From TREC to Watson: is open domain question answering a solved problem?

CONSTANTIN ORASAN

RESEARCH GROUP IN COMPUTATIONAL LINGUISTICS,

UNIVERSITY OF WOLVERHAMPTON, UKHTTP: / /WWW.WLV.AC.UK/~IN6093/

From TREC to Watson: is open domain question answering a

solved problem?

Constantin Orasan - KEPT 2011

Structure of the talk

4 July 2011

1. Brief introduction to QA

Video 1: Where are we now – IBM Watson

2. The structure of a QA system

Video 2: Watson vs. humans

3. Overview of Watson

4. QA from the point of view of users/companies

5. Conclusions


Information overload

4 July 2011

“Getting information off the Internet is like taking a drink from a fire hydrant”

Mitchell Kapor


What is question answering?

4 July 2011

A way to address the problem of information overload

Question answering aims at identifying the answer to a question posed in natural language in a large collection of documents

The information provided by QA is more focused than information retrieval

The output can be the exact answer or a text snippet which contains the answer

The domain took off as a result of the introduction of QA track in TREC, whilst cross-lingual QA as a result of CLEF


Types of QA systems

4 July 2011

open-domain QA systems: can answer any question from any collection+ can potentially answer any question- very low accuracy (especially in cross-lingual settings)

canned QA systems: rely on a very large repository of questions for which the answer is known+ very little language processing necessary- limited to the answers in the database

closed-domain QA systems: are built for very specific domains and exploit expert knowledge in them+ very high accuracy- can require extensive language processing and limited to one domain


Evolution of QA domain

4 July 2011

Early QA systems date as back as 1960s and were mainly front ends to

databases had limited usability

Open-domain QA emerged as a result of the increasing amount of data available to answer a question need to find and extract the answer developed last 1990s as a result of the QA track at Text

REtrieval Conferences emphasis on factoid questions, but other types of questions

were also explored CLEF competitions have encouraged development of cross-

lingual systems.


Where are we now?

4 July 2011

IBM and the Jeopardy ChallengeJeopardy! is an American quiz show where

participants are given clues and need to guess the question (e.g. if the clue is The Father of Our Country; he didn't really chop down a cherry tree the contestant would respond Who is George Washington?)

Watson is a QA system developed by IBMhttp://

www.youtube.com/watch?v=FC3IryWr4c8

http://www.youtube.com/watch?v=FC3IryWr4c8

http://www.youtube.com/watch?v=FC3IryWr4c8


Structure of an open domain QA system

4 July 2011

A typical open domain QA system consists of: Question processor Document processor Answer extractor (and validation)

Can have components for cross-lingual processing

Has access to several external resources


Question processor

4 July 2011

Produces an interpretation of the question Determines the Question Type (e.g. factoid, definition,

procedure, etc.) Determines the Expected Answer Type (EAT) On the basis of the question it produces a query Determines syntactic and semantic relations between

the words from the questions Expands the query with synonyms May perform translation of the keywords in the query

in the case of cross-lingual QA


Expected answer type calculation

4 July 2011

Relies on the existence of an answer type taxonomyThis taxonomy can be made open-domain by linking

to general ontologies such as WordNetThe EAT can be determined using rule-based as

well as machine learning approaches

Who is the president of Romania?Where is Paris?

Knowledge of domain can greatly improve the identification of EAT and help deal with ambiguities


Query formulation

4 July 2011

Produces a query from the question As a list of keywords As a list of phrases Identifies entities present in the question

Produce variants of the query by introducing morphological, lexical and semantic variations

Domain knowledge is very important for identification of entities and

generation of valid variations and vital in cross-lingual scenarios


Document processing

4 July 2011

Uses the query produced in the previous step to retrieve paragraphs which may contain the answer

It is largely domain independent as it relies on text retrieval engines

Ranks results, but this is largely independent of the QA task

For limited collections of texts it is possible to enrich the index with various linguistic information which can help further processing

When the domain is known, characteristics of the input files can improve the retrieval (e.g. presence of metadata)


Answer extraction

4 July 2011

Uses a variety of techniques to identify the answer of a question

The answer should have the type of EATVery often rely on previously created patterns (e.g.

When was the telephone invented? can be answered if there is a sentence that matches the pattern The telephone was invented in <date>),

Many patterns can express the same answer (e.g. the telephone, invented in <date>)

Relations identified in the question between the expected answer and entities from the question can be exploited by patterns


Answer extraction (II)

4 July 2011

Potential answers are ranked according to functions which are usually learned from the data

The ranking and validation of answers can be done using external sources such as the Internet

QA for well defined domains can rely on better patterns

The functions learned usually work well only on the type of data used for training


Open domain QA - evaluation

4 July 2011

Great coverage, but low accuracyFor example:

EPHYRA QA system in TRAC 2007 reports an accuracy of 0.20 for factoid questions (Schlaefer et al. 2007)

OpenEphyra was used for a cross-lingual Romanian – English QA system and we obtained 0.11 accuracy for factoid questions (Dornescu et al. 2008) – the best performing system for all cross-lingual QA tasks in CLEF 2008

The results are not directly comparable (different QA engines, tuned differently, different collections, different tasks)

But does it make sense to do open domain question answering?


How did Watson perform?

4 July 2011

http://www.youtube.com/watch?v=Puhs2LuO3Zc




How was this achieved?

4 July 2011

Starting point the Practical Intelligent Question Answering Technology (PIQUANT) developed by IBM to participate in TREC

Has been under development at IBM for more than 6 years by a team of 4 full time researchers

Was one of the top three to five in many TRECs

PIQUANT was performing around 0.33 on the TREC data

PIQUANT used a standard architecture for QA


How was this achieved? (II)

4 July 2011

Lots of extra work was put in the system: a core team of 20 researchers working for almost 4 years

PIQUANT system was enriched with a large number of modules for language processing

The processing was parallelised heavilyLots of components were developed to deal with

specific problems (lots of experts)Watson tries to combine deep and shallow

knowledgeHad access to large data sets and very good

hardware


Overview of Watson’s structure

4 July 2011


Hardware used

4 July 2011

Watson is a workload optimized system designed for complex analytics, made possible by integrating massively parallel POWER7 processors and the IBM DeepQA software to answer Jeopardy! questions in under three seconds. Watson is made up of a cluster of ninety IBM Power 750 servers (plus additional I/O, network and cluster controller nodes in 10 racks) with a total of 2880 POWER7 processor cores and 16 Terabytes of RAM. Each Power 750 server uses a 3.5 GHz POWER7 eight core processor, with four threads per core. The POWER7 processor's massively parallel processing capability is an ideal match for Watson's IBM DeepQA software which is embarrassingly parallel (that is a workload that is easily split up into multiple parallel tasks).

According to John Rennie, Watson can process 500 gigabytes, the equivalent of a million books, per second. IBM's master inventor and senior consultant Tony Pearson estimated Watson's hardware cost at about $3 million and with 80 TeraFLOPs would be placed 94th on the Top 500 Supercomputers list.

From: http://en.wikipedia.org/wiki/Watson_(computer)

http://en.wikipedia.org/wiki/Watson_(computer)#cite_note-nytmag-10

http://en.wikipedia.org/wiki/Watson_(computer)#cite_note-nytmag-10


Speed of answer

4 July 2011

In Jeopardy! an answer needs to be provided in 3-5 seconds

In initial experiments with running Watson on a single processor an answer was obtained in about 2 hours

The system was implemented using Apache UIMA Asynchronous Scaleout

Massively parallel architectureIndexes used to answer the questions had to

be pre-processed using Hadoop


Watson was not only NLP

4 July 2011

Betting strategyhttp://www.youtube.com/watch?v=vA9aqAd2iso

http://www.youtube.com/watch?v=vA9aqAd2iso




To sum up, Watson is:

4 July 2011

An amazing engineering projectA massive investmentResearch in many domains of NLPA big PR stuntA way to improve the IBM position in text

analytics

But it is not really a technology ready to be deployed

But was it a real progress in open-domain QA?

So is open domain QA a solved problem?

Can we really solve open domain QA?

Do we really need open domain QA?

Do we care?

QA from user perspective

Real user questions Are rarely open domain Can rarely be formulated in one go Do not always contain answers from only one

source

Companies Have very well defined needs Have access to previously asked questions Need very high accuracy Most of them cannot afford to invest millions of

dollars

The QALL-ME project

Question Answering Learning technologies in a multiLingual and Multimodal Environment (QALL-ME) – FP6 funded project on Multilingual and Multimodal Question Answering FBK, Trento, Italy – coordinator University of Wolverhampton, UK DFKI, Germany University of Alicante, Spain Comdata, Italy Ubiest, Italy Waycom, Italy

http://qallme.fbk.eu Has established an infrastructure for multilingual

and multimodal question answering

http://qallme.fbk.eu/

The QALL-ME project

demonstrators in domain of tourism – can answer questions in the domain of cinema/movies and accommodation.

E.g. What movies can I see in Wolverhampton this

week? How can I get to Novotel Hotel, Wolverhampton?

the questions can be asked in any of the four languages in the consortium

small scale demonstrator built for Romanian


QALL-ME framework

4 July 2011


The QALL-ME ontology

4 July 2011

All the reasoning and processing is done using a domain ontology

The ontology also provides the means of achieving cross-lingual QA

Determines the way data is stored in the database

Ontologies need to be developed for each domain

30

Part of the tourism ontology

MovieShow

Cinema

Movie

TicketPrice

DateTimePeriod

synposis

isInSitehasPrice

hasEventContent

hasPeriod

priceType

priceValue

Director

Star

Producer

Writer

Currency

GPSCoordinate

DirectionLocation

Contact

hasCurrency

TimePeriod

DatePeriod

startTimeendTime

endDate startDate

hasTimePeriod

hasDatePeriod

DirectionLocation

hasSiteFacility

hasContact

hasWriter

hasDirector

hasProducer

genre

name

hasPostalAddress

hasGPSCoordinate

PostalAddress

CinemaRoom

hasRoom

hasStar

certificate

SitePrice

Event

EventContentPeriod

subClassOfsubClassOf

subClassOf

subClassOfsubClassOf

SiteFacility

RoomFacility

hasRoomFacility

name description


Evaluation of the QALL-ME prototype

4 July 2011

For the cinema domain the accuracy ranged between 60% to 85% depending on the language

The system was tested on real questions posed by the users which were completely independent from the ones used to develop the system

The error were mainly caused by wrongly identified named entities, missing patterns and mistakes of the entailment engine

In an commercial environment this system can be revised every day in order to obtain much higher performance


Closed domain QA for commercial companies

4 July 2011

Closed domain QA has a certain appeal with companies

These companies normally have large databases of questions and answers from customers

The domain can be very clearly defined

In some cases the systems needed are actually canned QA systems


Interactive QA

4 July 2011

It is easy to ask:

Where can I eat paella tonight?

but what about:

What mobile phones are smart phones with a camera, have GPS, have touch screen, are produced by HTC or Nokia, but definitely not Samsung, have 1G internal memory ….

Interactivity is needed to deal with ambiguities, wrong answers, user feedback, etc.


Interactive QA

4 July 2011

“process where the user is a continual part of the information loop”

At intersection of Dialogue Systems and Question Answering

Initiate dialogue with the user in cases where there are too many or too few answers, or there is some ambiguity in the request

Can suggest follow up questionsProvide a more natural way of locating a needed

informationCompanies are interested in developing IQA systems

as a way for providing customer services

Hypothetical IQA session

[1] SYS: Hi![2] USER: Can you find me a smart phone with a camera?[3] SYS: There are [NUMBER][big number] of smart phones featuring a

camera, would you like to search for some [BRAND]?[4] USER: No, which have GPS?[5] SYS: [STILL BIG NUMBER] have GPS, are you interested in ones

having TOUCH SCREEN?[Suggesting new constraints][6] USER: Yes, it would be great.[7] SYS: [NUMBER] of Nokia phones, [NUMBER] HTC phones,

[NUMBER] Samsung phones, [NUMBER] of other brands.[8] USER: Ok, what about HTC phones?[9] SYS: [NUMBER] of them have [feature], [NUMBER] of them have

[feature].[10] USER: What are the ones with [feature]?[11] SYS: Here you are: [RESULTS].

26/05/2011Knowledge acquisition from Wikipedia for IQA

35


Answers from more than one source

4 July 2011

Many complex questions need to compose the answer to a question from several sources: List questions:

List all the cantons in Switzerland which border Germany

Sentiment questions:What features people like in Vista?

This is part of the new trend in “deep QA”Even though users probably really need such

answers, the technology is still at the stage of research projects


To sum up …

4 July 2011

Some researchers believe that search is dead and “deep QA” is the future

This was largely fuelled by IBM’s Watson’s winning the Jeopardy!

Watson is a fantastic QA system, but it does not solve the problem of open domain QA

For real applications we still want to focus on very well defined domains

We still want to have the user in the loop to facilitate asking questions

Watson may have revived the interest in QA


Watson is not always right

4 July 2011

but it kind of knows this ….

http://www.youtube.com/watch?v=7h4baBEi0iA




Thank you for your attention

4 July 2011

From TREC to Watson: is open domain question answering a solved problem?

Education

ibm deepqa

domain qa

open domain

qa track

solved problem

factoid questions

domain

ibm