Loading presentation...

Present Remotely

Send the link below via email or IM


Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.


Apple's Siri and the Semantic Web

University of Aberdeen - CS5010 - 2012

Andres Rodriguez Guapacha

on 30 October 2012

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Apple's Siri and the Semantic Web

Apple Siri and
the Semantic Web Use of Automatic speech recognition to transcribe human speech into text.
Using natural language processing (part of speech tagging, noun-phrase chunking, dependency & constituent parsing) to translate transcribed text into "parsed text"
Uses question and intent analysis to analyze parsed text, detecting user commands and actions. Functions of Siri An Example of Bad Language Difficult for Siri to process different accents Compares speech against a statistical model to estimate based on the sound you spoke, the order which is spoken and the letters which constitute it.
Many others language models are used to comprehend language Process of Siri Application at Server Based on the probability of sets
Documents are ranked based on the probability that generate the query
Best/Partial match How Language Model is used mathematically? Using data mashup technologies to interface with 3rd-party web services such as OpenTable, WolframAlpha, to perform actions, search operations, and question answering
Transforming output of 3rd party web services back into natural language text (eg. Today's weather report -> "The weather will be sunny")
Using TTS (text-to-speech) technologies to transform the natural language text from above into synthesized speech. Functions of Siri Level contd… Siri Speech Recognition and
Natural Language Processing Siri with Natural Language Processing Siri is an application that combines Voice Recognition with Natural Language Processing.
Siri tries to understand not just words but the intentions behind them. Process in Siri Click on Siri button and it loads the phrases of the application
Prompts the user to provide speech input to the Siri application software.
User speaks to the application
Sound of the Speech is compacted into Compact Digital signal
Streamed to recognizer
Recognizer returns a number of potential results after comparing and processing against a vocabulary. It uses your information.

Learn about your key place, Home, Office etc.

Uses Location-base Reminders

Learn about the key people in your life, Sisters, Brothers, Father, Mother etc. How Siri knows you?
Siri works with almost all your built-in applications.

Maps, Sports, Movies, Local Search, Post on FaceBook, FaceTime, Phone, Mail, Web Search and more. Applications Siri work with?
Make a Call.
find a business.
Get directions.
Schedule reminders and meetings
Search the web.
And more What types of things you ask Siri to do?
Speak to Siri as you would to a person — in a natural voice with a conversational tone.

E.g “Does it look like rain tomorrow?”.

No matter how you ask, Siri will tell you the forecast. Siri Understands! When you finish speaking Siri does the following:

Displays the text of what you said.

Provides a response.

It will also ask you a question about what is not clear. What happens after you ask Siri Questions

There’s more than one way to talk to Siri.

Headphones and Remote Microphone .

Bluetooth headsets. How to ask Siri something Intelligent personal assistant
Allow use of voice to send messages, schedule meetings, place phone calls and more.
Is not like traditional voice recognition software .
Understands your natural speech. What is Siri? What is Siri?
Using Siri HOW SIRI WORKS Siri understands and can speak many Languages in the world which includes:
- U.S (English, Spanish)
- U.K (English)
- Australia (English)
- France (French) and other more. Languages Support and Availability Siri works right out of the box.
learn about your accent.
It uses voice recognition algorithms.
Siri exposed to more variations of language as people use it.
It uses information from contacts, music library, calendars, and reminders.
. Does Siri work out of the box, or do you have to teach it? Avoid sending all the speech to Apple's servers
Support more accents and languages.
Add visual component that make the system to move from audio to visual-audio speech recognition.
Add new function that enable the user to provide an immediate and real time feed-back of the command. Suggestions to improve the performance Limited capability for the user in following up one request with another.

The dictation feature of Siri requires the communication with the server to analyse the speech and send back the text to the phone

The accuracy of the dictation function can be affected by irregular background noise, wind and the variable distance from speaker’s mouth to microphone. Limitations of Apple Siri Privacy and security concern

How to solve problems against natural language processing.

Restricted to support few accents and languages Limitations of Apple Siri Limitations of Apple Siri P(B|A) * P(A) P(B) P(A|B) = Amna Tannaf Al Saadi
Samuel Thampy
Ahmad Idris Tambuwal
Andres Rodriguez Guapacha From Siri to the Semantic Web "magical" NLP servers Semantic Web Stack Incorporates concepts from some CALO projects http://obitko.com/tutorials/ontologies-semantic-web/semantic-web-architecture.html Hipertext Web Standard Semantic Web Other Semantic Web technologies Find me a sushi restaurant select ?restaurant
from <...>
where { ... } {name:Kama Sushi, name:Live Sushi, ...} This is where Siri comes to play! Existing and known technologies:
URI: Helps identify different sources
Unicode: Keeps consistency across the web
XML: Fundamental for structured data
XMLNS: Helps using info from different sources Systems that help define the semantical meaning of the web:

"What is Sushi?" RDF: Semantical representation
RDFS / OWL: Groups RDF information around real-world concepts (ontologies, like "Food", "Restaurant", "Japan")
SPARQL / Semantic Web Services: Extraction mechanisms Still have to be approved / defined as a standard RIF / SWRL: Methods for describing the logic of inference in an ontology Cryptography: Source verification Trust: Not only to verify sources, but also to verify information and how that information was processed and if its relevant to the search User Interface / Other apps: Siri, source selection mechanisms, etc http://www.restaurants.co.uk/ http://www.hollywoodrestaurants.com/ {http://www.restaurants.co.uk/, <xml>...</xml>} {foaf:Person "Zooey", foaf:based_near:"Hollywood", ...}
Full transcript