This chapter explains significant speech and translation technologies for healthcare professionals. We first examine the progress of automatic speech recognition (ASR) and text-to-speech (TTS). Turning to machine translation (MT), we briefly cover fixed-phrase-based translation systems (“phraselators”), with consideration of their advantages and disadvantages. The major types of full (wide-ranging, relatively unrestricted) MT – symbolic, statistical, and neural – are then explained in some detail. As an optional bonus, we provide an extended explanation of transformer-based neural translation. We postpone for a separate chapter discussion of practical applications in healthcare contexts of speech and translation technologies.

Cross-language communication in healthcare is urgently needed. However, while development of the relevant linguistic technologies and related infrastructure has been accelerating, widespread adoption of translation- and speech-enabled systems remains slow. This chapter examines obstacles to adoption and directions for overcoming them, with emphasis on reliability and customization issues; discusses two major types of speech translation systems and their respective approaches to the same obstacles; surveys some healthcare-oriented communication systems, past and future; and concludes with an optimistic forecast for speech and translation applications in healthcare, tempered by due cautions.

This book offers state-of-the-art research on the design and evaluation of assistive translation tools and systems to facilitate cross-cultural and cross-lingual communications in health and medical settings. This book illustrates using case studies important principles of designing assistive health communication tools which are (1) detectability of errors to boost user confidence by health professionals; (2) adaptability or customizability for health and medical domains; (3) inclusivity of translation modalities (written, speech, sign language) to serve people with disabilities; and (4) equality of accessibility standards for localised multilingual websites of health contents. To summarize these key principles for promotion of accessible and reliable translation technology, we use the acronym I-D-E-A.

Automated speech translation, long a dream, has come into widespread use, as enterprises, application developers, and government agencies have become aware. Real-world S2ST applications have been tested locally over the past decade in consumer, healthcare, military, and humanitarian missions, and several projects aim to enable automatic cross-language communications at the 2020 Olympic Games to be held in Tokyo. Accordingly, this chapter provides a survey of the field’s technologies, approaches, companies, projects, and target use cases. (It is based on an industry report sponsored by the Translation Automation Users Society, released in 2017.) Sections examine the Past, Present, and Future of speech-to-speech translation. The first provides an orientation concerning issues in speech translation and a capsule history; the second snapshots technical achievements and representative participants in the burgeoning current scene; and the third speculates about future directions, with emphasis on platforms and form factors, big data, knowledge source integration, and the roles of human and automatic translators.

John Searle and other influential theorists have argued that machine translation (MT) and other natural language processing (NLP) programs can never appreciate meaning in the deepest sense – in other words, that they can never truly exhibit semantics. It is true that MT and many other NLP systems have made steady and impressive progress while use of explicit semantic processing has undergone a rise and fall. It is also true that consensus among researchers on the meaning of meaning has remained elusive. In this chapter, however, we observe renewed interest in semantic representation and processing. Moreover, we foresee gradual adoption of semantic approaches grounded upon audio, visual, or other sensor-based input. We distinguish such perceptually-grounded semantic approaches from most current methods, which have tended to remain perception-free. With respect to philosophical implications, we suggest that perceptually-grounded approaches to automatic natural language processing can display intentionality, and thus foster a truly meaningful semantics. As background for these predictions and suggestions, we survey the role of semantics in machine translation to date in terms of three paradigms: rule-based, statistical, and neural MT. A section on each paradigm discusses its treatment of semantics: rule-based methods have generally emphasized symbolic semantics; statistical methods have generally avoided semantic treatment or employed vector-based semantics; and neural methods have handled meaning as implicit within networks.