AWS re:Invent 2017

It used to be the case that we only spoke to computers in their language. But more and more often, we're interacting with them in ours. We are moving quickly into a world of computer conversation, and one in which, for many applications, the most natural interactions will be through spoken language. But how do you create engaging narrative and compelling, organic conversational interactions using the imprecise tools of speech recognition and intent resolution?  In this session, we look at the experience as a whole and take you through key learnings that you can use when building your skills. We cover issues like knowing your audience, creating compelling storylines, using a cast of characters, integrating voiceover, designing a soundscape, and finding those “magic moments”. For each of these, we share the design pattern, the backing AI or physiological science, and how to implement the experience with Alexa.

Last year, Capital One joined Alexa on stage to talk about their experience building their successful Alexa skill. Since that time, many lessons have been learned through customer feedback and new enhancements to the Alexa Skills Kit (ASK) such as the skills beta testing tool and the Alexa skill builder. How can you evolve your Alexa skill with more meaningful data sets outside of the existing intents? As the Alexa Skills Kit has grown its built-in library, what does it mean for your skill to support both ordinal (list) and numerical values? How can you handle new specifications without requiring wholesale code changes? Capital One has tackled all of these issues as well as embracing additional programming languages like TypeScript to ensure that response structures are validated against all schemas. With the arrival of multimodal devices such as the Echo Show, the opportunity for seamless customer interaction models across voice and visual has also arrived (big fonts, touch, video). Your customers can now transition back and forth between using their voice and their hands while engaging with your skill. Come learn direct from Capital One on the best way of providing extra contextual information using the new Alexa Skills Kit display directives but in more convenient ways to get things done.

In this advanced session, learn how to build Alexa-enabled devices that combine voice and visual responses in a meaningful way for consumers. The session covers the design methods and the hardware and software development resources for interactive multi-modal design. We also present some examples of products that are leading with such implementations.

Garbage in, garbage out. The quality of all machine learning solutions depends on the data used in training. Alexa developers are able to use advanced natural language understanding capabilities like built-in slot and intent training, entity resolution, and dialog management. This utterance data behind your skills is the most important contributor to the voice input experience. This session discusses how utterance data is processed by our systems, and what you can do as a developer to improve accuracy.

In this session, scientists from the Alexa team explore and discuss some of the AI challenges behind the Alexa Prize. Learn about the challenges of Automatic Speech Recognition (ASR), Natural Language Understanding (NLU) and conversational interaction through stories from the founding members of the team that also built Amazon Echo and Alexa. We'll address the early difficulties of designing the algorithms for noise reduction for close-talk, near field, and far-field Alexa devices, and methods and frameworks they use for ASR, NLU and conversational interaction.

Join us for the Golden Age of AI. The way that humans interact with machines is at an inflection point and conversational artificial intelligence (AI) is at the center of the transformation. Learn how Amazon is using machine learning and cloud computing to help fuel innovation in AI, making Alexa smarter every day. Alexa VP and Head Scientist Rohit Prasad presents the state of the science behind Amazon Alexa. He addresses advances in spoken language understanding and machine learning in Alexa, and shares how Amazon thinks about building the next generation of user experiences. He will announce the inaugural winner of the Alexa Prize and award the winning student team a check for $500,000.

Your Alexa skill could become the voice of your company to customers. How do you make sure that it conveys rich information, delivered with your brand's personality? In this session, Adam Long, VP of Product Management at Automated Insights, discusses natural language generation (NLG) techniques and how to make your Alexa response more insightful and engaging. Rob McCauley, Solutions Architect with Amazon Alexa, shows you how to put those techniques into action.

Join Alexa SVP Tom Taylor as we cover the state of the Alexa business, describe some early challenges, and share how we are approaching emerging trends. Voice experiences have transformed the way that customers interact with the world around them. We will introduce new capabilities to help developers better address opportunities in devices, the smart home, and voice. You will leave with an understanding of the vision behind Alexa that ties together the deep dives going on throughout re:Invent.

In this session, we will give you a complete picture of all the tools and techniques required to build complex, production-quality Alexa skills. You will leave this session knowing how to use Alexa's dialog management, entity resolution, and slot elicitation capabilities as well as how to process the results through a microservice with AWS Lambda.

Manufacturing companies collect vast troves of process data for tracking purposes. Using this data with advanced analytics can optimize operations, saving time and money. In this session, we explore the latest analytics capabilities to support your goals for optimizing the manufacturing plant floor. Learn how to build dashboards that connect to prediction models driven by sensors across manufacturing processes. Learn how to build a data lake on AWS, using services and techniques such as AWS CloudFormation, Amazon EC2, Amazon S3, AWS Identity and Access Management, and AWS Lambda. We also review a reference architecture that supports data ingestion, event rules, analytics, and the use of machine learning for manufacturing analytics.