|
 |
 |
| ||||
|
Technology Area | Market Needs | State-of-the-Practice | Current Technology | Issues to Consider | References Technology AreaCommunication Processing (CP) includes a person or device's ability to incorporate all sources of available information in order to communicate. Improving AAC-based Natural Language Processing (NLP) capabilities will impact important aspects of the augmented communicators life including daily living, education, employment, and societal inclusion. Advances in processing capabilities will enable clinicians to more effectively evaluate and monitor performance, modify performance parameters, reference recognized communication models and optimize and modify device performance in order to enhance communication and literacy. The availability of contextual information (e.g. interlocutor speech, location, time, own speech) would enable AAC devices to incorporate more advanced information processing capabilities available for current computer systems. [ Top of Page ] Market NeedsCommunication Processes can be summarized as the ability of an individual to process input from varying sources in the environment simultaneously and translate that input at various levels in order to produce communicative interaction. Devices have been developed to interact with and facilitate some of the systems that humans use in communication. Natural language processing, for the purpose of this white paper, is the ability of a device to decode, integrate, process, and encode all sources of information/stimuli (visual, auditory, tactile, etc.) for the purpose of communication using an established set of symbols (pictures, letters, words), structure, and meaning and communicative function. Overall improved processing capabilities may be the result of advances in NLP and/or other computational techniques. In addition, it is also important to consider the Human Computer Interaction (HCI) and the relation between communication processes and natural language processes. HCI is significant when considering device and human interaction for two main reasons 1) in terms of the way in which humans employ machines to complete tasks and 2) the ability of machines to amplify human ability. Individuals using AAC are already unique with respect to their limited verbal, vocal, and gestural abilities to conduct social interactions. [1] Communication processing is associated with two main roles: 1) communication and 2) literacy. Expressive communication is the ability to efficiently and clearly formulate, encode and convey ideas, feelings and attitudes to other individuals through interaction. Critical aspects of communication include content (what is said), form (message structure), function (communication purpose) and delivery (speed and manner of performance). Communication is imperative in all daily living, educational, work, and recreational activities. Literacy competency includes the ability to build complex, complete, logical, detailed, and grammatically correct verbal and written dialogue for complex use of language (e.g. poetry, technical writing, thesis, and dissertations, etc.). It is also the capability of individuals to comprehend and express oneself through the media of the culture (e.g. text, internet, video). Augmented communicators have identified factors such as improved communication rate, literacy, ease of use, and ease of optimization as high priority needs. This type of technological improvement can help address these needs and can also have a major impact in areas such as increased educational and employment opportunities, independence, and inclusion in social interactions. An advanced processing system incorporated into AAC devices would also enable the user to gain literacy skills, which could increase their ability to participate in a variety of interactions. AAC devices with improved processing capabilities would also allow individuals to communicate with more naturalness and efficiency. Proper implementation of communication models could promote both literacy and interactive communication. Clinicians have identified a need to link AAC-based processing capabilities with a clinical model of language acquisition and processing. The acquisition process, however, is only one-dimensional in that it deals with vocabulary, mapping of meaning and language structure onto the device. There is a whole other realm, which has to do with using the device for communication purposes. Few perspectives exist for that, but several relevant models of human communication and information processing could provide a guide to device designers for making better devices. Models that address certain areas of communication such as language acquisition, language processing, language representation and Human Computer Interaction are currently available but not fully implemented into device design. [2,3,4,5,6,7] Clinicians are responsible for selecting AAC devices and then modifying (i.e. programming) device parameters to fit the needs, capabilities and applications of the user. The clinician seeks to efficiently track and evaluate a user's communication and literacy skills while employing the AAC device. Research has emerged to demonstrate that children with severe motor and speech disorders learn language differently than normally developing children, wherein, their lack of speech puts them at a disadvantage for learning language. [8] Conclusions drawn from this research may also apply to individuals with mental retardation or autism, persons reacquiring language and individuals with cognitive deficits. AAC-based processing capabilities that facilitate language acquisition would be especially useful. Improved service delivery is a further benefit of the enhancement of communication processing in devices because it allows the clinician to train and modify the device with greater ease and efficiency. This delivery, however, is sometimes limited by appropriate funding. HMO's, Medicare, and Medicaid offer the subscriber only a limited number of rehabilitation services. Two things currently are happening: 1) the patient receives the prescribed number of treatment sessions, is dismissed and receives no further treatment or 2) patient finishes prescribed number of treatment sessions and has to independently fund rehabilitation services. In both scenarios patients are not receiving the full training needed to use the device. A clinician cannot realistically optimize device performance and the user is forced to assume this role. Currently AAC devices can be difficult to program and understand which can take up a significant amount of the clinician and user's time for treatment and training. Contextual information could be used by devices to optimize communication performance (e.g. efficiency, clarity). For example, time and location context might be provided by the Global Positioning System. The topical context (subject of discussion) might be provided by recognizing the augmented communicator's speech history or the words uttered by their communication partner (interlocutor speech recognition). AAC devices could adapt in response to this contextual information in order to optimize communication rate (e.g. adapt word prediction list to context) and clarity (i.e. the best word gets selected rather than a sufficient word). [ Top of Page ] State-of-the-PracticeThere are many models of language development and processing that include all aspects of language such as phonology (structure and sequence of speech sounds), morphology (rules that govern meaning at the intra-word level, e.g. bat vs. bats), syntax (grammatical rules for the combination of words), semantics (meaning of words and phrases), and pragmatics (use of language) [2,9]. Referring to these models, researchers and clinicians seek to establish developmental stages and milestones used as markers for both assessment and treatment in communication disorders. A clear link between AAC device parameters and capabilities and an accepted model of communication and literacy would enable the clinician to efficiently and systematically optimize AAC device performance. In addition, AAC device users are at risk for passive communicative interactions. Researchers have shown that communicative attempts made by children who use AAC can be either ignored or redirected by the adults with whom they interact. Furthermore, children have a difficult time in obtaining rewarding responses to their communicative attempts and their opportunities for communication are severely limited. The device may also limit their output capabilities by using structure and grammar that is different than the norm. Children who use AAC devices with a symbolic representation system fail to achieve appropriate levels of literacy needed in an educational setting. Communication Processing can provide this population with a means by which they can develop language that closely resembles that of a normally developing child. AAC devices already employ natural language processes. Word prediction systems, for example, have been developed to learn the user's word selections so that after a sequence of letters are selected; the words most commonly chosen by the user will be presented. Word prediction software systems include word processor programs with predicative word capabilities, word prediction programs that helps the user create grammatically correct sentences, and systems that can predict words for the user during text entry by learning the user's frequently typed words, etc. Other technology developers are working towards the improvement of word prediction in AAC devices. Research has focused on analyzing the way in which non-impaired individuals predict words, because it is shown that humans are more efficient and accurate when predicting appropriate words than computers. By incorporating natural word prediction skills into AAC devices it increases the rate and accuracy of which the device predicts words for the user. This new technology corresponds with the ideas of natural language processing. AAC systems frequently employ graphical interfaces (text or other graphical symbol (icons, photographs) and nonsymbolic materials (color, shape)) to navigate through their device and to retrieve words and messages. Advances in communication processing could be used to facilitate the grammatical utterance construction or to speed up the retrieval of whole utterances. GPS systems are being used worldwide by military, government, and even more recently by AAC users to determine the location of an individual or device. Two types of GPS systems can be used: 1) Standard Positioning Service (SPS) and 2) Precise Positioning Service (PPS). SPS allows for predictable positioning accuracy of approximately 100 meters with a time transfer accuracy within 340 nanoseconds. PPS, which is not available to the general public, is able to predict positioning within at least 22 meters and can time transfer within 200 nanoseconds. [10] Researchers are continuing to refine and improve the GPS technology to provide more accurate and precise information. GPS input would allow AAC devices to process and use environmental information to provide context to their interactions. GPS hand-held systems are available on the market to all consumers, and are used for recreational activities and travel. An example of a GPS system that is currently being used in an assistive device is Atlas Speaks (talking map) and Strider (GPS access system). This system is a voice output electronic mobility aid for individuals with visual impairments. It incorporates a GPS system receiver that lets users learn about the physical layout of a neighborhood, city, or state and navigate from location to location. [11] Interlocutor speech recognition allows an AAC device to recognize input from individuals in the environment and use the information to process appropriate responses that relate both contextually and grammatically. In addition, systems are being developed that can be used in noisy environments, which have been addressed as specific problem areas for augmented communicators. Currently technology is being developed that can recognize the context of an interlocutor's speech and use that context, vocabulary, and grammar to create a series of appropriate responses and questions that can be used by the augmented communicator. The implementation and use of a uniform assessment tool, which provides measurable data for clinicians to use to determine therapeutic effectiveness of AAC devices, is a growing trend in the field of AAC. Traditionally, SLP's have tracked outcome measurements using pen, paper and video, which can be cumbersome and expensive. Recent advances in performance assessment and remote data logging permit the researcher, developer and clinician to collect detailed logs of user-device performance. With the development of analysis software and validation of the clinical measures, and the adoption of a standard logfile format by manufacturers, advances could be made in the objective analysis of communication and device performance. [12] [ Top of Page ] Issues to ConsiderThe Need
State-of-the-Practice
Future Technology and Products
[ Top of Page ] References
[ Top of Page ] |
|
||
|
| |
