How Is Speech Recognition Systems Used in Innovative and Useful Ways - Essay Example

How is speech recognition systems used in innovative and useful ways. Do you think we will eventually eliminate the need for humans in common telephone interactions? Is this good or bad? 11-12-2011 How is speech recognition systems used in innovative and useful ways. Do you think we will eventually eliminate the need for humans in common telephone interactions? Is this good or bad? According to Obermaisser in a thesis proposed by him about ‘speech recognition system’, he emphasizes the need of human and machine interaction in this era of development. He is of the idea to increase number of words in the speech recognition systems so that the system do not become while depending only on the speakers. For the technological enhancements in this field, Obermaisser suggests the idea of using Domain-Separated Hidden Markov Models (DHMMs) that are an improved version of Dynamic Time Warping ( DTW) that were used classically. Alejandro Acero in his book ‘acoustical and environmental robustness in automatic speech recognition’ enlightens us about the early problems in ‘Unconstrained Automatic speech recognition’ (ACR) systems. While “Sphinx”, developed by CMU, is a speaker independent and large vocabulary system, Acero have some other ideas. Acero discuses the various factors that affect speech systems, some of them being constant and others are assumed. Such as the equipments that are being used, the amount of room reverberation and the difference caused by the difference in pitch of variety are people are some of the very relevant constant problems. Robert describes about the digitalized working of computer operated systems in the book “Computer Recognition Systems 4”. A spoken sentence is recognized after it passes through three stages of analysis. These three stages include 1) the given words are recognized, 2) words are arranged in ascending order and then filtered 3) then the overall sentences are recognized. The task of word recognition is performed by using models like HMM (hidden markov models). Then the words are filtered by a ‘Natural gram’ language solution software which is implied. The last of the three stages of sentence recognition is resolved by a “hidden markov model” made up of sentences. A chapter in the book “automatic speech recognition on mobile device and communication networks” discus channel characterization and modeling. Frame stand interleaving is also argued as a substitute to overcoming the consequence of burst channel crossing out. The chapter ends with an example of current standards for channel coding policies designed for distributed speech recognition (DSR). Blade Kotelly discuss the psychology of people interacting through speech recognition systems in his book” The art and business of speech recognition: creating the noble voice”. It happens in our material world that two of the speech recognition software of the same standards is launched at the same time. But still we observe that one becomes more popular than the others. Researchers say that the things are not only depend on the voice quality. Men wants features that he uses as in day to day life. Written words and symbols also need to be used. Abstract: Its morning of the last working day, Friday, and you’re thinking about gratifying a week’s worth of hard work by switching off the lights of the workplace for a lunch time in the sun. You consider that eating alone would not be fun, so you plan to call your friend to accompany you. So you take out your innovative iPhone smart and just press a single button. At the foot of the screen, a small microphone appears and says: ‘How can I help you? Then you speak out: “Send text to James: Let’s have lunch together.” Instantaneously, your proposed message becomes visible on the screen, laden into a shortened account of the cell phone’s standard messaging form. James’s name has occupied the necessary spot. And then off it departs. Have you ever experienced talking to your personal PC? Where it in reality recognized what you whispered and after that did something as a product? If YES, then youve worked with a technology known as to all as ‘speech recognition’. Inventing a machine that understands human manners, mainly the ability of speaking as you would expect and responding properly to spoken language, has made curiosity amongst engineers and scientists for centuries. Today speech technologies are available in the market for a restricted but interesting range of everyday jobs. These technologies enable equipments to respond in the approved manner and dependably to human voices, and offer useful and important services. Introduction: Vocal conversations are the most important of communications that men use in his social life. This account for the fact that “speech” is the most common and the most natural way communication in the world, and also on the World Wide Web. No matter from where the sound is being received, either it be TV broadcasts, our owned family videos, a program on the radio system, or even a simple day to day chat, the quality of our speech recognition software is that they can easily transform the human input into transcripts that can be worked on by computer systems or smart phones. In the last 7 decades, computers have astonished us with their capability to conduct simple and random calculations at amazing speeds. With every passing minute, the computer’s ability to practice these calculations has highly developed to a point where today, we are questioning it with our most urgent, difficult problem that are mapping worldwide climate change, atomic physics, and material engineering sciences and further at the atomic level and past. Stuff that would take hundreds of individual’s lifetimes to work out on their very own. Every part of these phenomena functions under a firm series of laws. Our aptitude to understand them is merely as sufficient as our ability to assemble a big enough statistics set to correspond to them. Aside from the infrequent once in a lifetime discovery that redefines those regulations. Such as when scientists exposed bacteria some year’s back that survived on arsenic, an element that is not amongst the six known elements that constitute life, still the rules remains constant. System speech recognition is the field of computer studies that contracts with designing computer based systems that can identify spoken words and manipulate them. Here, note that voice detection implies only that the PC can take dictation, not that it is aware of what is being whispered. To comprehend personal languages, falls under a special field of computer science called as ‘natural language processing’. A variety of voice recognition soft wares are available today in the market. The most powerful one of them can recognize limitless words. Though, they generally require a comprehensive guidance session during which the workstation system becomes accustomed to a singular voice. These systems are known as ‘speaker dependent’ systems. A lot of systems also want that the speaker speak gradually and distinctly and divide each word with a tiny pause. These are known as ‘discrete speech’ systems. In recent times, great steps have been made in ‘continuous speech’ systems that are voice recognition method that allow you to talk naturally. There are now a variety of continuous-speech systems available in the market for PC’s. Because of their boundaries and high prizes, voice recognition systems have conventionally been used only in a small number of specialized circumstances. For example, these systems are helpful in instances when the client is not capable to use a keyboard to enter data because his or her hands are busy or disabled. As an alternative to typing commands, the user can plainly speak hooked on a headset. However, as the cost is decreasing and performance is being enhanced, speech recognition systems are incoming in the mainstream and are being used as a substitute to keyboards. Problem Statement: Voice recognition software starts with a record of pre programmed sound prototypes. However, the actual client speech shows a discrepancy. A users pronunciation of a specified word can alter, the class of the microphone assembling the sound patterns may be deprived, and ambient sound can all change the sound pattern for a specific word. Voice recognition facility works finest only after the software has gathered data concerning each users verbal communication patterns. This way the software has an introductory wisdom curve. The software becomes most successful with time and primarily makes countless faults. The voice recognition software programmers and engineers have worked out method to give each word, phrase and sentence, standards of strength, volume, and distinction. But obviously we can see how rapidly that can set hurdle for it self. In fact it’s not an easy task. However, in doing so, voice recognition can as well get very close to recognize the sole speaker. This has functions as well for training the software for correctness amongst users, changing users without being educated and improving accuracy with unfamiliar users. Every year voice recognition system has improved verily- the future of ‘voice recognition’ is near. Either it is a small or large trading organization, a business telephone system has become an essential necessity. Relatively than an ordinary phone arrangement, today organizations insist a business telephone system with modern technology that helps to boost yield and improve client service. High end business telephone schemes are the outcome of the increased demand for systems with progressive technology. Voice over internet protocol (VoIP), and other voice processing and CPU telephony integration solutions are a few of the innovated technologies in business telephone system. Usually, these innovate technologies are greatly beneficial for groups that make extensive use of the phone. With the ability to carry integrated voice plus data applications, VoIP are measured the best choice to the analogue phone connection. Through this scientific development, employees can compose phone calls and launch faxes over IP-based data networks flanked by branch offices. But the most noteworthy advantage of VOIP is that it has reduced telephony expenses. Hence, this knowledge has become the most well-liked and widely used high-end business phone system in organizations internationally. This rapid technological advancement makes us ponder on two very fundamental questions. Will we eventually eliminate the need for human beings in common telephone communications? And either it’s good or a bad factor? Preferred or Relevant Business Information Systems Model:  Speech recognition systems are relevantly used in business and technological fields. Speech recognition is used to infer the callers answer to voice prompts. Software recently developed named as Interactive voice response or simply IVR have made the interaction between human voice and computer systems possible. The two main ranges of speech recognition are used in IVR (Interactive voice response): that stand upon well defined grammars, and that based on statistically qualified language models. Supplementary technologies comprise using Text-To-Speech (TTS) to speak intricate and dynamic information, like e-mails, news reports or weather statistics. TTS is computer produced manufactured speech that is now not the computerized voice that is traditionally linked with computers. Genuine voices craft the speech in fragments that are merged together before being acted in to the caller. An IVR can be positioned in following different ways: 1. Used by costumers as installed equipments at their workplaces 2. Equipments are involved at workplaces where large telephone networks are operated and there is a constant need of switching from one user to another. There are a number of ways IVR uses its calling system. In the field of telecommunications, an audio response unit (ARU) is a tool that provides synthesized voice responses to DTMF (Dual-tone multi-frequency signaling) key presses by handing out calls that have foundation on: (a) The caller’s originator input (b) Information acknowledged from a database (c) In order of the incoming call, like the time of day. Methodology, Best Practices, Successful model that works: The types of systems that are implemented in speech recognition system are often very sophisticated but still influential. The systems often make use of certain mathematical operations and the laws of “probability and combination” to find the best results. A speech group manager, Garofolo, declares that the two essentials that dominate this field today are either HMM or The Neural networking system. These two methods make use of complex mathematical functions and logical operations. Let’s have a closer look at HMM, which is defined as the most common of the method by the elite people using speech recognition systems. In HMM, every phone is assembled in a chain structure which eventually ends up into a single word. During this time, the system assigns values for every detail and considers every step by the mathematical methods. In this era of internet, speech recognition race has accelerated and the champs are those who collect the most data. With these things people like Watson, Siri, and Dragon are starting to see that some relations are going to occur in some linguistic manner. Nahamoo said: “The question is, what aspects of these technologies is going to become a keyboard, a commodity where it’s unclear where the business is, versus applications that exist that are the underpinning of business?” The risk is high. As globalization becomes the financial norm, the potential of technology to run and manipulate language is supreme. It’s not just in user electronics that large call center process continue to utilize speech recognition to process desires and time crunched doctors around the globe rely on it to say aloud notes over the phone for record. The big subject here is we are entering an age where more people understand that speech recognition is no longer this put in on that makes text. The Internet made of stuff, has always guaranteed a fully connected world, but only implied at what could be proficient with it. In the upcoming, you could be asking a robotic teller appliance for a balance, striking your smart home to lower the warmth by a couple of degrees or educate your kitchen’s coffee machine to formulate six cups on the brawny side. Even enhanced, your subsequent call to customer service may in fact detect how annoyed you are and react accordingly. Above all, the most inspiring feature of modern speech recognition is that classification will constantly and mechanically learn to advance itself. As smart phones propagate around the world, linked TV boxes make their approach into the living room and smart cars hit the streets, companies like Microsoft, Google and Nuance will observe their datasets cultivate and grow with innovative sources of speech, from a Midwestern mother shouting over her kids at the mall to a Bangladeshi cab driver transferring text messages to associates as he pace around New York City. Conclusion: The world that we are living now is advancing every second. It’s just a matter of time when speech recognition systems implied in smart phones and else where, will prevail the world. Every individual will be benefited by this technological outcome and actions will soon be reduced to commands. This will be the new world as I see it. References: Acero, A. (1993). Acoustical and environmental robustness in automatic speech recognition. Boston: Kluwer Academic Publishers. Nah, Y., Obermaisser, R., Puschner, P., Rammig, F. J., IFIP WG 10.2 International Workshop Software Technologies for Embedded and Ubiquitous Systems, International Workshop Software Technologies for Embedded and Ubiquitous Systems, & SEUS. (2007). Software technologies for embedded and ubiquitous systems: 5th IFIP WG 10.2 international workshop, SEUS 2007, Santorini Island, Greece, May 7-8, 2007 ; revised papers. Berlin [u.a.: Springer. Burduk, R., Kurzynski, M., Wożniak, M., & Zolnierek, A. (2011). Computer Recognition Systems 4. Berlin, Heidelberg: Springer Berlin Heidelberg. Tan, Z.-H., & Lindberg, B. (2008). Automatic speech recognition on mobile devices and over communication networks. London: Springer. Kotelly, B. (2003). The art and business of speech recognition: Creating the noble voice. Boston, MA: Addison-Wesley. Read More