dialogue modeling 2. indirect requests can i have a cup of coffee? one approach to dealing with...

29
Dialogue Modeling 2

Upload: rhoda-copeland

Post on 19-Jan-2018

215 views

Category:

Documents


0 download

DESCRIPTION

Understanding an Indirect Request Asking if I am able to get a cup of coffee Interested in my ability to perform the action Being able to perform the action is a prerequisite for performing it Must want me to perform that action

TRANSCRIPT

Page 1: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Dialogue Modeling 2

Page 2: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Indirect Requests “Can I have a cup of coffee?”

One approach to dealing with these kinds of requests is by plan-based inference

Page 3: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Understanding anIndirect Request

Asking if I am able to get a cup of coffee

Interested in my ability to perform the action

Being able to perform the action is a prerequisite for performing it

Must want me to perform that action

Page 4: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Extends to other types of indirect requests too

“My soup is cold.”

People don’t like cold soup.

I have cooked the meal, andHave access to the stove

You want me to reheat your soup.

Page 5: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

But requires lots of hand coded rules

E.g. :– All of the inferences rules we discussed

yesterday and for the previous example:

– People don’t like cold soup– Knowledge that putting soup over an

oven will heat it.– Knowing I am able to reheat soup

Page 6: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Instead we can use surface level techniques :

Cue-based Inference andKnowledge Poor Methods

Page 7: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Lecture Outline Lexical cue-based Methods for Interpreting

Dialogue Acts

Finite State Methods– Dialogue Managers– HMMs for Interpreting Dialogue Acts

Other Surface Level Dialogue Managers– Frame Based– Agenda Based

Page 8: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Cue-Based Methodsfor Interpretation

Instead of resolving fully, use surface features– “Can you”, “Please” a Request– “Are you” yes or no question– “Tell me” Request

Page 9: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Context used also S : “Would you like to make an

appointment with John on Tuesday” U : “Yes” agreement

S : “Your appointment is with John on Tuesday.”

U : “Yes” grounding

Page 10: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

How would you build these models?

If you had a marked-up training set…

You could collect all of the sentences which you knew to be a particular dialogue act, then you could train a lexical mode :

– D = argmax_d P(d | W) = argmax_d P(d) P (d | W)

=

Page 11: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

And Context?

Use a finite-state framework

Page 12: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Lecture Outline Lexical cue-based Methods for Interpreting

Dialogue Acts

Finite State Methods– Dialogue Managers– HMMs for Interpreting Dialogue Acts

Other Surface Level Dialogue Managers– Frame Based– Agenda Based

Page 13: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Finite State Machines

a

bb

a

b

a

What is the language accepted by this machine?Equivalent to regular languages.

Is this the minimal machine for this language?

Page 14: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Finite State Dialogue Model

Do you want a pizza?

yes

no

What size?large

medium

small

Bye!

What toppings?

Page 15: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Pros Simple to write Very robust Quick

System must direct entire conversation User actions very limited

Cons

Page 16: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Use Variables to shrink Automaton

Do you want a pizza?

What size?

Size=response

What toppings?

Page 17: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Lecture Outline Lexical cue-based Methods for Interpreting

Dialogue Acts

Finite State Methods– Dialogue Managers– HMMs for Interpreting Dialogue Acts

Other Surface Level Dialogue Managers– Frame Based– Agenda Based

Page 18: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

HMMs Hidden Markov Models

– Probabilistic Finite State Automaton

Are you hungry?

Maybe : .8

Maybe : .2Ok

Would you like a bagel?

Page 19: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

HMMs So after one turn, don’t necessarily

know what state you’re in

Are you hungry?

Maybe : .8

Maybe : .2Ok

Would you like a bagel?

Page 20: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Dialogue Model

opening

closing

suggest constrain

accept reject

.76 .23

.18 .77

.46

.18 .22

.36

.19

.63

.99

Page 21: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

What State Are you In? Combination of what words have been

spoken :– P(E|D) : probability of words being

spoken given that you’re in state D And what a reasonable dialogue looks

like– P(D) : probability of being in a dialogue

state D

Page 22: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

This model provides context

By knowing what state you were previously in, makes it easier to determine what the next state will be.

– Different lexical frequencies in every state

– Different probabilities for each sequences of dialogue states

Page 23: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Lecture Outline Lexical cue-based Methods for Interpreting

Dialogue Acts

Finite State Methods– Dialogue Managers– HMMs for Interpreting Dialogue Acts

Other Surface Level Dialogue Managers– Frame Based– Agenda Based

Page 24: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Frame-Based Dialogue Managers

For appointment scheduler, there are a few distinct things need to be decided

How would you write a finite state machine to get that information?

People : ____Time : _____

Location : _____

Page 25: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

This either :– Imposes a very rigid structure– Neglects useful shared structure

Want to flexibly ask for different pieces of the entire bit of information

Page 26: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Frame Driven Model

Solution : have a model where a frame drives the dialogue

Page 27: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Frame-Driven DialogueUser : I’d like to

schedule an appointment.

System : Who is the other party?

User : John

Calls up appointment frame

People :____Time : ____

Location : ___

People : JohnTime : __

Location : ___

Page 28: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Multiple ResponsesUser : I’d like to

schedule an appointment.

System : Who is the other party?

User : John, sometime on tuesday

Calls up appointment frame

People :____Time : ____

Location : ___

People : JohnTime : TuesdayLocation : ___

Page 29: Dialogue Modeling 2. Indirect Requests Can I have a cup of coffee?  One approach to dealing with these kinds of requests is by plan-based inference

Grounding w/Frames– After each utterance,ground the request

User : John, sometime on TuesdaySystem: Ok where do you want to meet John

on Tuesday?– Alternatively can conform when form is

finishedSystem : Ok, I am scheduling you in Room 332

with John on Tuesday at 4pm