In the Heart of Silicon Valley
Talking to Computers: Moving into the Mainstream
Conversing with computers using human language—speech or text—is now a technology that all companies must incorporate in their business activities, according to the 30-year-old not-for-profit industry organization Applied Voice Input Output Society (AVIOS). Marie Meteer, President, AVIOS, summarized, “No company can be without a web site today. Automating natural conversations with customers is quickly becoming the next ‘must have’.”
Evolution Continues: Will computers and humans be inseparable?
Judy cries in her crib not long after birth. A device called a “baby tender” on her crib recognizes the cry. It begins playing soothing music. It signals the mother’s smartphone. She acknowledges the alert but must finish a task before coming. She speaks through the device to the baby: “I’m coming, sweetheart. Be there soon.” The baby calms.
Speech Understanding: The Next Leap in Speech Recognition
Deep learning and the data that feeds it will drive speech tech’s commercial use
Early attempts to commercialize speech recognition technology were limited by the state of that technology. One typically had to find an application where hands- or eyes-free control of a system was required, and the actions controlled were restricted to a small vocabulary. One early application that found limited success directed a headset-wearing worker to a particular bin to pick up an order, with the worker saying things like “out of stock” or “next.”
Conversational Interaction creates conversation!
I thought the conference was very well organized and the content was extremely informative. I liked the balance between larger research organizations (e.g. IBM, google, Amazon, Intel) and small companies. Many good examples of chatbots and conversational UI already in use. Good coverage of a wide range of applications, from social-public service to commercial, agricultural, and so on. Several enterprise-type chatbots for us from Oracle to consider. Great perspectives from Bill and the various NLU and technical people on what works, what the trends are. Very interesting additional information on internationalization, hearing of the digital assistants, and other factors that can make or break your implementation. I went to about 30 presentations over the two days. It's very convenient to have the meals right there so you don't waste time going out. I met a number of people I will be following up with.
What is a conversation?
Conversation: A major trend in the evolution of the user interface
Arguably, the Graphical User Interface (GUI) drove the growth of personal computers by making their use more intuitive. Today, the basics of the WIMP interface (Windows, Icons, Menu, and Pointing device) have transitioned beyond the PC to many other devices and channels—smartphones and web sites being prime examples.
But the GUI is getting overburdened and less intuitive as it supports an increasing number of features and devices, with problems such as complex and long menus and the need to navigate many pages—each with too many choices—to accomplish a task. The problem is compounded when the interaction is on the small screens of mobile devices and the pointing device is a finger.
The goal of the Conversational User Interface (CUI) is to transform this complexity into simply “say or type what you want.” Conversational interaction can be essentially the same irrespective of device or application, creating a uniform experience across devices. The conversation is sometimes conducted through a “digital assistant” or a “virtual agent” that may even exhibit an engaging personality. Home speakers and some automotive systems have shown that flexible interaction with digital systems by voice is possible even when a screen isn’t used.