AI Supporting AAC Pictographic Symbol Adaptations.

The phenomenal increase in technological capabilities that allow the design and training of systems to cope with the complexities of natural language and visual representation in order to develop other formats is remarkable. It has made it possible to make use of image to image and text to image technologies to support those with disabilities in ways not previously explored. It has opened the world of adaptations from one picture to another in a design style of a user's choosing. Automated text simplification alongside graphical symbol representations to enhance understanding of complex content is already being used to support those with cognitive impairments and learning difficulties. Symbol sets have become embedded within applications as dictionaries and look up systems, but the need for flexibility and personalization remains a challenge. Most pictographic symbols are created over time within the bounds of a certain style and schema for particular groups such as those who use augmentative and alternative forms of communication (AAC). By using generative artificial intelligence, it is proposed that symbols could be produced based on the style of those already used by an individual or adapted to suit different requirements within local contexts, cultures and communities. This paper explores these ideas at the start of a small six-month pilot study to adapt a number of open licensed symbols based on the symbol set's original style. Once a collection has been automatically developed from image to image and text descriptions, potential stakeholders will evaluate the outcomes using an online voting system. Successful symbols will be made available and could potentially be added to the original symbol set offering a flexible personalized approach to AAC symbol generation hitherto not experienced by users.

Exploring Practical Metrics to Support Automatic Speech Recognition Evaluations.

Recent studies into the evaluation of automatic speech recognition for its quality of output in the form of text have shown that using word error rate to see how many mistakes exist in English does not necessarily help the developer of automatic transcriptions or captions. Confidence levels as to the type of errors being made remain low because mistranslations from speech to text are not always captured with a note that details the reason for the error. There have been situations in higher education where students requiring captions and transcriptions have found that some academic lecture results are littered with word errors which means that comprehension levels drop and those with cognitive, physical and sensory disabilities are particularly affected. Despite the incredible improvements in general understanding of conversational automatic speech recognition, academic situations tend to include numerous domain specific terms and the lecturers may be non-native speakers, coping with recording technology in noisy situations. This paper aims to discuss the way additional metrics are used to capture issues and feedback into the machine learning process to enable enhanced quality of output and more inclusive practices for those using virtual conferencing systems. The process goes beyond what is expressed and examines paralinguistic aspects such as timing, intonation, voice quality and speech understanding.

Co(0.85)Se/Multi-Walled Carbon Nanotube Composite as Alternative Cathode Catalyst for Microbial Fuel Cells.

The development of low-cost and high-performance cathode catalyst remains a great challenge for constructing large-scale microbial fuel cells (MFCs). Here, the multi-walled carbon nanotube (MWCNT) supported cobalt selenide (Co0.85Se) composite was synthesized by a simple and effective hydrothermal process for the first time. Compared with uncomposited electrodes such as MWCNTs and Co0.85Se, the as-prepared Co0.85Se/MWCNT composite electrode exhibited significantly improved electrochemical performance and durability. The maximum power density of Co0.85Se/MWCNT composite electrode is 243.6 mW m­2, which is 12.2 fold higher than that of unmodified electrode and more than twice as much as those observed for MWCNTs cathode (114.1 mW m­2) and Co0.85Se cathode (119.8 mW m­2). This work may provide not only the fundamental studies on carbon supported transition-metal selenide but also a new kind of promising alternative electrode in the technology of power generation from MFCs.