Tag Archives: DevOps

This past September, 10 university teams, from schools around the world, officially began their journey with the Alexa Prize Socialbot Grand Challenge 3. Teams spent part of their summer attending an intensive boot camp to learn what it means to be customer obsessed and how to integrate their ideas into the Alexa Prize infrastructure. Now ready to compete, teams have built an AI (called a socialbot), at scale, capable of conversing coherently and engagingly with humans for 20 minutes. Their socialbots should be able to discuss popular topics and news events such as entertainment, fashion, politics, sports, and technology.

Even as innovators create AI that can drive cars, play board games, and diagnose diseases, the seemingly intuitive task of everyday social conversation remains an elusive goal. Undaunted, the student teams have geared up to tackle the challenge. As they build their socialbots, they will encounter esoteric problems like context modeling and dialog planning as well as exoteric problems like user experience and conversational engagement. The Alexa Prize provides a unique opportunity for industry and academia to come together to address these challenges.

We invite you to join the students on their journey and help them along the way. You can interact with their socialbots simply by saying, “Alexa, let’s chat” on any Alexa-enabled device in the US-English marketplace. You will be connected to one of the 10 socialbots at random. Chat about your favorite topics, say “Stop” when you are done, and share your feedback.

It is still early days in this pioneering journey, and we know the teams will appreciate your help and patience as they learn, iterate, and build new interactive conversational experiences. The socialbots will improve quickly based on your use and feedback. We hope you’ll enjoy interacting with them.

The student teams are working hard, having fun, and well on their way to making history. We look forward to having you join the Alexa Prize community to support them along the way.

Follow the #AlexaPrize hashtag and bookmark our site for updates.

Amazon.FallbackIntentが日本、フランス、イタリア、スペインで使用可能となりました。さらに、AMAZON.PhoneNumber（ベータ）とAMAZON.Ordinal（ベータ）のビルトインスロットタイプも合わせて使用可能となりました。

Alexaスキルの開発者であれば、スキルのインテントに幅広いバリエーションのサンプル発話を含めることで、発話のハンドリング精度が向上することはご存知だと思います。（さらに詳しい情報については、こちらのドキュメントをご参照ください)

しかし、インタラクションモデルがどれほど堅牢であっても、スキルのインテントとまったく一致しない発話を受け取ることがあります。 これらの種類のリクエスト（ドメイン外リクエスト）が意図していないインテントにマッピングされると、予期しない応答が発生する可能性があります。このような問題を解決するために、Alexa Skills Kit は昨年、AMAZON.FallbackIntent を英語ロケールとドイツ語のビルトインライブラリーに追加しました。

スキルのインタラクションモデルに FallbackIntent を追加すると、スキルの開発者とユーザーの両方にメリットをもたらします。 スキルユーザーに対しては、FallbackIntent がドメイン外リクエストを適切にルーティングすることで、より関連性の高い応答が提供されるようになります。 通常これは、ユーザーのリクエストがスキルでは対応できないものであることをユーザーに知らせ、スキルの機能に戻るためのガイダンスを提供するプロンプトとなります。

スキル開発者の場合、インテントリクエスト履歴 APIを使用することで、FallbackIntent カテゴリ内に表示されるドメイン外リクエストを分析することができます。これにより、スキルに期待されている発話や機能を判断し、必要に応じてドメイン外モデルを微調整することが可能となります。

以下は、AMAZON.FallbackIntentがどのように機能するかを示す例です。

ユーザー： アレクサ、トリッププランを開いて
Alexa： 旅行を計画しましょう。目的地はどこですか？

ユーザー： 札幌の天気を教えて （この発話は、トリッププランのどのインテントとも一致しません）
Alexaはこの発話をインテントにマッピングできる確率が低いと判断したため、スキルにAMAZON.FallbackIntentを送信します。

Alexa： トリッププランスキルではその質問にはお答えできませんが、フライトを調べたり、旅行の計画の手伝いをしたりすることはできます。どこに行きますか？ （応答にスキルの名前を含めて、ユーザーを正しくガイドします）

AMAZON.FallbackIntentはどのように機能しますか？

AMAZON.FallbackIntentがスキルに追加されると、Alexaの機械学習を使用してスキルの既存のインテントとサンプル発話を自動的に評価し、そのスキルに固有のドメイン外モデルを作成します。 AMAZON.FallbackIntentは、サンプル発話を追加し、拡張することもできます。これにより、開発者は必要に応じてドメイン外モデルを微調整できます。 Alexaがある発話をカスタムインテントのいずれかにマッピングできる確率が高いと判断すると、そのインテントが選択されます。一方、発話がスキルのドメイン外（スキルでは意味をなさない）である確率が高いと判断すると、AlexaはAMAZON.FallbackIntentを選択します。すべてのインテントに対して信頼度が非常に低い場合（音声が不明瞭な場合や雑音が入った場合など）は、かわりに再プロンプトが出される可能性があります。その後、スキルバックエンドは AMAZON.FallbackIntent リクエストを受け取り、適切な指示やガイダンスで応答することができます。

AMAZON.FallbackIntent のドメイン外モデルは、スキルのモデルと、既存のサンプルデータのテンプレートから自動的に生成されます。 ドメイン外モデルはモデルのビルドごとに再生成され、他のインテントへの変更を自動的に組み込みます。（インタラクションモデルの改善に応じて自動的に更新されます。）

AMAZON.FallbackIntentをスキルに追加する方法

Alexa 開発者コンソール上でAMAZON.FallbackIntentを追加し、インタラクションモデルを更新します。

その後、対応するIntent handler をスキルバックエンドに追加します。

他のビルトインインテントと同様に、バックエンドにハンドラを実装する必要があります。 通常これは、ユーザーのリクエストがスキルでは対応できないものであることをユーザーに知らせ、そのスキルの機能に戻るためのガイダンスを提供するプロンプトとなります。また、プロンプトでは、自分がどのスキルと対話しているのかをユーザーが再度認識できるようにする必要があります。以下は ASK SDK for Node.jsを使用したサンプルです。

さらに詳しい情報については、こちらのテクニカルドキュメントを参照してください。

AMAZON.PhoneNumber (ベータ) AMAZON.Ordinal (ベータ) が利用可能になりました

ビルトインスロットタイプを使用して電話番号と序数値をキャプチャする機能が、すべてのロケールに拡張されました。

AMAZON.PhoneNumber (ベータ) は、電話番号を表す数字または単語を句読点なしの文字列形式に変換し、国際電話番号、国番号、および地域の電話番号形式をサポートします。

AMAZON.Ordinal (ベータ) は、序数を表す単語を数字に変換します (例：「三番目」から 3)。音声による要素の選択など、複数のユースケースで役立ちます。 これにより、ユーザーはデバイス画面上のビジュアルコンポーネントをよりシームレスに操作できます。 たとえば、Ordinalインテントは、リスト選択をサポートすることができます（例：「二番目のアイテムを選択する」、「四番目のアイテムを選択する」）

関連コンテンツ

• AMAZON.FallbackIntent テクニカルドキュメント
• AMAZON.FallbackIntentとインテントリクエスト履歴
• スロットタイプリファレンス
• Alexa Skills Kit (ASK)
• Alexa 開発者フォーラム

Kixi is a German-language video on demand app designed to offer educationally valuable content for children and families.

We talked to Sascha Stradtmann, head of acquisitions and programming at Kixi, to learn more about Kixi’s service and their reasons for publishing it on Amazon Fire tablets and Amazon Fire TV.

Please tell us more about Kixi and your service.

“Kixi was originally founded under the brand-name Kinderkino in 2011 at a time when video on demand (VoD) was far from mainstream in Germany. Ever since then, Kixi started offering educationally valuable movies, series, cartoons, instructive films and documentaries for children and families. Our licensors are small independent production companies as well as major studios. We offer our partners a turn-key solution for the publishing of their content on our platforms. Kixi is a convenient way for our partners to monetize their content in an economic and efficient way. We use the newest state-of-the-art technology for encoding and content management, so we are able to achieve a lot with a surprisingly small team.

“In addition to the Kixi apps for Fire TV and Fire tablet, we also have an app for our owned and operated VoD platform.”

Why did you choose to publish your app Kixi on the Amazon Appstore and for Fire tablets and Fire TV?

“Amazon offers customers various ways to watch their favorite content – whether at home or on the go. So for us, it’s been a no-brainer to develop for the Amazon Appstore.”

Why did you choose to implement the in-app purchasing API?

“By implementing the in-app purchasing API, we did not have to code a payment service or back-end by ourselves. Instead, we were able to use a reliable ready-made solution.”

How do Amazon customers fit into your target group? What is the value of Amazon customers compared to other audiences?

“With the Amazon Appstore, we reach a wide set of customers and users can choose from a rich app catalog, including a great selection for family and kids. Therefore, the Amazon Appstore is the perfect place for us to engage with our target audience.

How do you optimize the app and content for families?

“We optimize our content strategy two ways: First, which individual title is watched the most and which the least. We try to match the respective successful genre or content type when acquiring new content. Second, we align our content release to days and dates on which VoD is consumed the most.”

How does your app perform on Amazon Appstore?

“Our apps for Fire TV and Fire tablets perform very well! In fact, the Amazon Appstore currently generates the most turnover and revenue for Kixi.” 

What is your experience developing and publishing your app for Amazon devices?  

“It was pretty straightforward. Amazon offers a good environment and framework for developers. Our tips for other developers: Just go for it!”

Getting started with Amazon Appstore

Amazon offers several tools and documents to get you started. You can test your existing APK within minutes to help you understand potentially required adaptations (if any). The Developer Portal also offers a variety of extensive documents to help you implement Amazon specific services, such as Login with Amazon, Amazon Device Messaging or the In-App Purchasing API.

We sat down with eOne to hear more about what drives their kids and family success on Amazon Fire tablets, and what they learned along the way about Amazon Appstore customers.

Entertainment One’s (eOne) ‘Family & Brands’ division creates kid’s content that connects with families around the world by carefully selecting, crafting, and nurturing the best children’s content into global brands.

Why did you initially choose to launch your apps and games on Amazon Appstore?

“Our digital services are an extension of our content and key to engaging with our brands’ fanbase, so it’s important for us to give our brands maximum visibility by making our apps available to as many children and families as possible. Amazon Fire tablets have been growing in popularity, especially in the kids’ space, and with part of our portfolio already available on Amazon Video, having our apps on the Amazon Appstore only seemed like a natural progression.”

What makes Amazon Appstore a valuable store for eOne’s apps?

“We are fortunate that our brands have huge recognition with Amazon consumers, but it’s also important for us to guarantee a good user experience. That’s why we really value their presence on the Amazon Appstore. The easy access to kids content made available by the Appstore family tab on Fire tablets ¹ makes it a very attractive store for our brands. The high level of content curation and commitment to offering a rich and versatile app catalogue for family and kids makes Amazon Appstore a great store for us to engage with our target audience.”

Is there a specific audience behaviour that is unique to Amazon Appstore?

“We have noticed that Amazon Appstore is the store on which our premium apps perform best. This suggests that Amazon consumers value the peace of mind brought by an upfront payment with no further transactions nor advertising.”

What is your approach and vision for content innovation and content breadth?

“Nowadays, consumers expect to have access to content anywhere, at any time and their appetite for it never stops growing. eOne quickly innovated in this area at a time when mobile devices were on the rise to give fans a wealth of high-quality digital content they could enjoy on-the-go or at home. Our digital content strategy has evolved and expanded over the years to include a broad portfolio of free and premium apps, as well as our subscription app, World of Peppa Pig. A lot of parents don’t want to spend hours on the Appstore looking for new content for their children. This app gives families a rich experience that combines games, educational play, videos and creative activities under one roof in a format that is regularly updated. Our apps are directly inspired by the themes, narratives and characters from our shows and feature the original sounds and music that the fans will recognise, and we think that makes our collection of apps truly engaging for our audience.”

How do you determine which monetisation model is appropriate for your apps and games?

“We always strive to make the best content available to as many children and families as possible and provide them a wide a range of options. The main challenge for any children’s app developer is adapting to the ever-changing needs and behaviours of our audiences across the world. By offering different apps models, whether it is free-to-download, subscription or premium apps, we want to make sure that all families can engage with our content in the way that works best for them.

“In our experience, Amazon Appstore customers have favoured the premium model for the peace of mind it gives parents and the great user experience it offers children. That is truly remarkable in an area that has been generally shifting towards free-to-download apps over the years.

“We are also seeing a very positive response to our subscription app World of Peppa Pig. There is clearly a demand for an app that is regularly updated with new content and combines lots of games, videos, episodes on-the-go, and activities under one safe umbrella. “

How do you leverage marketing opportunities with Amazon Appstore?

“To support the adoption of our subscription app World of Peppa Pig, we ran a discount exclusively for Amazon Appstore customers on our yearly subscription plan for the two weeks following Prime Day. The unique nature of the offer allowed us to obtain support from the Amazon Appstore team, which was incredibly helpful to give the app visibility over the promotion period. We promoted the offer on owned channels, the Amazon Appstore storefront and inside the app which generated a 123% increase on installs across Amazon devices.

“The best timing to run campaigns is when people purchase new Amazon devices, as they naturally look for content to download. The days immediately following key device activation dates like Black Friday, the holidays, and Prime Day are perfect opportunities for us to bring our apps to new Amazon device owners’ attention. That’s why we focus most of our marketing efforts after such dates and try to combine it with the release of discounts and new content.”

¹ Available for Appstore customers in US, DE and UK.

Getting started with Amazon Appstore 

Amazon offers several tools and documents to get you started. You can test your existing APK within minutes to help you understand potentially required adaptations (if any). The Developer Portal also offers a variety of extensive documents to help you implement Amazon specific services, such as Login with Amazon, Amazon Device Messaging or the In-App Purchasing API.

So, you’ve published a skill to the Alexa Skills Store and now customers are using it. What’s next? With the metrics dashboard, you can measure how customers are interacting with your skill so you can continue improving their experience over time. To help you get the most out of the metrics dashboard, this post walks through some of the top questions we get from the developer community about it.

1. Why is the number of Unique Customers and User Enablements different?

• User Enablements refers to the total number of unique users who enabled the skill. Some users might enable your skill, but not launch it. It’s important to note that user accounts are not unique to the device. If a single user enabled the skill from both Alexa app and Alexa Skills Store, it will be counted as one user enablement.
• Unique Customers are determined by Alexa user accounts, not devices. If a customer launches your skill on multiple devices that are all registered to the same user account, this is counted as one Unique Customer.

Another reason why these numbers are different is that some users might have enabled your skill, but not launched it yet. Only users who launch a skill are included in the count of Unique Customers.

2. How can I customize my skill dashboard to view the analytics data over time?

You can customize your reporting by selecting the Custom range button and choosing the time frame that you want to see. You can also use the Time Interval controls to change the date range (today, yesterday, last 7 days, last 30 days, or by selecting a custom date range). Using the Aggregation Period drop-down list, you can change how the data is grouped (every 15 minutes, hourly, daily, or weekly).

3. Why does the number of Unique Customers change depending on how the data is aggregated?

The number of Unique Customers changes depending on how you aggregate the data: per hour, per day, or per week. For example, consider the following scenario in which several customers use the same skill in the same day:

• Customer A uses the skill at 1:00 AM.
• Customers B and C both use the skill at 11:00 AM.
• Customer B uses the skill again at 3:00 PM.

When you aggregate the data by hour, you see one Unique Customer at 1:00 AM, two Unique Customers at 11:00 AM, and one unique customer at 3:00 PM. However, when you aggregate the data by day, you see a total of three Unique Customers. Thus, adding together the number of Unique Customers per hour does not necessarily give you the number of Unique Customers per day, because it doesn’t account for repeat users. The metrics that you see on the Analytics page do account for repeat customers, which can help explain why you see differences when you aggregate the data differently.

The Summary section of the Analytics page shows the total number of Unique Customers for the entire date range that you select.

4. Can I see the total metrics for all of a skill’s locales or languages in one place?

Yes. In the language drop-down list in the upper left, choose All to see combined metrics for all of the languages and locales in which your skill is available:

5. How long does it take for the data on the Analytics page to update?

All metrics data is refreshed every two hours, with the exception of the retention metrics (refreshed weekly) and the enablements metrics (refreshed daily).

6. Can I export the data?

Yes, you can export charts, grids, and other metrics visualizations as PNG or JPEG images. You can also download the raw data in CSV format.

Got More Questions?

Don’t worry, we’re here to help. If you run into more questions while using the metrics dashboard, a good place to start is the frequently asked questions about skill metrics. You can also post questions and chat with other Alexa skill developers on our developer forum, or connect with us through our contact form.

You can monitor the performance of your skill through usage metrics available to you in on the Analytics page in the developer console. These metrics give you useful information about how customers are using your skill, and can help guide your decisions on how to design your skill in order to improve the experience.

Related Content

• Frequently Asked Questions
• Alexa developer console
• Developer Forum
• Explore the Metrics Dashboard

Teach Your Monster to Read is part of The Usborne Foundation, which was founded by children’s publisher, Peter Usborne MBE with his children Nicola and Martin.

Peter Usborne has over 40 years of experience creating children’s books and Nicola and Martin also bring with them a wealth of experience creating books and interactive educational resources for children. Together, they wanted to use this knowledge to first address the ongoing crisis in early literacy, which sees too many children fail to learn to read. The goal was to combine education, research and technology to make this difficult task both interactive and fun.

With this in mind, Teach Your Monster to Read, a suite of online literacy games, was born. The games create a fun, playful environment in which over 15 million children have already tackled their first steps of reading.

The success of Teach Your Monster to Read has led the team behind the game to find new ways to help improve the lives of children through learning and education. They aim to create exciting learning experiences through engaging stories, technology, people led design and of course, fun.

They have now also turned their focus to the growing issues with childhood obesity in the UK. Their aim is to create new, interactive ways to help children learn about healthy food.

We had the opportunity to talk to Alex Goss, product manager of Teach your Monster How to Read, and Jonathan Skuse, the game’s co-designer and co-developer, about their game and their experience and reasons for publishing their game on the Amazon Appstore.

Why did you choose to publish your game on the Amazon Appstore and for Amazon Fire tablets?

“We chose to publish Teach Your Monster to Read on the Amazon Appstore because, like any app developer, we want to be where the attention is. Kids play games on all sorts of devices, so it was important to us that we help kids learn to read whether they’re playing in the car on their dad’s phone, sitting at the kitchen table playing on a tablet, or in school during an ICT lesson.”

How do Amazon customers fit into your target group? What is the value of Amazon customers compared to other audiences?

“We don’t think there’s one sort of Amazon customer, so there isn’t one answer to this. However, many parents with young kids use Fire tablets. What’s special about these tablets is that there’s also a dedicated Kids Edition tablet. We’ve also released a version of Teach Your Monster to Read that’s available in FreeTime Unlimited, Amazon’s subscription service for kids content. Amazon designing a tablet built from the ground up for kids goes hand-in-hand with the efforts we put into creating impactful educational games.”

How does your game perform on Amazon Fire Tablets?

“Teach Your Monster to Read performs well on the Amazon Appstore. We think that comes down to excellent QA, listening to feedback from Amazon customers, and working closely with the Amazon Appstore team.”

What is your experience developing your game for Amazon devices?

“Our game started out as a Flash web game about 8 years ago, and it’s now available on all app stores, plus there’s an HTML version on the web. It’s fair to say that it’s come a long way since then and with over 150 million plays, it’s exceeding everyone’s expectations! Developing the game for Amazon has needed little overhead. Amazon’s devices behave in a coherent way; meaning we can develop with confidence due to low fragmentation.”

What is your experience publishing and maintaining the game on Amazon devices? Any tips for other developers?

“It’s been quite straightforward to maintain a stable app on Amazon devices. Fire Kids Edition tablets are a great choice for parents who want to give kids a tablet experience designed just for them while also having peace of mind that the content their kids are viewing is kid-friendly. For those not currently publishing to Amazon devices, we’d say it’s extremely likely that you’re missing out on a big opportunity to reach your audience.”

Getting started with Amazon Appstore

Amazon offers several tools and documents to get you started. You can test your existing APK within minutes to help you understand potentially required adaptations (if any). The Developer Portal also offers a variety of extensive documents to help you implement Amazon specific services, such as Login with Amazon, Amazon Device Messaging or the In-App Purchasing API.

We are delighted to announce that In-Skill Purchasing (ISP) is now generally available for Alexa skills published in France, Italy and Spain! That means you can now earn money by offering premium content in your Alexa skills published in those countries as well. For example, users can purchase virtual game products, unlockable features and add-on content while seamlessly interacting with the voice experience. You define your offering and price, and Alexa handles the voice-first purchasing flow. We also provide self-service tools to manage your in-skill products and optimize your sales performance over time.

ISP currently supports one-time purchases for entitlements that unlock access to features or content in your skill, subscriptions that offer access to premium features or content for a period of time, and consumables, which can be purchased, depleted and purchased again.

One-time purchases unlock access to features or content in a skill. Customers purchase the content once for an entitlement that does not expire. A few local examples are:

• Mentalshock (Italy): an engaging interactive story where players have to explore a futuristic world full of danger and intrigue. The first part is free and the second is unlockable for €1.99. The skill also offers hints for €0.99.
• Hablando se aprende – las tildes (Spain): This fun game lets you guess where the accent mark (á, é, í, ó, ú) should be placed in a word. For only €0.99 the skill also provides corrections and tips to learn more effectively.
• Ciné Quizz (France): this original quiz has 200 movie’s extracts in its database and is perfect to play with friends and family.

With an in-skill subscription, you can offer access to premium features or content for a period of time. Subscriptions support free trials (1 to 31 days) and two billing cycles: monthly or yearly. Examples include:

• Trova Telefono (Italy): a handy skill that calls your phone when you can’t find it around your home. You can now get unlimited uses for €4.99 per year.
• Sleep Sounds: Rain (available in France, Italy and Spain): Skill plays calming rain sound loops to help you fall asleep faster, sleep better, relax, meditate, or drown out distracting noises. With the premium subscription you can enjoy Premium sound quality and enhanced looping to listen longer.
• Rueda de letras (Spain): this fun vocabulary game allows to unlock new game categories for only €1.99/week or €3.99/month.

Consumables can be purchased, depleted, and purchased again. This feature is often used to monetize game hints, extra lives or special items for an adventure game. Examples of skills using consumables include:

• Quiz Treccani (Italy): this devilishly challenging quiz around Italian grammar and etymology offers three daily games, and an additional five games can be unlocked for €0.99.
• Quiz Vero o Falso (Italy): the free version of this trivia allows 3 wrong answers everyday, and additional 10 lives can be unlocked for €0.99.
• Lemonade Stand (available in France, Italy and Spain): Lemonade Stand is a simple business simulation. The game is easy to play, players run a lemonade stand and make business decisions every day. It is possible to protect the lemonade stand during a thunderstorm by buying a 10 umbrellas consumable package for €0.99.

A skill may use different product types together e.g. individual one-time purchases for unlocking specific features/content, and a subscription to unlock all premium features/content.

Getting Started with In-Skill Purchasing

To help you get started with In-Skill Purchasing, we have put together a free guide which you can consult. You can download the guide in your preferred language (Click on the link to go to the preferred version: French, Italian, Spanish). Other useful resources:

• How to design for In-Skill Purchasing
• Sample Code on GitHub: Premium Hello World, Premium Facts

Resources

• Starter Kit: How to Make Money with Alexa Skills: France, Italy and Spain
• Best practices for designing a premium customer experience
• Developer Story: With In-Skill Purchasing, Gal Shenar Sets His Growing Voice Business Up for Long-Term Success
• Developer Story: Musicplode Media Uses In-Skill Purchasing to Turn Its “Beat the Intro” Voice Game into a Hit for Alexa Customers
• Developer Story: Hypno Therapist Alexa Skill Uses In-Skill Purchasing to Reach More Customers and Build a Business
• Test your in-skill purchase before you submit it for certification
• Set up tax forms for payment purposes
• Measure skill usage and earnings with the analytics dashboard within the Developer Console

The Alexa Web API for Games (Preview) introduces web technologies and tools to create visually rich and interactive voice-controlled game experiences. With the Alexa Web API for Games, you’ll be able to build for multimodal devices using HTML and web technologies like Canvas 2D, WebAudio, WebGL, JavaScript, and CSS. You can apply to participate here.

Build Games with Web Technologies

With Alexa Web API for Games, you can take advantage of your knowledge of existing web technologies to create high-end voice based games with HTML/CSS/Javascript or 3D experiences with WebGL. You can build a custom Alexa skill that starts a web application (web app). The web app will be displayed on the device and can communicate with the skill to handle voice requests and react to local events like the microphone opening and closing. Customers can interact with the web app using a variety of modes such as voice through the skill, touch on devices with touch screens, and remotes on Fire TVs.

Ticket to Ride, Masamune’s Room, Lemonade Stand, and Goal Scorer are game skills that were built using Alexa Web API for Games:

• Ticket to Ride (UK, US, FR) and Ticket to Ride Europe (UK, US, FR): New companion game for the popular Ticket to Ride board game by Days of Wonder. You can ask Alexa to provide interactive tutorials, keep score or even play the game with you.
• 政宗の部屋 (Masamune’s Room) (JP): Based on the mobile game Ikemen Sengoku: Romances Across Time, Masamune’s Room is a life simulation game from Cybird, Inc. You can experience living with Japanese Sengoku warlord, Masamune Date.
• Lemonade Stand (DE): Business simulation game from Krögoor Entertainment that allows the customer to run a lemonade stand and make business decisions every day. You can play the game in single player mode or with up to eight people.
• Elfmeterschießen (Goal Scorer) (DE): Sports game that lets customers enjoy the excitement of a soccer penalty shoot-out as if they were playing against the best teams at the World Cup.

Alexa Web API for Games will be supported on Echo Show devices and select Fire TVs*. If you want to build visual experiences that can reach all Alexa multimodal devices today, use the Alexa Presentation Language (APL).

Developer Preview Details

Developers in the preview will have access to the Alexa.Presentation.HTML interface and the Alexa JavaScript API. To be considered for the program, you must meet the following criteria:

1. You want to build a game or game-like experience that ensures that voice is the central mode of interaction in the player experience and adds to the experience in a material and innovative way.
2. You are willing to support your game on Echo Show devices and Fire TVs.
3. You are willing to provide a fallback experience for devices that don’t support Alexa Web API for Games.

Please follow the Alexa developer blog to get the latest updates regarding Alexa Web API for Games.

Apply for the Developer Preview

You can apply to participate in the Alexa Web API for Games developer preview by telling us about your game in this short survey, and we’ll notify you if your application is selected.

^{*Alexa Web API for Games will be supported on Fire TV Stick 4K, Fire TV with 4K Ultra HD and Alexa Voice Remote, and Fire TV Cube.}

Alexa Web API for Games（プレビュー版）は、視覚的にリッチで、音声でインタラクティブに制御できるゲームを作成するためのウェブ技術とツールを導入します。Alexa Web API for Games を使うと、HTMLやCanvas 2D、WebAudio、WebGL、JavaScript、CSSなどのウェブ技術を使用してマルチモーダルデバイス向けにスキルを開発できます。開発者プレビューのお申し込みはこちら（英語）。

ウェブ技術を用いたゲームの開発

Alexa Web API for Gamesを使用すると、既存のウェブ技術のノウハウを活用して、ハイエンドのHTMLおよびJavaScriptベースの音声ゲームを作成できます。この技術を用いてカスタムスキルを開発すると、ウェブアプリが起動し、デバイスの画面に表示され、スキルを通して音声リクエストを処理することができます。

ユーザーはスキルによる音声、タッチスクリーンつきデバイスへのタッチ、Fire TVのリモコンなど、さまざまな手段でウェブアプリを操作できます。

Ticket to Ride、政宗の部屋、Lemonade Standは、Alexa Web API for Gamesを使用して構築されたスキルです。

• Ticket to Ride (英国、米国、フランス)とTicket to Ride Europe (英国、米国、フランス)： Days of Wonderによる人気のボードゲーム、Ticket to Rideをサポートするスキルです。インタラクティブなチュートリアルの提供、スコアの維持、ゲームのプレイを依頼することができます。
• 政宗の部屋(日本)： 株式会社サイバードによる恋愛シミュレーションゲーム「イケメン戦国◆時をかける恋」を元にしたスキルで、伊達政宗との生活を体験できます。このスキルではリッチな視覚および音声コンテンツを追求しており、時間帯や季節によって背景が変わります。
• Lemonade Stand（ドイツ、米国、フランス、イタリア、スペイン）：Krögoor Entertainmentによる、ビジネスシミュレーション型のゲームです。ユーザーがレモネードスタンドを運営し、ビジネス上の意思決定を行います。プレイヤーモードまたは最大8人でゲームをプレイすることができます。
• Elfmeterschießen (Goal Scorer) (ドイツ): 世界で一番強いサッカーチームに対して、PK戦をするような興奮を楽しめるスポーツゲームです。

Alexa Web API for GamesはEcho Showデバイスや一部のFire TV*でサポートされています。すべての画面付きデバイスで表示可能な視覚コンテンツを提供する場合は、 Alexa Presentation Language（APL）を使用してください。

開発者プレビューの詳細

プレビューに選ばれた開発者はAlexa.Presentation.HTML インターフェースとAlexa JavaScript APIにアクセスできます。このプログラムへ登録するには、以下の条件を満たす必要があります。

1. プレイヤーのエクスペリエンスにおいて音声がインタラクションの中心となるようにする。ゲームまたはゲームに似たエクスペリエンスを実質的で、イノベーティブな方法で提供する。
2. Echo ShowデバイスおよびFire TV向けをサポートしたゲームを提供する。
3. Alexa Web API for Gamesをサポートしていないデバイスについては、代替となるエクスペリエンスを提供する。

最新のAlexa Web API for Gamesに関するニュースは、Alexa開発者ブログでご確認ください。 

開発者プレビューに申し込む

Alexa Web API for Gamesの開発者プレビューへの登録を申請するには、短いアンケートであなたのゲームについて教えてください。プレビューに選ばれた場合は、お知らせいたします。

*Alexa Web API for GamesをサポートしているFire TVシリーズはFire TV Stick 4K、Fire TV with 4K Ultra HD、Fire TV Cubeになります。

We’re excited to introduce a new Alexa skills training course called How to Build a Multimodal Alexa Skill. This self-paced course is free and available to anyone ready to build skills using Alexa Presentation Language (APL). APL is a design language and set of tools that make it easy to create visually rich Alexa skills for devices with screens. There are more than 100 million Alexa devices, including Echo devices such as the Echo Spot and Echo Show, Fire TVs, Fire Tablets, and devices from other manufacturers with Alexa built-in like the Lenovo Smart Tab devices and LG TVs. If you’re unfamiliar with building skills, start with the Cake Walk course. If you are familiar with building skills, you can get started now with the How to Build a Multimodal Alexa Skill course.

Enhance Your Voice Experience with New Modalities

When you build a multimodal Alexa skill, you combine voice, touch, text, images, animation, audio, and video in a single experience. Adding new modalities can enhance voice experiences and make skills more engaging and interactive for customers. You can bring your branding visually to customers and provide unique interactions not possible with a voice-only design, helping your skill to stand out to your customers.

In the Cake Walk course, you build a skill that counts down the days until it is your birthday. When it is your birthday, Cake Walk will deliver a special message. In the new course, you will make your own version of this skill, enhanced with images, animations, text, and a special birthday video. Here’s a sneak peak.

Course Components

The course will introduce the important concepts of multimodal skill design:

• Why Build Multimodal Alexa Skills

• Build Your First Visuals with the APL Authoring Tool

• Modify Your Backend to Add LaunchRequest Visuals

• Create More Visuals with Less

• Add Animation and Video Control

• Wrapping Up & Resources 

Each part includes step-by-step instructions so you can easily apply the course material to your skill.

More Training Opportunities to Enhance Your Skills

Once you have completed the Cake Walk series, continue to learn with these additional training materials:

• Alexa Design Guide: Learn the principles of situational voice design so that you can create voice-first skills that are natural and customer-centric.

• How to Design for In-Skill Purchasing: Learn best practices, tips, and resources for In Skill Purchasing (ISP) design.

Get Started with How to Build a Multimodal Alexa Skill

Click here to get started. And let us know what you think! You can also join us on our Twitch stream on November 21st, 2019 at 9:00 AM PST. Feel free to reach out to me on Twitter @JoeMoCode to share comments and feedback, or send your APL questions to the ASK Developer Forums.

We are excited to announce that starting today developers residing in India can start earning money through the Alexa Developer Rewards Program for highly engaging skills published to the India Alexa Skills Store.

Eligible skill categories include:

• Education & Reference
• Food & Drink
• Games, Trivia & Accessories
• Kids
• Health & Fitness
• Lifestyle
• Music & Audio
• Productivity

We evaluate customer engagement and determine skill rewards on a monthly basis. You can increase customer engagement and potentially earn more rewards by publishing more engaging skills to the India Alexa Skills Store, and by improving your existing skills. Read about these best practices on how to promote your Alexa skills and reach more customers.

There is no need to apply for the program. If your skill qualifies for rewards, you will receive an email from the Alexa team. The email will cover the next steps for your registration and payment account setup. Make sure your country/region is correctly set to India in the profile section of the Alexa Developer Portal.

To date, the Alexa Developer Rewards program has paid millions of dollars to developers in more than 20 countries. Visit the Alexa developer blog to read how the Alexa Developer Rewards program has helped developers like Gal Shenar, David Markey, and Oscar Merry earn money and build their voice businesses.

We’re excited to see what voice experiences Indian developers will build next.

March 4 marked the kickoff of the third Alexa Prize Socialbot Grand Challenge, in which university teams build socialbots capable of conversing on a wide range of topics and make them available to millions of Alexa customers through the invitation “Alexa, let’s chat”. Student teams may begin applying to the competition now, and in the next six weeks, the Alexa Prize team will make a series of roadshow appearances at tech hubs in the U.S. and Europe to meet with students and answer questions about the program.
 
As the third Alexa Prize Socialbot Grand Challenge gears up, the Alexa science blog is reviewing some of the technical accomplishments from the second. An earlier post examined contributions by Amazon’s Alexa Prize team; this one examines innovations from the participating university teams.

The 2018 Alexa Prize featured eight student teams from four countries, each of which adopted distinctive approaches to some of the central technical questions in conversational AI. We survey those approaches in a paper we released late last year, and the teams themselves go into even greater detail in the papers they submitted to the latest Alexa Prize Proceedings. Here, we touch on just a few of the teams’ innovations.

Handling automatic speech recognition (ASR) errors

Conversations with socialbots can cover a wide range of topics and named entities, which makes automatic speech recognition (ASR) more difficult than it is during more task-oriented interactions with Alexa. Consequently, all the teams in the 2018 Alexa Prize competition built computational modules to handle ASR errors.

Several teams built systems that prompted customers for clarification if ASR confidence scores were too low, and several others retained alternative high-scoring ASR interpretations for later re-evaluation.

Gunrock, the team from the University of California, Davis, that won the 2018 challenge, built a system for correcting ASR errors that uses the double-metaphone algorithm, an algorithm that produces standardized representations of word pronunciations. When the ASR system assigned a word a low confidence score, Gunrock’s error correction module would apply the double-metaphone algorithm to the word and then search a database of metaphone-encoded pronunciations for a partial match. Those pronunciations are grouped according to conversation topic, which lets the module take advantage of contextual information.

So, for instance, the metaphone representation of the phrase “secure holiday” is SKRLT, which doesn’t occur in Gunrock’s database. But SKRLT is a substring of the metaphone representation APSKRLTS, which does occur in the database. So Gunrock’s system would correct SKRLT to APSKRLTS and return the corresponding English phrase: “obscure holidays”.

Knowledge graphs

Carrying on a conversation requires knowledge, and most teams chose to encode their socialbots’ knowledge in graphs. A graph is a mathematical object consisting of nodes, usually depicted as circles, and edges, usually depicted as line segments connecting nodes. In a knowledge graph, the nodes might represent objects, and the edges might represent relationships between them. Several teams populated their knowledge graphs with data from open sources such as DBPedia, Wikidata, Reddit, Twitter, and IMDB, and many of the teams built their graphs using the Neptune graph database service from Amazon Web Services.

 ^{An example of a knowledge graph built with Amazon Web Services’ Neptune service}

Alana, the team from Heriot-Watt University in Scotland and the third-place finisher in the 2018 challenge, used Neptune to build a knowledge graph that encodes all the information in the Wikidata knowledge base, plus some additional data from the DBpedia knowledge base. When the Alana socialbot identifies a named entity in a conversation, it begins a context-constrained exploration of the graph, assembling a subgraph of linked concepts.

If someone chatting with the Alana bot mentioned the movie E.T., for example, Alana’s linked-concept generator would follow the Wikidata link from E.T. to the entry for Drew Barrymore, who appeared in the film, but not to the entry for Sweden, which is the second country in which the movie was released. Then, once it has built up a database of linked concepts, the Alana socialbot selects one at random to serve as the basis for a conversational response.

Natural-language understanding (NLU) for open-domain dialogue

Amazon researchers provided the student teams with default modules for doing natural-language understanding (NLU), or extracting linguistic meaning from raw text, but most teams chose to supplement them or, in some cases, supplant them with systems tailored specifically to the demands of conversational AI. Student teams built their own modules to classify utterances according to intent, or the goal the speaker hopes to achieve, and dialogue act, such as asking for information or requesting clarification; to identify the topics of utterances; and to assess the sentiments expressed by particular choices of phrasing, among other things.

Most of the NLU literature focuses on relatively short, goal-directed utterances. But in conversations with socialbots, people will often speak in longer, more complex sentences. So Gunrock built an NLU module that splits longer sentences into smaller, semantically distinct units, which then pass to additional NLU modules.

To train the segmentation module, Gunrock used movie-dialogue data from the Cornell Movie-Quotes Corpus, which had been annotated with a special tag (“”) to indicate breaks between semantically distinct units. On a test set, the module was 95.25% accurate, and an informal review indicated that it was accurately segmenting customers’ remarks. For example, the raw ASR output “Alexa that is cool what do you think of the Avengers” was segmented into “Alexa that is cool what do you think of the Avengers ”.

Dialogue management

The outputs of the NLU modules, along with any other utterance data the teams deem useful, pass to the dialogue management module, which generates an array of possible responses and selects one to send to Alexa’s voice synthesizer. 

Alquist, the team from the Czech Technical University in Prague and the runner-up in the 2018 challenge, used a hybrid code network (HCN) for its dialogue manager. An HCN combines a neural network with handwritten code that reflects the developers’ understanding of the problem space. HCNs can dramatically reduce the amount of training data required to achieve a given level of performance, by sparing the network from having to learn how to perform tasks that are easily coded.

In Alquist’s case, the added code has two main functions: it filters out suggested responses that violate a set of handwritten rules about what types of responses should follow what types of utterances, and it inserts context-specific data into responses selected by the neural net. So, for instance, the neural network might output the response “That movie was directed by {say_director}”, where {say director} is an instruction to a complementary program that has separately processed data from the NLU modules.

Customer experience and personalization

All of the teams had to address the question of when to switch conversation topics and how to select new topics, but Iris, the team from Emory University, built a machine learning model that predicted appealing topics on the basis of conversational history — what topics a customer had previously accepted and rejected and what types of interactions he or she had previously engaged in. Iris trained their model on data from their socialbot’s past interactions with customers.

In tests, Iris compared their model to the simple heuristic of suggesting new topics in order of overall popularity and found that, on average, their model’s recommendations were 62% more likely to be accepted.

It was a pleasure to work with the student teams who competed in the 2018 Alexa Prize and a privilege to witness their innovative approaches to a fundamental problem in artificial-intelligence research. We can’t wait to see what the next group of teams will come up with!

Behnam Hedayatnia is an applied scientist in the Alexa AI group.

Papers: 

• “Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize”
• The 2018 Alexa Prize Proceedings

Alexa science

Related:

• Contextual Clues Can Help Improve Alexa’s Speech Recognizers
• Public Release of Fact-Checking Dataset Quickly Begins to Pay Dividends
• How Alexa Knows “Peanut Butter” Is One Shopping-List Item, Not Two
• How Alexa Is Learning to Converse More Naturally
• How Alexa Can Use Song-Playback Duration to Learn Customers’ Preferences
• The 2019 Alexa Prize Roadshow
• Exciting Changes for the Alexa Prize Socialbot Grand Challenge 3
• Alexa Prize Socialbot Grand Challenge 3 Application Period Now Open