hidden Markov model (HMM) for speech recognition. James Baker had learned about HMMs from

2856-513: 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 operating system. L&H was an industry leader until an accounting scandal brought an end to

2975-439: 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 , 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

3094-470: A substantial amount of data be maintained by the EMR (now more commonly referred to as an Electronic Health Record 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

3213-467: 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. The 1980s also saw the introduction of the n-gram language model. Much of the progress in the field is owed to the rapidly increasing capabilities of computers. At

3332-701: A whole. Trademark protection does not apply to utilitarian features of a product such as the plastic interlocking studs on Lego bricks. The earliest examples of use of markings date back to around 15,000 years ago in Prehistory . Similar to branding practices, the Lascaux cave paintings in France, for instance, depict bulls with marks, which experts believe may have served as personal marks to indicate livestock ownership. Around 6,000 years ago, Egyptian masonry featured quarry marks and stonecutters' signs to identify

3451-493: Is "escalator," which was once a trademark. In contrast, patents have a fixed term, typically lasting 20 years from the filing date, after which the invention enters the public domain. Copyrights generally last for the life of the author plus an additional 50 to 70 years (depending on the jurisdiction), after which the protected work enters the public domain. Although intellectual property laws such as these are theoretically distinct, more than one type may afford protection to

3570-417: Is a form of intellectual property that consists of a word, phrase, symbol, design, or a combination that identifies a product or service from a particular source and distinguishes it from others. Trademarks can also extend to non-traditional marks like drawings, symbols, 3D shapes like product designs or packaging, sounds, scents, or specific colors used to create a unique identity. For example, Pepsi®

3689-958: 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 emerged as an attractive acoustic modeling 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 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

Dictaphone - Misplaced Pages Continue

3808-501: Is a registered trademark associated with soft drinks, and the distinctive shape of the Coca-Cola® bottle is a registered trademark protecting Coca-Cola's packaging design. The primary function of a trademark is to identify the source of goods or services and prevent consumers from confusing them with those from other sources. Legal protection for trademarks is typically secured through registration with governmental agencies, such as

3927-588: Is an artificial neural network 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 modeling 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

4046-467: 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 ( ARRA ) provides for substantial financial benefits to physicians who utilize an EMR according to "Meaningful Use" standards. These standards require that

4165-493: 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

4284-411: 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.,

4403-574: Is incapable of learning the language due to conditional independence 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 expanded on the work with extremely large datasets and demonstrated some commercial success in Chinese Mandarin and English. In 2016, University of Oxford presented LipNet ,

4522-435: Is inherently distinctive (able to identify and distinguish a single source of goods or services), often falling into categories such as suggestive, fanciful, or arbitrary, and is therefore registerable. In contrast, weak trademarks tend to be either descriptive or generic and may not be registerable. The registration process typically begins with a trademark clearance search to identify potential conflicts that could prevent

4641-503: Is required to act as the "basic mark." In the international application, the trademark owner can designate one or more Madrid System Member countries for protection. Each designated country’s trademark office will review the Madrid application under its local laws to grant or refuse protection. In the United States, for example, a trademark must first be registered or pending with the U.S. Patent and Trademark Office (USPTO) to serve as

4760-466: Is subject to various defenses, such as abandonment, limitations on geographic scope , and fair use. In the United States, the fair use defense protects many of the interests in free expression related to those protected by the First Amendment . Fair use may be asserted on two grounds, either that the alleged infringer is using the mark to describe accurately an aspect of its products, or that

4879-735: 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, 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. 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

Dictaphone - Misplaced Pages Continue

4998-413: 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 or verify the identity of a speaker as part of

5117-606: Is used to refer to both trademarks and service marks. Similarly, the World Intellectual Property Organization (WIPO) defines a trademark as a sign capable of distinguishing the goods or services of one enterprise from those of other enterprises. WIPO administers the Madrid Protocol , which allows trademark owners worldwide to file one application to register their trademark in multiple countries. Almost anything that identifies

5236-687: The American Graphophone Company (founded in 1887) which itself later evolved into Columbia Records (founded as the Columbia Phonograph Company in 1889). The name "Dictaphone" was trademarked in 1907 by the Columbia Graphophone Company , which soon became the leading manufacturer of such devices. This perpetuated the use for voice recording of wax cylinders , which had otherwise been eclipsed by disc-based technology . Dictaphone

5355-498: The Common European Framework of Reference for Languages (CEFR) assessment criteria for "overall phonological control", intelligibility outweighs formally correct pronunciation at all levels. In the health care 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,

5474-515: 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

5593-647: The Paris Convention and the Madrid Protocol , simplify the registration and protection of trademarks across multiple countries. Additionally, the TRIPS Agreement sets minimum standards for trademark protection and enforcement that all member countries must follow. The term trademark can also be spelled trade mark in regions such as the EU, UK, and Australia, and as trade-mark in Canada. Despite

5712-575: 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

5831-592: The United States Patent and Trademark Office (USPTO) or the European Union Intellectual Property Office (EUIPO). Registration provides the owner certain exclusive rights and provides legal remedies against unauthorized use by others. Trademark laws vary by jurisdiction but generally allow owners to enforce their rights against infringement, dilution, or unfair competition. International agreements, such as

5950-461: The University of Montreal 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

6069-706: The computer science , linguistics and computer engineering fields. The reverse process is speech synthesis . Some speech recognition systems require "training" (also called "enrollment") where an individual speaker reads text or isolated vocabulary 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 interfaces such as voice dialing (e.g. "call home"), call routing (e.g. "I would like to make

SECTION 50

#1732787082485

6188-646: The "basic mark" necessary for Madrid filings. The trademark registration process with the USPTO generally follows these steps: Trademark owners seeking protection in multiple jurisdictions have two options: the Paris route, under the Paris Convention , or the Madrid System , which is administered by WIPO . The Paris route, covering 180 countries and also known as the "direct route," requires filing separate applications with each country’s IP office. In contrast,

6307-517: The DNN based on context dependent HMM states constructed by decision trees were adopted. 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 deep learning , in recent overview articles. One fundamental principle of deep learning

6426-628: The EARS program: IBM , a team led by BBN with LIMSI and Univ. of Pittsburgh , Cambridge University , and a team composed of ICSI , SRI and University of Washington . EARS funded the collection of the Switchboard telephone speech corpus containing 260 hours of recorded conversations from over 500 speakers. The GALE program focused on Arabic and Mandarin broadcast news speech. Google 's first effort at speech recognition came in 2007 after hiring some researchers from Nuance. The first product

6545-429: The European Union requires "genuine use" of the mark within a continuous five-year period following registration to maintain the trademark, with non-use potentially resulting in revocation. The trademark owner must enforce their rights to preserve the trademark's distinctiveness , prevent trademark infringement, and avoid dilution. Enforcement after registration generally involves: Trademark infringement occurs when

6664-468: The Madrid System streamlines the process by allowing a single Madrid application, built on an existing or applied-for national or regional registration (the "basic mark"), to extend protection to up to 131 countries. Unlike patents and copyrights , which have fixed expiration dates, trademark registrations typically have an initial term of 10 years and can be renewed indefinitely, as long as

6783-577: The Roman Empire. Other notable trademarks that have been used for a long time include Stella Artois , which claims use of its mark since 1366, and Löwenbräu , which claims use of its lion mark since 1383. The first trademark legislation was passed by the Parliament of England under the reign of King Henry III in 1266, which required all bakers to use a distinctive mark for the bread they sold. The first modern trademark laws emerged in

6902-528: The Trademark Act in 1905. The Lanham Act of 1946 updated the law and has served, with several amendments, as the primary federal law on trademarks. The Trade Marks Act 1938 in the United Kingdom set up the first registration system based on the "intent-to-use" principle. The Act also established an application publishing procedure and expanded the rights of the trademark holder to include

7021-492: The Trademark Electronic Search System (TESS) in 2023. A comprehensive clearance search involves checking the USPTO database for federally registered and applied-for trademarks, state trademark databases, and the internet to see if someone else has already registered that trademark or a similar one. The search should also include looking at both words and designs. To search for similar designs in

7140-486: The UK Patent Office for the first time. Registration was considered to comprise prima facie evidence of ownership of a trademark and registration of marks began on 1 January 1876. The 1875 Act defined a registrable trade mark as a device or mark, or name of an individual or firm printed in some particular and distinctive manner; or a written signature or copy of a written signature of an individual or firm; or

7259-400: The USPTO database, design search codes must be used. WIPO ’s Global Brand Database provides international access to trademarks and emblems. Trademark owners can either maintain protection at the national level or expand internationally through the Madrid System by building on their national registration. To pursue international protection, a national registration or pending application

SECTION 60

#1732787082485

7378-584: The United States, trademark rights are established either (1) through first use of the mark in commerce, creating common law rights limited to the geographic areas of use, or (2) through federal registration with the U.S. Patent and Trademark Office (USPTO), with use in commerce required to maintain the registration. Federal registration with the USPTO provides additional benefits, such as: Trademark law grants legal protection to "distinctive" trademarks, which are marks that allow consumers to easily associate them with specific products or services. A strong trademark

7497-544: The Year can identify herself as such on her website. Speech-recognition Speech recognition is an interdisciplinary subfield of computer science and computational linguistics that develops methodologies and technologies that enable the recognition and translation 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

7616-459: The alleged infringer is using the mark to identify the mark owner. One of the most visible proofs that trademarks provide a limited right in the U.S. comes from the comparative advertising that is seen throughout U.S. media. An example of the first type is that although Maytag owns the trademark "Whisper Quiet" for its dishwashers, makers of other products may describe their goods as being "whisper quiet" so long as these products do not fall under

7735-438: The barring of trademark use even in cases where confusion remained unlikely. This Act served as a model for similar legislation elsewhere. The oldest registered trademark has various claimants, enumerated below: Trademark protection can be acquired through registration and/or, in certain countries, through use. Globally, the most common method for establishing trademark rights is registration. Most countries operate under

7854-575: 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

7973-415: 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 and Google Brain and Bahdanau et al. of

8092-507: 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 and Navdeep Jaitly of the University of Toronto in 2014. The model consisted of recurrent neural networks and a CTC layer. Jointly, the RNN-CTC model learns the pronunciation and acoustic model together, however it

8211-420: The company and all its subsidiaries, including Dictaphone, were forced into bankruptcy protection . In early 2002, Dictaphone emerged from bankruptcy as a privately held organization , with Rob Schwager as its chairman and CEO . In 2004, it was split into three divisions: In June 2005, Dictaphone Corporation announced the sale of its Communications Recording Systems to NICE Systems for $ 38.5 million. This

8330-438: The company in 2001. The speech technology from L&H was bought by ScanSoft which became Nuance in 2005. Apple originally licensed software from Nuance to provide speech recognition capability to its digital assistant Siri . 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

8449-488: 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 phonemes (so that phonemes with different left and right context would have different realizations as HMM states); it would use cepstral normalization to normalize for

8568-519: The correctness of the learner's pronunciation and ideally their intelligibility to listeners, sometimes along with often inconsequential prosody such as intonation , pitch , tempo , rhythm , and stress . Pronunciation assessment is also used in reading tutoring , for example in products such as Microsoft Teams and from Amira Learning. Automatic pronunciation assessment can also be used to help diagnose and treat speech disorders such as apraxia . Assessing authentic listener intelligibility

8687-450: 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 hidden Markov models combined with feedforward artificial neural networks . Today, however, many aspects of speech recognition have been taken over by

8806-463: 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 , 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

8925-447: The demise of the dedicated word-processor, and the division was closed. In 1995, Pitney Bowes sold Dictaphone to the investment group Stonington Partners of Connecticut for a reported $ 462 million. Dictaphone thereafter sold a range of products that included speech-recognition and voicemail software with limited success as the market only existed among some early adopters despite its vertical markets' enhancements. In 2000, Dictaphone

9044-595: The different spellings, all three terms denote the same concept. In the United States, the Lanham Act defines a trademark as any word, phrase, symbol, design, or combination of these things used to identify goods or services. Trademarks help consumers recognize a brand in the marketplace and distinguish it from competitors. A service mark , also covered under the Lanham Act, is a type of trademark used to identify services rather than goods. The term trademark

9163-452: The end of the DARPA program in 1976, the best computer available to researchers was the PDP-10 with 4 MB ram. It could take up to 100 minutes to decode just 30 seconds of speech. Two practical products were: 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

9282-462: 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

9401-587: 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 Google DeepMind achieving 6 times better performance than human experts. In 2019, Nvidia launched two CNN-CTC ASR models, Jasper and QuarzNet, with an overall performance WER of 3%. Similar to other deep learning applications, transfer learning and domain adaptation are important strategies for reusing and extending

9520-445: The first hardware-independent dictation-management software system with integrated speech-recognition and workflow management. In 2008, iSpeech AG took over the activities and products of the former Calison AG. In February and March 2006, the remainder of Dictaphone was sold for $ 357 million to Nuance Communications (formerly ScanSoft ), ending its short tenure as an independent company that had begun in 2002. This, in effect, closed

9639-487: 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 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

9758-409: The identification of products and services which meet the expectations of consumers as to the quality and other characteristics. Trademarks may also serve as an incentive for manufacturers, providers, or suppliers to consistently provide quality products or services to maintain their business reputation. Furthermore, if a trademark owner does not maintain quality control and adequate supervision about

9877-557: 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 modeling 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

9996-868: The last decade to the test and evaluation of speech recognition in fighter aircraft . Of particular note have been the US program in speech recognition for the Advanced Fighter Technology Integration (AFTI) / F-16 aircraft ( F-16 VISTA ), the program in France for Mirage aircraft, and other programs in the UK dealing with a variety of aircraft platforms. In these programs, 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

10115-574: The late 19th century. In France, the first comprehensive trademark system in the world was passed into law in 1857 with the "Manufacture and Goods Mark Act". In Britain, the Merchandise Marks Act 1862 made it a criminal offense to imitate another's trade mark 'with intent to defraud or to enable another to defraud'. The passing of the Trade Marks Registration Act 1875 allowed formal registration of trademarks at

10234-414: The main application of this technology is computer-aided pronunciation teaching (CAPT) when combined with computer-aided instruction for computer-assisted language learning (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

10353-532: The manufacture and provision of products or services supplied by a licensee, such "naked licensing" will eventually adversely affect the owner's rights in the trademark. For US law see, ex. Eva's Bridal Ltd. v. Halanick Enterprises, Inc. 639 F.3d 788 (7th Cor. 2011). This proposition has, however, been watered down by the judgment of the House of Lords in the case of Scandecor Development AB v. Scandecor Marketing AB et al. [2001] UKHL 21; wherein it has been held that

10472-415: The mark remains in continuous use in commerce. If the trademark owner stops using the mark for too long (typically three to five years, depending on the jurisdiction), the trademark rights may be lost. For example, in the United States, trademark rights are based on use in commerce. If a mark is not used for three consecutive years, it is presumed abandoned and becomes vulnerable to challenges. Similarly,

10591-476: The mere fact that a bare license (the equivalent of the United States concept of a naked license) has been granted did not automatically mean that a trademark was liable to mislead. By the same token, trademark holders must be cautious in the sale of their mark for similar reasons as apply to licensing. When assigning an interest in a trademark, if the associated product or service is not transferred with it, then this may be an "assignment-in-gross" and could lead to

10710-453: The mid-seventies as the "Thought Tank". The recording medium did not need to be moved from where the dictation took place to the location such as a typing pool where the typists were located. This was normally operated via a dedicated in-house telephone system, enabling dictation to be made from a variety of locations within the hospital or other organizations with typing pools. One version calculated each typist's turnaround time and allocated

10829-548: 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. Automatic pronunciation 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,

10948-413: The next piece of dictation accordingly. Dictaphone was prominent in the provision of multi-channel recorders , used extensively in the emergency services to record emergency telephone calls (to numbers such as 911, 999, 112) and subsequent conversations. Additionally, Dictaphone at one point expanded its product line to market a line of electronic (desktop and portable) calculators. In 1979, Dictaphone

11067-433: The original LAS model. Latent Sequence Decompositions (LSD) was proposed by Carnegie Mellon University , MIT and Google Brain to directly emit sub-word units which are more natural than English characters; University of Oxford and Google DeepMind extended LAS to "Watch, Listen, Attend and Spell" (WLAS) to handle lip reading surpassing human-level performance. Typically a manual control input, for example by means of

11186-445: 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

11305-421: The possibility of perpetual rights, provided the trademark is continuously used and renewed. However, if a trademark is no longer in use, its registration may be subject to cancellation. Trademarks can also lose protection through genericide , a process where a trademark becomes so widely used to refer to a category of goods or services that it loses its distinctiveness and legal protection. A well-known example

11424-426: 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

11543-457: 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

11662-517: The pronunciation, acoustic, and language model . 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 Google and Apple (as of 2017 ) are deployed on

11781-404: 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 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

11900-484: 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 , 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

12019-527: The registration of the trademark. A comprehensive clearance search can help avoid costly and time-consuming issues, such as refusal to register, opposition or cancellation proceedings, or a trademark infringement lawsuit. In the United States, the USPTO maintains a publicly accessible database of registered trademarks. This database can be searched using the Trademark Search system, which replaced

12138-425: The relevant national authority. Using the ® symbol for unregistered trademarks is misleading and can be treated as unfair business practice. It may also result in civil or criminal penalties. A brand is a marketing concept that reflects how consumers perceive a product or service. It has a much wider meaning and refers to the proprietary visual, emotional, rational, and cultural image that customers associate with

12257-426: The same article. For example, the particular design of a bottle may qualify for copyright protection as a non-utilitarian [sculpture], or trademark protection based on its shape, or the ' trade dress ' appearance of the bottle as a whole may be protectable. Titles and character names from books or movies may also be protectable as trademarks while the works from which they are drawn may qualify for copyright protection as

12376-560: The same benchmark, which was funded by IBM Watson speech team on the same task. Both acoustic modeling and language modeling are important parts of modern statistically based speech recognition algorithms. Hidden Markov models (HMMs) are widely used in many systems. Language modeling is also used in many other natural language processing applications such as document classification or statistical machine translation . Modern general-purpose speech recognition systems are based on hidden Markov models. These are statistical models that output

12495-405: The same category of goods the trademark is protected under. An example of the second type is that Audi can run advertisements saying that a trade publication has rated an Audi model higher than a BMW model since they are only using "BMW" to identify the competitor. In a related sense, an auto mechanic can truthfully advertise that he services Volkswagens , and a former Playboy Playmate of

12614-535: 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 distances represented themselves as

12733-486: The source of goods or services can serve as a trademark. In addition to words, slogans, designs, or combinations of these, trademarks can also include non-traditional marks like sounds, scents, or colors. Under the broad heading of trademarks, there are several specific types commonly encountered, such as trade dress, collective marks, and certification marks: To maintain distinctiveness , trademarks should function as adjectives, not as nouns or verbs, and be paired with

12852-454: 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 modeling 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

12971-488: The stone's origin and the workers responsible. Wine amphorae marked with seals were also found in the tomb of Pharaoh Tutankhamun , who ruled ancient Egypt more than 3,000 years ago. Over 2,000 years ago, Chinese manufacturers sold goods marked with identifying symbols in the Mediterranean region. Trademarks have also been discovered on pottery, porcelain, and swords produced by merchants in ancient Greece and

13090-462: The training data. Examples are maximum mutual information (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

13209-624: 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 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

13328-520: Was acquired by the then-leading Belgian voice-recognition and translation company Lernout & Hauspie for nearly $ 1 billion. Lernout & Hauspie provided the voice-recognition technology for Dictaphone's enhanced voice-recognition-based transcription system. Soon after the purchase, however, the SEC raised questions about Lernout & Hauspie's finances, focusing on the supposedly skyrocketing income reported from its East Asian endeavors. Subsequently,

13447-540: Was considered a great bargain in the industry and came after NICE was ordered to pay Dictaphone $ 10 million in settlements related to a patent-infringement suit in late 2003. In September 2005, Dictaphone sold its IVS business outside the United States to a private Swiss group around its former VP Martin Niederberger, who formed Dictaphone IVS AG (later Calison AG) in Urdorf , Switzerland and developed "FRISBEE",

13566-525: Was important to the manufacturer for reducing the size of portable recorders. Walter D. Fuller became the director of the company in 1952. In 1969 he was appointed as chairman. In Japan, JVC was licensed to produce machines designed and developed by Dictaphone. Dictaphone and JVC later developed the picocassette , released in 1985, which was even smaller than a microcassette but retained a good recording quality and duration. Dictaphone also developed "endless loop" recording using magnetic tape, introduced in

13685-578: Was introduced during the later part of 2009 by Geoffrey Hinton 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). A Microsoft research executive called this innovation "the most dramatic change in accuracy since 1979". In contrast to

13804-431: Was purchased by Pitney Bowes and kept as a wholly owned but independent subsidiary . Dictaphone bought Dual Display Word Processor, a stiff competitor to Wang Laboratories , the industry leader. In 1982, it marketed a word processor from Symantec . The hardware sold for $ 5,950 in 1982. The software was an additional $ 600. The advent of the personal computer, MS-DOS , and general-purpose word-processing software saw

13923-426: Was spun off into a separate company in 1923 under the leadership of C. King Woodbridge. In 1947, having relied on wax-cylinder recording to the end of World War II , Dictaphone introduced its Dictabelt technology. This cut a mechanical groove into a Lexan plastic belt instead of into a wax cylinder. The advantage of the Lexan belt was that recordings were permanent and admissible in court. Eventually IBM introduced

14042-521: Was the first person to take on continuous speech recognition as a graduate student at Stanford University in the late 1960s. Previous systems required users to pause after each word. Reddy's system issued spoken commands for playing chess . 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

14161-449: 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 neural networks and denoising autoencoders are also under investigation. A deep feedforward neural network (DNN)

#484515