At one point or another, every one of us has been faced with difficulties while trying to search for information. Whether you are using the Internet, a file system on your computer's hard drive, database, or a global information system of your company, searching for information can be quite a tedious task.
The problems that you will encounter are numerous and even though you are searching with the aid of keywords, it's all to no avail. It is times like this where modern technology can be considered as a double edge sword. How so? Well, since the memory of a single PC is virtually equivalent to the amount of text data that is stored in a real land base library, searching for a certain piece of information is like searching for a needle amidst a bunch of hay. As for the Internet, the problem is much worse. That is why it is quite necessary to create search systems and technologies that would simplify as well as accelerate access to the necessary information.
There are various search technologies already available and selecting one that is suitable would depend directly on the specific tasks that you would need to solve in the future. So just imagine if you are caught in a deadline, desperate for that one data and at the same time your hands are tied to the phone or on your computer? Do you have to keep on halting your searches simply because you cannot type those keywords? Not necessarily so.
Have you heard of voice-activated Internet surfing and searching? Believe it or not that technology is now well within our grasp. As a matter of fact, Google had recently made announcements regarding their plans to offer free software online that would allow users to search the Internet with the use of voice-activated tools. Pretty cool is it not?
Indeed, but not quite. Those who are familiar with voice-activated software that is currently available either for business correspondence or in writing schoolwork is probably aware of at least one drawback in this type of technology. The drawback is actually minor and can easily be solved. You see voice activated software requires special microphones, or else the background noises and vibrations will make it difficult for your voice-activated software to comprehend your requests. Simply put, you will not be able to search around the Internet this way without it.
Another thing, the environment is also significant. If you want your voice interface to work properly, then your environment must be peaceful enough. If you are near a railroad or an airport, forget it. The noise and vibrations would simply mess up the words that you spoke into the microphone and would only confuse the voice-activated software. But then again, those are just minor setbacks and can easily be conquered with the right tools and location. Do not let it keep you from experiencing the next big thing in the Internet - voice interface for search engines.
Just what is a voice interface for search engines? Well, this is one of the most promising search technologies. According to its patent, voice interface for search engines is a system that would provide search results from a voice search query. How does it work? The theory is simple enough. Naturally, the system would receive a voice search query from a user, then it will derive one or more recognition hypotheses. Each one is associated with a weight, from the voice search query, and constructs a weighted boolean query using the recognition hypotheses. Then the system will provide the weighted boolean query to a search system and then provides the results of the search system to a user.
The voice interface is would be beneficial to the information retrieval systems and, more specifically, to a system and method for supporting voice queries in information retrieval systems. The main objective of this technology is to help those people in their searches. The voice interface to a search engine will allow them to perform searches without the need to type words, since it recognizes spoken queries and are guaranteed to return highly relevant search results.
There are several problems that exist in designing satisfactory voice interfaces. One of which is that current speech recognition technology has high word error rates, particularly for large vocabulary sizes. Meaning a change in intonation may confuse the command.
Other speech recognition applications a recognizer can use context, like a dialogue history in order to create certain expectations and guide the recognition. This is another problem. Voice search queries do not have such context. Voice queries tend to be very short and are on the order of only a few words or single word. Thus, there is little information from the voice itself in which to make a voice recognition determination.
Fortunately, the current patent application associated in voice interfaces for search engines addresses these problems. According to the said patent, by limiting the scope of the voice queries to a very narrow range will solve those problems. How would that work? Well, the user will be prompted to choose a small number of choices each time. For instance, at the initial menu, the user may choose from "news," "stocks," "weather," or "sports." After choosing one category, the system will again offer another small set of choices until the user reach the piece of information that he or she is searching for.
By limiting the number of possible speeches, the problem in speech recognition task would decrease to a level that is satisfactory enough. However, it is still likely that this approach might result in severe deficiencies. For instance, since the user still needs to go through several voice menus, it might be slow to use. Furthermore, if the user's information need do not match any of the category, then it would be quite difficult to find the information that he or she desires. In the end it might cause some frustration on the part of the user, especially since he or she must adjust his or her interactions to the rigid, mechanical structure of the system. Therefore, a more effective voice interface for search engines is still at large.
Technology 
