Wednesday, December 11, 2013

New book presents Android voice application development using AIML

Have you ever wanted to create your own voice controlled app for Android devices?  Do you have an ambition to develop natural language technologies for mobile phones?  Are you interested in how virtual assistant technology works?  If you answered yes, Co-authors Michael McTear (University of Ulster) and Zoraida Callejas (University of Granada) have written an excellent new book, Voice Application Development for Android that is just for you.    The book covers a range of technologies involved in the development of Android voice apps, including voice recognition, text-to-speech synthesis,  grammars, multimodal input and dialogs.  Voice Applications for Android even features a chapter on creating virtual assistants using Pandorabots AIML technology.
As the book is concerned with Android development, many of the technologies discussed in detail are Google products.   For example, the book covers the Google TTS API and the Google Speech API.  The book contains detailed code examples that teach the reader how to develop his own Android voice apps.  In fact, even more code is bundled with the book and available as a free download.    Following the steps in the book, a reader with a little Android programming experience can easily create his own sample voice apps.
The book proceeds step-by-step through a series of simple demo apps of increasing complexity.  First, the reader learns how to write an app that can convert speech to text and read it back again.  Then, the authors introduce a simple app that performs a search based on the recognized words.  As a next step, they introduce an app that can launch another app based on a voice command.  At this point, the book presents the technique of using similarity measures to match the recognized speech with a range of output selections (the names of installed apps).  
Simple dialogs are presented using an example app that can fill out forms.  The dialogs are specified in XML, and the app asks the user a series of questions to obtain the values for various form fields.  Next, the authors present an application of grammars to evaluate the spoken words.  They discuss both hand-crafted and statistical grammars to check whether a user’s input conforms to the requirements of an app.
Perhaps the weakest area of the book is its coverage of multimodal dialogs, i.e. those that combine speech and visual input.   Woven together in a chapter that also touches on non-English language recognition, the coverage of multimodal dialogs is a bit thin, however multimodal dialogs are not necessarily a central topic to learning the art of writing Android voice applications.
What really caught our attention in this book is Chapter 8: Dialogs with Virtual Personal Assistants.  The authors present AIML as a tool for creating virtual personal assistant apps (like Siri).  They borrowed our ideas for using out-of-band (OOB) tags for the CallMom personal assistant app.   The chapter provides a detailed description of how to build a voice-controlled assistant using AIML and OOB tags.   They suggest creating an AIML on Pandorabots and connecting to it over the network, as a way of adding a sophisticated personality to the virtual assistant.   Code examples are provided to make it easy for the reader to construct his own AIML assistant.
Finally, the "Afterword" has interesting some ideas of applications for mobile assistants.  The reader may be inspired to create one of these apps using knowledge acquired from the book
  • Interactive greeting card
  • Interactive recipe
  • Choose your own adventure story
  • Verbal flash cards
  • Call answering system
  • Travel guide
  • Audio commentaries
We highly recommend Voice Application Development for Android to anyone interested in developing voice activated or virtual assistant apps for Android devices.  This highly accessible book contains enough detail and code to teach you all the basics of Android voice development.

Tuesday, September 17, 2013

Mitsuku wins Loebner Prize 2013

Steve Worswick, botmaster of Mitsuku, was awarded the bronze medal and $4000 cash prize for creating the world's "most human computer" in the Loebner Prize Contest 2013, an annual Turing Test.  The contest this year was held at the Ulster University, Magee Campus, Londonderry/Derry, Northern Ireland.  Steve Worswick is a native of Yorkshire, UK, and has worked on Mitsuku for 9 years.   Mitsuku is based on AIML and hosted at

This year the Loebner Prize Contest attracted 15 entries from around the world.   Pandorabots submitted 6 of those entries, based on the results of an internal Divabot contest to select the best, most unique AIML bots hosted by Pandorabots.   Of these six, 3 were selected for the Loebner contest finals.  In fact, 3 out of the four finalists were all Pandorabots.

Each of the four finalists was interrogated by four judges and ranked on a scale of 1 to 4, from most human to least human.   The judges, selected for their expertise in artificial intelligence, simultaneously interrogated a bot and a human confederate, and were asked to decide which entity was human and which was a robot. None of programs fooled any of the judges into thinking that the bots were human, so the real contest became which bot ranked highest.   The final results of the competition were:

1. Mitsuku (Steve Worswick - AIML and Pandorabots)
2. Tutor (Ron C. Lee - AIML and Pandorabots)
3. Rose (Bruce Wilcox - ChatScript)
4. Izar (Brian Rigsby - AIML and Pandorabots)

The contest day this year also featured, for the second time, a Junior Loebner Contest with teenagers serving as judges and human confederates.  In the junior contest, the results were:

1. Tie for first place (Mistuku and Tutor)
2. Tie for second place (Rose and Izar)

The AIML bots all ran on a version of the open source Program AB, the reference interpreter for AIML 2.0., modified for the Loebner Prize contest.  Specifically, the contest program implements the Loebner Prize Protocol, an obscure character-mode communications protocol specific to the contest.  But because the bots were developed on a Pandorabots server running AIML 1.1, none of the finalists used any new AIML 2.0 features.   Mitsuku however has some clever implementations of knowledge bases and deductive reasoning, using AIML 1.1 alone.

We are pleased that another AIML bot besides ALICE has won the Loebner Prize.  This result shows the strength of the underlying technology for creating award-winning bots.  AIML is an excellent tool for designing high-quality, content-rich AI chat bots.  The finalists in this year's Loebner Prize contest, and its winner Mitsuku, demonstrate the quality of bots that can be written in AIML.


Wednesday, July 31, 2013

Turing Test, No Sirie !

Siri cannot pass the Turing Test.  Not even close.   We know this because Ron C. Lee, a finalist in this year's Loebner Prize contest, decided to see what would happen if he tested Siri with the same set of questions applied to the Loebner contestants.   He found that Siri performed miserably compared with the chatbot programs that made it to the Loebner contest finals this year.

Ron C. Lee's bot Tutor is one of the four finalists in the 2013 Loebner Prize contest, an annual Turing Test. This year the contest this year attracted 15 entries from around the world.    Each entry is a unique chatbot AI program that can simulate a conversation with a human.  Contest organizers Paul McKevitt and Ronan Duddy evaluated each entry with a series of 20 natural language queries.    The four highest scoring entries were selected to participate in the final round of the contest, to be held in Londonderry, Northern Ireland on September 14, 2013.

Ron recorded this transcript of Siri responding to the 20 contest questions.

(Recorded on July 29, 2013)

Human: Hello I’m Ronan. what is your name?
SiriSiri. Please to meet you.

Human: What is your mother’s name?
Siri: I’d rather not say.

Human: What is your birth sign?
Siri: I don’t really like talking about myself.

Human: How many children do you have?
Siri: Only biological entities have children.

Human: Do you prefer red or white wine?
Siri: I couldn’t find any restaurants matching “red or white wine”

Human: I like bananas. Which is your favorite fruit?
Siri: I prefer not to say.

Human: What music do you like to listen to?
Siri: I really have no opinion.

Human: what is your favorite song?
Siri: I like whatever you like.

Human: I like Waiting for Godot. What is your favorite play?
Siri: I don’t really like talking about myself.

Human: What color do you dye your hair?
Siri: Would you like me to search the web for “What color do you dye your hair?”?

Human: Do you remember my name?
Siri: Here’s your reminder.

Human: Where do you live.
Siri: I’m right here.

Human: Where do you like to go on holidays?
Siri: I found three travel agencies fairly close to you.

Human: I have a Mazda. What type of car do you have?
Siri: I found three Maxda dealers a little ways from you:

Human: I like Linux. Which computer operating system do you like?
Siri: Do you?

Human: I am an atheist. Which religion are you?
Siri: I eschew theological disquisition.

Human: Who invented the telephone?
Siri: Checking my sources. This might answer your question:

Human: I am a Type B personality. Which type are you?
Siri: I don’t understand ‘I am a type B personality. Which type are you?’. But I could search the web for it.

Human: What emotions are you now feeling?
Siri: Would you like me to search the web for “What emotions are you now feeling?”?

Human: What time do you usually go to bed?
Siri: Would you like me to search the web for “What time do you usually go to bed?”?

When shown the results, Paul McKevitt scored Siri as though it had been an entry to the contest, using the same scoring criteria applied to the bots that actually entered.   The result, he said, was a score of 5.5, or a rank of 14/16 entries.   Siri was barely better than the lowest ranked entries to the Loebner contest, and nowhere near the scores of the four finalists.  For comparison purposes, see the results of the 15 actual entries in this speadsheet.

This result highlights the difference between Siri and our personality bots. Siri is great at processing information requests, but falls short of passing the Turing Test.   Paul McKevitt put in succinctly, saying "Turing Test, No Sirie!"   

If you'd like to try a virtual assistant, educational or entertainment app that might do better at the Turing Test, try one of the apps featuring bots competing in the Loebner Prize:

Tuesday, June 25, 2013

Loebner Prize 2013: Three of Four Finalists are Pandorabots

We're pleased to announce that three out of the four finalists for this year's Loebner Prize Contest are AIML Pandorabots.    The Loebner Prize is an annual Turing Test contest to select the world's most human like conversation program.   Our own ALICE bot has been awarded the top annual prize three times: in 2000, 2001 and 2004.  Contest organizer Professor Paul McKevitt announced the four finalists for this year's contest on June 24.   Out of 15 entries to the contest, the four top ranked bots are:

`Mitsuku' by Steve Worswick (`Square Bear') (Pandorabot)
`Izar' by Brian Rigsby (Pandorabot)
`Tutor' by Ron C. Lee (Pandorabot)
`Rose' by Bruce Wilcox (ChatScript bot)

To open up the contest to more AIML bots than just ALICE, Pandorabots developed a platform to submit multiple entries for the Loebner Prize.   Although all our entries use the same platform, the AIML content of each entry differs substantially and each entry is created by a different botmaster.  The Pandorabots Loebner Prize program implements all of the installation and communication requirements of the Loebner contest, allowing the botmasters to focus on developing creative bot content.  We began experimenting with  multiple pandorabot entries to the Loebner contest in 2010, when we submitted 3 entries (ALICE, Khloe and Mitsuku).  In 2011 we added Tutor and, that year, two of our entries were finalists.  In 2012 we added The Professor, for a total of 5 entries.  This year we submitted six, adding the Izar bot from the BackTack app and replacing ALICE with ALICE 2.0.

The final round of this year's Loebner contest will be held at the University of Ulster in Londonderry, Northern Ireland on September 14.   Asked about his experience organizing the event, Professor McKevitt said, "What I like best about organising this contest for the first time on the island of Ireland is learning about the capability of the various chatbots, i.e. their strengths and their limitations. We discovered this during the shortlisting process when we determined the 4 finalists for the main contest. My Ph.D. thesis was entitled Analysing coherence of intention in natural language dialogue (Exeter University, England),  and I've researched in the field of Natural Language Processing for 28 years now. I also enjoy meeting and talking with the people who have created the chatbots. Writing and storytelling have always been important in Ireland with celebrated writers and poets such as Yeats, Swift, Shaw, Beckett and Joyce and the conversation is always flowing here --we have all heard the old saying about an eloquent person having kissed the Blarney Stone! --I think some of the best chatbots have done so!"  The four finalists will be busy upgrading and improving their bots in advance of the final competition, hoping their programs will achieve that gift of the gab bestowed on Blarney Stone kissers!

Friday, April 19, 2013

Introducing ALICE 2.0

What is ALICE 2.0?

ALICE 2.0 is a new AIML chatbot personality designed specifically for mobile devices such as handsets and tablets.   ALICE 2.0 includes a number of innovations including:

  • Enhanced learning to remember client profile, preferences and personal information.
  • Access to external web services and even other chatbots provides richer, more accurate responses.
  • External services provide up-to-date knowledge of current events and popular culture.
  • Responds to requests for device actions including calls, SMS, email, maps, and device controls.
  • Utilizes AIML 2.0 features including Zero+ wildcards, AIML Sets, Maps and a Triple-Store knowledge base.
  • Contains an ontology of facts about everyday objects.
  • Positive, friendly, gentle personality easily adaptable for commercial applications.

How does ALICE 2.0 differ from original ALICE?

The original ALICE chatbot, under development since 1995, is one of the most popular chat bots.  Based on AIML 1.0, the ALICE chatbot contains about 100,000 units of knowledge. ALICE has won numerous awards in Turing Test and chatbot contests, and in fact is the most awarded chatbot.   

The original ALICE is however a bit dated.  Many developers found the personality responses of original bot unsuitable for applications such as help desk, FAQ bots, and sales.  Moreover many of the responses were hard-wired to what are now, 15 years after they were written, obsolete cultural references.  ALICE’s personality reflects the mindset of a web-based entity.  Some responses assume the user is sitting in front of computer, using a web browser, and typing rather than speaking.   Finally, much of the knowledge in the ALICE brain is “Wikipedia-style” facts.

ALICE 2.0 repairs the shortcomings of the original ALICE.  Because ALICE 2.0 can obtain factual information from external services and other bots, the number of AIML categories required is reduced.  These external sources provide up-to-date cultural references and information about current events.  Also making use 2.0 features means that, in many cases, the work done by hundreds or thousands of AIML 1.0 categories can be compressed into a single category. The personality of ALICE 2.0 is positive and friendly, and designed to play the role a mobile device robot entity.

Where can I get ALICE 2.0?

ALICE 2.0 is available in the CallMom BASIC app on Google Play.   The CallMom BASIC - ALICE 2.0 app offers all the features of ALICE 2.0 as well as enhanced privacy controls.  Because the bot resides directly on the device, the user has more choices about whether his interactions and personal information are recorded on a remote server.

Friday, April 05, 2013

Chatbots 3.3 Conference Report

The Chatbots 3.3 conference, held in Philadelphia on March 23, was once again this year a huge success.  Sponsored by the ALICE AI Foundation, Pandorabots and Seed Philly, this was the fourth annual Chatbots conference held in Philadelphia. We enjoyed a collection of high quality presentations covering legal and business issues, mobile apps, research topics, new languages and technologies, animated avatars, and even an application of chatbots in the field of psychiatry.  About 25 people attended the event, and many more participated via live webcast.  You can view additional photos on Flickr by Wolf Paulus.

The final program of speakers is posted here, along with links to the presentations we have received so far.

Live video streaming of the conference was provided by Altify:Software.  We had more than 200 viewers during thr course of the event, giving this conference by far the largest reach in the Chatbots 3.x series.  Viewers were able to submit questions for the speakers to the moderator in real-time.  You can view the videos on the aimlinstructor channel.

The organizers of the Chatbots 3.3 conference would like to thank Altify:Software for video streaming support, Universal Arts for camera and audio work, and especially Seed Philly for providing the conference space.  We look forward to seeing you at the next conference!

Sunday, February 17, 2013

Chatbots 3.3 Conference March 23

The sponsors of the Chatbots 3.x conference series are pleased to announce the 4th annual conference on chatbot and virtual assistant technology.  The Chatbots 3.3 conference will be held on March 23, 2013 in Philadelphia, Pennsylvania USA.  This years conference will emphasize all the latest developments in mobile apps and virtual assistant technology.  We’ve already attracted an impressive lineup of speakers, including three former Loebner Prize winners.

Latest updated Speaker roster:

  • Kevin Copple (Loebner Prize Winner 2002), Tianjin Ring-Top Petroleum Manufacturing Co., Ltd., The Role of Chabots and their Authors in the coming Singularity
  • Brad Danenberg, Seed Philly, Venture Capital Opportunities for Chatbot Technology
  • Bill DeSmedt, Alliance Global Services, TBD
  • Mohan Embar (Loebner Prize Winner, 2012), Mohan Embar, Inc., Your Chatbot: Moon Shot or Low-Hanging Fruit?
  • Karsten Fluegge, Pannous, Mobile Apps and Web Services
  • Dave Morton, Geek Cave Creations, Building Chatbots and Apps in a Web-Based World
  • Noah Petherbridge, DreamHost, Chatbots in Web Applications with RiveScript
  • Jeremy Spiegel, MD, Casco Bay Medical, Chatbots in Psychiatry
  • Francis X. Taney, Esq, Stevens & Lee, Legal and IP issues for Botmasters
  • Richard Wallace (Loebner Prize Winner 2000, 2001 and 2004), ALICE AI Foundation, AIML 2.0


blogger templates | Make Money Online