Adap%ve(Methods(for(User1Centered( Organiza%on(of(Music(Collec%ons( - - PowerPoint PPT Presentation

Data & Knowledge Engineering Group

Adap%ve(Methods(for(User1Centered( Organiza%on(of(Music(Collec%ons(

Doctoral(Thesis(Defense(–(Sebas%an(Stober( Magdeburg(|(November(15,(2011(

The(Vision(

! an(intelligent(soIware(to(help(me(

! organize(my(music(collec%on(

((no(simple(structuring(by(meta1data(but(by(similarity)(

! find(music,(I(like(to(listen(to(in(a(specific(moment(

((no(query1result1lists(but(explora%on)(

! How(should(the(soIware(compare(these(songs?(

! melody,(mood,(%mbre,(lyrics,(tempo,(dynamics…( ! mode,(instrumenta%on,(key,(harmonics,(rhythm,(meter(…( ➟ music(has(many(facets(–(How(important(is(each(one?(

➟ How(can(the(soIware(learn(how(I(compare(songs?(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 2(

The(1st(Problem(

listening(example:(

The(Thesis(

!

introduc%on(to(music(informa%on(retrieval((MIR)(

!

state(of(the(art(in(adap%ve(MIR(

!

fundamental(techniques(

!

data1adap%ve(feature(extrac%on(

!

user1adap%ve(genres(

!

context1adap%ve(music(similarity(

!

focus1adap%ve(visualiza%on(

!

bisocia%ve(music(discovery(

!

gaze1controlled(adap%ve(focus(

!

conclusion(&(outlook(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 3(

Music(Informa%on(Retrieval((MIR)(

“the%interdisciplinary%science%% % % % %of%retrieving%informa5on%from%music”%

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 4( “His(Master's(Voice”((Francis(Barraud)(

[wikipedia](

music(data( user(

Music(Informa%on(Retrieval((MIR)(

“the%interdisciplinary%science%% % % % %of%retrieving%informa5on%from%music”%

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 5(

[wikipedia](

core(retrieval(system(

feature( extrac%on( presenta%on(

similarity( preferences( retrieval(model(

structuring( ranking( classifica%on(

index(/(DB(

querying(

(music(data(

(user(

user(interface( data(interface(

a(typical(MIR(system:(

MIR(Challenges([Downie’03](

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 6(

(((((users(of(MIR(systems(have(different(musical(backgrounds( and(have(varying(informa%on(needs(( music(is(mul%1cultural( music(informa%on(has(many(facets(( ((((((((((and(can(be(represented(in(mul%ple(ways(( music((can(be(experienced(in(many(ways(( (((((((((((leading(to(different(percep%ons(

1

& \ \ D . D . Q % \ \ . . E . ! . . Q .

3

E . !

Adap%ve(Systems(

A(system(is((context)(adap5ve%iff(

1) it(behaves(different(in(different(contexts(given(the(same(

input([based(on(Broy(et(al.(‘09]( AND(

2) the(respec%ve(adapta%on((i.e.,(the(difference(in(behavior)(

is(goal1driven(in(that(it(aims(to(op%mize(the(system’s( behavior(in(the(given(context(according(to(some(pre1 defined(measure.(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 7(

Adaptable(➞(Adap%ve(System(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 8(

adaptable system

USER

INPUT OUTPUT

adaptation logic

control parameters evaluate

adaptive system environment

context model context sensing

ADAPTIVE(MUSIC(SIMILARITY(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 9(

core retrieval system

feature extraction presentation

similarity preferences retrieval model

structuring ranking classification

index / DB

querying

music data user

user interface data interface

Goal(/(Problem(Formula%on(

Learn(mul%1facet(music(similarity(measures(( that(reflect(the(user’s(informa%on(need(and(context!(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 10(

Adaptable(Model(of(Similarity(

! objects(of(interest(are(described(by(various(features((

! capture(different(aspects(of(similarity( ! may(not(be(equally(important(for(comparison(

! distance(facet((

=((set(of)(feature(s)(( +(distance(measure(

! non1nega%ve:

(d(a,b)(≥(0([and(d(a,b)(=(0(iff(a=b](

! symmetric:

(d(a,b)(=(d(b,a)(

! op%onally:(

(fulfills(triangle(inequality(

➟ distance(=(weighted(linear(sum(of(facet(distances(

! weights(non1nega%ve,(constant(weight(sum( ! direct((manual)(adapta%on(possible(

(simple(&(understandable(model)(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 11(

"objec%ve( "subjec%ve(

context(model(

control(parameters(

adaptable(system(

OUTPUT( INPUT(

System(Design(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 12(

distance(measure(

USER(

control(facet(weights( evaluate(

adap%ve(system(

rel.(distance(constraints(

user(ac%ons(/( expert(annota%ons(

• bjec%ve(

subjec%ve( adapta%on(logic(

environment(

context(sensing( facet(distances( aggregated(distance(

General(Adapta%on(Approach(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 13(

• p%miza%on(problem:(
• find((valid)(weights(that(

a) sa%sfy(all(constraints( b) minimize(error((#viola%ons)( derive(

rela%ve(distance(constraints( user(ac%ons( classifica%on(problem:(

• learn(linear(classifier(for((

+(training(examples(d(s,a)%<%d(s,b)% –(training(examples(d(s,a)%>%d(s,b)%

• weights(are(defined(by(the(separa%ng(

hyperplane(

[Cheng(&(Hüllermeier(‘08](

as(constraints( as(training(examples(

Facet(Weight(Adapta%on(Approaches(

! Gradient(Descent((op%miza%on)(

! directly(minimizes(error((constraint(viola%ons)( ! problem:(may(get(stuck(in(local(minimum(

! Quadra%c(Programming((op%miza%on)(

! minimizes(weight(change(subject(to((

! hard(weight(bounds(and( ! hard(or(soI(distance(constraints((addi%onal(slack(variables)(

! con%nuity((no(abrupt(changes)(

! Linear(Support(Vector(Machine((classifica%on)(

! maximizes(margin((between(+(and(–(training(examples)( ! favors(“stable”(solu%ons(( ! problem:(soI(weight(constraints(may(be(violated((neg.(weights)(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 14(

Applica%ons(&(Considered(User(Ac%ons(

! Liederenbank([ISMIR’09](

! classifying(Dutch(folk(songs( ➟ class(annota%ons((by(experts)(

! BeatlesExplorer([AMR’08](

! structuring(the(Beatles(dataset( ➟ moving(songs(to(other(cells( ➟ correc%ng(similarity(rankings(

! MusicGalaxy([CMMR/SMC’10](

! exploring(media(collec%ons( ➟ tagging(objects(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 15(

[hvp://www.liederenbank.nl](

FOCUSDADAPTIVE(VISUALIZATION(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 16(

core retrieval system

feature extraction presentation

similarity preferences retrieval model

structuring ranking classification

index / DB

querying

music data user

user interface data interface

Mo%va%on(

! generate(an(overview(of(a(music(collec%on(for(explora%on( ! idea:(use(dimensionality(reduc%on(techniques(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 17(

Islands(of(Music( [Pampalk(et(al.(2003]( MusicMiner( ([Mörchen(et(al.(2005]( SoundBite(for(Songbird( ([Lloyd(2009](

high1dimensional( feature(space( 2D(display(

Focus1Adap%ve(SpringLens(

temporarily(fix(/(highlight(( the(neighborhood(in(focus(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 18(

Focus1Adap%ve(SpringLens*(

! mul%1focus(fish1eye(distor%on(highlights(nearest(neighbors(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 19(

! primary(lens(

! controlled(by(user( ! enlarges(region(of(interest( ! more(space(for(details( ! preserves(context(

! secondary(lenses(

! data1driven( ! highlight(nearest(neighbors(( ! show(“wormholes”( ! neighbors(come(closer(

*based(on(SpringLens(non1linear(distor%on(technique([Germer(et(al.(‘06](

context(model(

control(parameters(

adaptable(system(

OUTPUT( INPUT(

System(Design(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 20(

SpringLens(

USER(

control(lens(parameters( evaluate(

adap%ve(system(

focus(model(

region(of(interest( (primary(lens)(

adapta%on(logic(

environment(

context(sensing( dataset(projec%on( distorted(projec%on(

posi%ons( neighborhoods(

User1Centered(Development(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 21(

Jan’10( Mar’10( Apr’10( May’10( Feb’10(

Evalua%on(Results(

usability(:(comparison(with(tradi%onal(panning(&(zooming( N=30( usefulness(:(%(of(annotated(objects(by(focus(region( N=914((

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 22(

in primary focus at the time of marking.

focus region primary

• ext. primary

secondary none same topic 37.75 4.27 30.74 4.38

• ther topic

4.49 13.24 2.08 no focus 3.06 total 37.75 8.75 43.98 9.52

1 2 3 4 5 6 7 P&Z SL both

helpfulness

1 2 3 4 5 6 7 P&Z SL both

simplicity

1 2 3 4 5 6 7 P&Z SL both

intuitivity

MusicGalaxy(Prototype((Demo)(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 23(

Similarity(Space(+(Linked(Data((Graph)(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 24(

projec%on:( content1based(similarity( nearest(neighbors:( graph(traversal(

Further(Applica%ons(

! gaze1supported(interac%on(

! eye1tracker( ! touch&%lt(device(

! complex(distor%ons(

! gaze1based((aven%on(map)( ! model1driven(

! other(media(types(&(applica%ons(

! images,(video(clips,(text(documents( ! knowledge(base(explora%on( ! travel(recommenda%on(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 25(

[Stellmach(et(al.(2011](

CONCLUSIONS(&(OUTLOOK(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 26(

Main(Thesis(Contribu%ons(

• 1. a(systema%c(overview(on(the(state(of(the(art(in(adap%ve(music(

retrieval((

• 2. pioneering(work(on(automa%c(listening(context(logging(for(

emerging(idiosyncra%c(genres((

• 3. a(general(approach(for(adap%ve(mul%1facet(music(similarity(
• 4. the(focus1adap%ve(SpringLens(visualiza%on(technique((
• 5. two(prototype(applica%ons(as(demonstrators(of(the(proposed(

adap%ve(techniques((

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 27(

Conclusion(

! Has(the(vision(finally(been(realized?(

! not(fully( ! s%ll(a(good(deal(of(soIware(engineering(has(to(be(done(;1)( ! v1.0(at(CeBIT(2012?(

! How(can(long(term(adapta%ons(be(supported?(

! life1long(learning(/(companion( ! model(driI((change(of(preferences)( ! decay(of(constraint(importance(

! Can(we(build(more(expressive(models(of(similarity(without(

sacrificing(simplicity?(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 28(

Some(Ques%ons(for(Future(Work(

THANK(YOU(FOR(( YOUR(ATTENTION!(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 29(

References(

Broy,(M.(et(al.:(“Formalizing(the(no%on(of(adap%ve(system(behavior.”(In:(Proc.%of%the% 2009%ACM%Symposium%on%Applied%Compu5ng%(SAC’09).%2009,(pp.(1029–1033(( Cheng,(W.(and(Hüllermeier,(E.:(“Learning(Similarity(Func%ons(from(Qualita%ve(Feedback.”( In:(Proc.%of%the%9th%European%Conf.%on%Advances%in%CaseJBased%Reasoning%(ECCBR’08).% 2008,(pp.(120–134( Downie,(J.:(“Music(informa%on(retrieval.”(In:(Annual%review%of%informa5on%science%and% technology%37.1((2003),(pp.(295–340(( Germer,(T.(et(al.:(“SpringLens:(Distributed(Nonlinear(Magnifica%ons.”(In:(Eurographics% 2006%J%Short%Papers.%2006,(pp.(123–126(( Lloyd,(S.:(“Automa%c(Playlist(Genera%on(and(Music(Library(Visualisa%on(with(Timbral( Similarity(Measures.”(MA(thesis.(Queen(Mary(University(of(London(( Mörchen,(F.(et(al.:(“Databionic(Visualiza%on(of(Music(Collec%ons(According(to(Perceptual( Distance.”(In:(Proc.(of(ISMIR’05.(2005,(pp.(396–403(( Pampalk,(E.(et(al.:(“Exploring(music(collec%ons(by(browsing(different(views.”(In:(Proc.(of( ISMIR’03.(2003,(pp.(201–208(( Stellmach,(S.(et(al.:(“Designing(Gaze1supported(Mul%modal(Interac%ons(for(the( Explora%on(of(Large(Image(Collec%ons.”(In:(Proc.%of%1st%Int.%Conf.%on%Novel%GazeJ Controlled%Applica5ons%(NGCA’11).(2011,(pp.(1–8(

November(15,(2011( Sebas%an(Stober(1(Adap%ve(Methods(for(User1Centered(Organiza%on(of(Music(Collec%ons(|(Doctoral(Thesis(Defense,(OvGU(Magdeburg( 30(