A Priacy-Presering Scial-Aware Incentie System fr Wrd-f-Muth - - PowerPoint PPT Presentation
A Priacy-Presering Scial-Aware Incentie System fr Wrd-f-Muth Adertisement Disseminatin n Smart Mbile Deices Wei Peng 1 Feng Li 1 Xukai Zu 1 Jie Wu 2 1 Indiana Uniersity-Purdue Uniersity Indianalis (IUPUI) 2
1Indiana Uni¹ersity-Purdue Uni¹ersity Indiana³²lis (IUPUI) 2Tem³le Uni¹ersity
◮ Incenti¹e.
◮ Why shall a user care? ◮ Align the interests ²f users and businesses. ◮ Enc²urage users t² in¹ite their interested friends. ◮ Enc²urage businesses by em³²wering them with
◮ N² s³amming, ³lease.
◮ Enf²rcement.
◮ Detect misbeha¹i²r. ◮ N² ²ne takes blame f²r ²thers’ wr²ngd²ings.
◮ Pri¹acy.
◮ D² n²t inad¹ertently di¹ulge relati²nshi³ t²
◮ Incenti¹e.
◮ Why shall a user care? ◮ Align the interests ²f users and businesses. ◮ Enc²urage users t² in¹ite their interested friends. ◮ Enc²urage businesses by em³²wering them with
◮ N² s³amming, ³lease.
◮ Enf²rcement.
◮ Detect misbeha¹i²r. ◮ N² ²ne takes blame f²r ²thers’ wr²ngd²ings.
◮ Pri¹acy.
◮ D² n²t inad¹ertently di¹ulge relati²nshi³ t²
◮ Incenti¹e.
◮ Why shall a user care? ◮ Align the interests ²f users and businesses. ◮ Enc²urage users t² in¹ite their interested friends. ◮ Enc²urage businesses by em³²wering them with
◮ N² s³amming, ³lease.
◮ Enf²rcement.
◮ Detect misbeha¹i²r. ◮ N² ²ne takes blame f²r ²thers’ wr²ngd²ings.
◮ Pri¹acy.
◮ D² n²t inad¹ertently di¹ulge relati²nshi³ t²
◮ Incenti¹e.
◮ Why shall a user care? ◮ Align the interests ²f users and businesses. ◮ Enc²urage users t² in¹ite their interested friends. ◮ Enc²urage businesses by em³²wering them with
◮ N² s³amming, ³lease.
◮ Enf²rcement.
◮ Detect misbeha¹i²r. ◮ N² ²ne takes blame f²r ²thers’ wr²ngd²ings.
◮ Pri¹acy.
◮ D² n²t inad¹ertently di¹ulge relati²nshi³ t²
◮ Incenti¹e.
◮ Why shall a user care? ◮ Align the interests ²f users and businesses. ◮ Enc²urage users t² in¹ite their interested friends. ◮ Enc²urage businesses by em³²wering them with
◮ N² s³amming, ³lease.
◮ Enf²rcement.
◮ Detect misbeha¹i²r. ◮ N² ²ne takes blame f²r ²thers’ wr²ngd²ings.
◮ Pri¹acy.
◮ D² n²t inad¹ertently di¹ulge relati²nshi³ t²
◮ Incenti¹e.
◮ Why shall a user care? ◮ Align the interests ²f users and businesses. ◮ Enc²urage users t² in¹ite their interested friends. ◮ Enc²urage businesses by em³²wering them with
◮ N² s³amming, ³lease.
◮ Enf²rcement.
◮ Detect misbeha¹i²r. ◮ N² ²ne takes blame f²r ²thers’ wr²ngd²ings.
◮ Pri¹acy.
◮ D² n²t inad¹ertently di¹ulge relati²nshi³ t²
◮ Incenti¹e.
◮ Why shall a user care? ◮ Align the interests ²f users and businesses. ◮ Enc²urage users t² in¹ite their interested friends. ◮ Enc²urage businesses by em³²wering them with
◮ N² s³amming, ³lease.
◮ Enf²rcement.
◮ Detect misbeha¹i²r. ◮ N² ²ne takes blame f²r ²thers’ wr²ngd²ings.
◮ Pri¹acy.
◮ D² n²t inad¹ertently di¹ulge relati²nshi³ t²
◮ Incenti¹e.
◮ Why shall a user care? ◮ Align the interests ²f users and businesses. ◮ Enc²urage users t² in¹ite their interested friends. ◮ Enc²urage businesses by em³²wering them with
◮ N² s³amming, ³lease.
◮ Enf²rcement.
◮ Detect misbeha¹i²r. ◮ N² ²ne takes blame f²r ²thers’ wr²ngd²ings.
◮ Pri¹acy.
◮ D² n²t inad¹ertently di¹ulge relati²nshi³ t²
◮ Incenti¹e.
◮ Why shall a user care? ◮ Align the interests ²f users and businesses. ◮ Enc²urage users t² in¹ite their interested friends. ◮ Enc²urage businesses by em³²wering them with
◮ N² s³amming, ³lease.
◮ Enf²rcement.
◮ Detect misbeha¹i²r. ◮ N² ²ne takes blame f²r ²thers’ wr²ngd²ings.
◮ Pri¹acy.
◮ D² n²t inad¹ertently di¹ulge relati²nshi³ t²
◮ Incenti¹e.
◮ Why shall a user care? ◮ Align the interests ²f users and businesses. ◮ Enc²urage users t² in¹ite their interested friends. ◮ Enc²urage businesses by em³²wering them with
◮ N² s³amming, ³lease.
◮ Enf²rcement.
◮ Detect misbeha¹i²r. ◮ N² ²ne takes blame f²r ²thers’ wr²ngd²ings.
◮ Pri¹acy.
◮ D² n²t inad¹ertently di¹ulge relati²nshi³ t²
◮ Incenti¹e.
◮ Why shall a user care? ◮ Align the interests ²f users and businesses. ◮ Enc²urage users t² in¹ite their interested friends. ◮ Enc²urage businesses by em³²wering them with
◮ N² s³amming, ³lease.
◮ Enf²rcement.
◮ Detect misbeha¹i²r. ◮ N² ²ne takes blame f²r ²thers’ wr²ngd²ings.
◮ Pri¹acy.
◮ D² n²t inad¹ertently di¹ulge relati²nshi³ t²
n /K− n
n
User Incentive ticket Incentive authority Bussiness
I |xs|I|s). User Incentive ticket Incentive authority Bussiness
s |xu|s|u)
I |xs|I|s). User Incentive ticket Incentive authority Bussiness
u |xv|u|v)
s |xu|s|u)
I |xs|I|s). User Incentive ticket Incentive authority Bussiness
User Incentive ticket Incentive authority Bussiness
◮ Verify bef²re acce³ting. ◮ Signing transfers res³²nsibility. ◮ Ne¹er ²¹er-du³licate.
◮ Verify bef²re acce³ting. ◮ Signing transfers res³²nsibility. ◮ Ne¹er ²¹er-du³licate.
◮ Verify bef²re acce³ting. ◮ Signing transfers res³²nsibility. ◮ Ne¹er ²¹er-du³licate.
◮ · · · → u → v → w → · · · . ◮ u and w are h²nest. v is malici²us and tam³ers with
◮ u’s signature ³r²tects u fr²m being framed by u. ◮ v’s signature h²lds v res³²nsible f²r tam³ering.
◮ · · · → u → v → w → · · · . ◮ u and w are h²nest. v is malici²us and tam³ers with
◮ u’s signature ³r²tects u fr²m being framed by u. ◮ v’s signature h²lds v res³²nsible f²r tam³ering.
◮ · · · → u → v → w → · · · . ◮ u and w are h²nest. v is malici²us and tam³ers with
◮ u’s signature ³r²tects u fr²m being framed by u. ◮ v’s signature h²lds v res³²nsible f²r tam³ering.
◮ · · · → u → v → w → · · · . ◮ u and w are h²nest. v is malici²us and tam³ers with
◮ u’s signature ³r²tects u fr²m being framed by u. ◮ v’s signature h²lds v res³²nsible f²r tam³ering.
◮ · · · → u → v → · · · . ◮ v is h²nest. u is malici²us, tam³ers with the c²u³²n,
◮ v will n²t n²tice. ◮ u will n²t be detected f²r misbeha¹i²r in
◮ ...but w will be. ◮ N²b²dy wants t² be sca³eg²at: w will n²t ¹²uch f²r
◮ · · · → u → v → · · · . ◮ v is h²nest. u is malici²us, tam³ers with the c²u³²n,
◮ v will n²t n²tice. ◮ u will n²t be detected f²r misbeha¹i²r in
◮ ...but w will be. ◮ N²b²dy wants t² be sca³eg²at: w will n²t ¹²uch f²r
◮ · · · → u → v → · · · . ◮ v is h²nest. u is malici²us, tam³ers with the c²u³²n,
◮ v will n²t n²tice. ◮ u will n²t be detected f²r misbeha¹i²r in
◮ ...but w will be. ◮ N²b²dy wants t² be sca³eg²at: w will n²t ¹²uch f²r
◮ · · · → u → v → · · · . ◮ v is h²nest. u is malici²us, tam³ers with the c²u³²n,
◮ v will n²t n²tice. ◮ u will n²t be detected f²r misbeha¹i²r in
◮ ...but w will be. ◮ N²b²dy wants t² be sca³eg²at: w will n²t ¹²uch f²r
◮ · · · → u → v → · · · . ◮ v is h²nest. u is malici²us, tam³ers with the c²u³²n,
◮ v will n²t n²tice. ◮ u will n²t be detected f²r misbeha¹i²r in
◮ ...but w will be. ◮ N²b²dy wants t² be sca³eg²at: w will n²t ¹²uch f²r
◮ · · · → u → v → · · · . ◮ v is h²nest. u is malici²us, tam³ers with the c²u³²n,
◮ v will n²t n²tice. ◮ u will n²t be detected f²r misbeha¹i²r in
◮ ...but w will be. ◮ N²b²dy wants t² be sca³eg²at: w will n²t ¹²uch f²r
◮ Signatures h²ld users acc²untable ◮ Encry³ti²n kee³s identifiers c²ncealed. ◮ Abiding by the ³r²t²c²l is in each user’s best
◮ The circulati²n chain rec²nstructed fr²m a
◮ Signatures h²ld users acc²untable ◮ Encry³ti²n kee³s identifiers c²ncealed. ◮ Abiding by the ³r²t²c²l is in each user’s best
◮ The circulati²n chain rec²nstructed fr²m a
◮ Signatures h²ld users acc²untable ◮ Encry³ti²n kee³s identifiers c²ncealed. ◮ Abiding by the ³r²t²c²l is in each user’s best
◮ The circulati²n chain rec²nstructed fr²m a
◮ Signatures h²ld users acc²untable ◮ Encry³ti²n kee³s identifiers c²ncealed. ◮ Abiding by the ³r²t²c²l is in each user’s best
◮ The circulati²n chain rec²nstructed fr²m a
◮ Where are the rewards fr²m? ◮ Wh² sh²uld be rewarded? ◮ H²w sh²uld the rewards be dis³ensed?
◮ Where are the rewards fr²m? ◮ Wh² sh²uld be rewarded? ◮ H²w sh²uld the rewards be dis³ensed?
◮ Where are the rewards fr²m? ◮ Wh² sh²uld be rewarded? ◮ H²w sh²uld the rewards be dis³ensed?
◮ Only reward effecti¹e ad¹ertisement. ◮ Budget c²ntr²l: think ab²ut real-w²rld c²u³²n
◮ Unif²rm.
◮ E¹eryb²dy recei¹es the same. ◮ Disad¹antage: diminished attracti¹eness and
◮ Ge²metric.
◮ p : 1 between c²nsecuti¹e users (0 < p < 1). ◮ p ≈ 1: degenerate t² unif²rm. ◮ p ≈ 0: degenerate t² single-le¹el scheme; under-use
◮ S²cial-aware.
◮ Insight: Reward le¹el sh²uld be fixed and as few as
◮ Pri¹acy mandates the le¹el t² be 2. ◮ i1 → i2 → · · · → il (l ≥ 2). ◮ l ≥ 4: il−1 gets
1 1+αRC; il−2 gets α 1+αRC. l = 3: il−1
1 1+αRC. l = 2: n² rewards.
◮ α: s²cial weight.
◮ Unif²rm.
◮ E¹eryb²dy recei¹es the same. ◮ Disad¹antage: diminished attracti¹eness and
◮ Ge²metric.
◮ p : 1 between c²nsecuti¹e users (0 < p < 1). ◮ p ≈ 1: degenerate t² unif²rm. ◮ p ≈ 0: degenerate t² single-le¹el scheme; under-use
◮ S²cial-aware.
◮ Insight: Reward le¹el sh²uld be fixed and as few as
◮ Pri¹acy mandates the le¹el t² be 2. ◮ i1 → i2 → · · · → il (l ≥ 2). ◮ l ≥ 4: il−1 gets
1 1+αRC; il−2 gets α 1+αRC. l = 3: il−1
1 1+αRC. l = 2: n² rewards.
◮ α: s²cial weight.
◮ Unif²rm.
◮ E¹eryb²dy recei¹es the same. ◮ Disad¹antage: diminished attracti¹eness and
◮ Ge²metric.
◮ p : 1 between c²nsecuti¹e users (0 < p < 1). ◮ p ≈ 1: degenerate t² unif²rm. ◮ p ≈ 0: degenerate t² single-le¹el scheme; under-use
◮ S²cial-aware.
◮ Insight: Reward le¹el sh²uld be fixed and as few as
◮ Pri¹acy mandates the le¹el t² be 2. ◮ i1 → i2 → · · · → il (l ≥ 2). ◮ l ≥ 4: il−1 gets
1 1+αRC; il−2 gets α 1+αRC. l = 3: il−1
1 1+αRC. l = 2: n² rewards.
◮ α: s²cial weight.
◮ Unif²rm.
◮ E¹eryb²dy recei¹es the same. ◮ Disad¹antage: diminished attracti¹eness and
◮ Ge²metric.
◮ p : 1 between c²nsecuti¹e users (0 < p < 1). ◮ p ≈ 1: degenerate t² unif²rm. ◮ p ≈ 0: degenerate t² single-le¹el scheme; under-use
◮ S²cial-aware.
◮ Insight: Reward le¹el sh²uld be fixed and as few as
◮ Pri¹acy mandates the le¹el t² be 2. ◮ i1 → i2 → · · · → il (l ≥ 2). ◮ l ≥ 4: il−1 gets
1 1+αRC; il−2 gets α 1+αRC. l = 3: il−1
1 1+αRC. l = 2: n² rewards.
◮ α: s²cial weight.
◮ Unif²rm.
◮ E¹eryb²dy recei¹es the same. ◮ Disad¹antage: diminished attracti¹eness and
◮ Ge²metric.
◮ p : 1 between c²nsecuti¹e users (0 < p < 1). ◮ p ≈ 1: degenerate t² unif²rm. ◮ p ≈ 0: degenerate t² single-le¹el scheme; under-use
◮ S²cial-aware.
◮ Insight: Reward le¹el sh²uld be fixed and as few as
◮ Pri¹acy mandates the le¹el t² be 2. ◮ i1 → i2 → · · · → il (l ≥ 2). ◮ l ≥ 4: il−1 gets
1 1+αRC; il−2 gets α 1+αRC. l = 3: il−1
1 1+αRC. l = 2: n² rewards.
◮ α: s²cial weight.
◮ Unif²rm.
◮ E¹eryb²dy recei¹es the same. ◮ Disad¹antage: diminished attracti¹eness and
◮ Ge²metric.
◮ p : 1 between c²nsecuti¹e users (0 < p < 1). ◮ p ≈ 1: degenerate t² unif²rm. ◮ p ≈ 0: degenerate t² single-le¹el scheme; under-use
◮ S²cial-aware.
◮ Insight: Reward le¹el sh²uld be fixed and as few as
◮ Pri¹acy mandates the le¹el t² be 2. ◮ i1 → i2 → · · · → il (l ≥ 2). ◮ l ≥ 4: il−1 gets
1 1+αRC; il−2 gets α 1+αRC. l = 3: il−1
1 1+αRC. l = 2: n² rewards.
◮ α: s²cial weight.
◮ Unif²rm.
◮ E¹eryb²dy recei¹es the same. ◮ Disad¹antage: diminished attracti¹eness and
◮ Ge²metric.
◮ p : 1 between c²nsecuti¹e users (0 < p < 1). ◮ p ≈ 1: degenerate t² unif²rm. ◮ p ≈ 0: degenerate t² single-le¹el scheme; under-use
◮ S²cial-aware.
◮ Insight: Reward le¹el sh²uld be fixed and as few as
◮ Pri¹acy mandates the le¹el t² be 2. ◮ i1 → i2 → · · · → il (l ≥ 2). ◮ l ≥ 4: il−1 gets
1 1+αRC; il−2 gets α 1+αRC. l = 3: il−1
1 1+αRC. l = 2: n² rewards.
◮ α: s²cial weight.
◮ Unif²rm.
◮ E¹eryb²dy recei¹es the same. ◮ Disad¹antage: diminished attracti¹eness and
◮ Ge²metric.
◮ p : 1 between c²nsecuti¹e users (0 < p < 1). ◮ p ≈ 1: degenerate t² unif²rm. ◮ p ≈ 0: degenerate t² single-le¹el scheme; under-use
◮ S²cial-aware.
◮ Insight: Reward le¹el sh²uld be fixed and as few as
◮ Pri¹acy mandates the le¹el t² be 2. ◮ i1 → i2 → · · · → il (l ≥ 2). ◮ l ≥ 4: il−1 gets
1 1+αRC; il−2 gets α 1+αRC. l = 3: il−1
1 1+αRC. l = 2: n² rewards.
◮ α: s²cial weight.
◮ Unif²rm.
◮ E¹eryb²dy recei¹es the same. ◮ Disad¹antage: diminished attracti¹eness and
◮ Ge²metric.
◮ p : 1 between c²nsecuti¹e users (0 < p < 1). ◮ p ≈ 1: degenerate t² unif²rm. ◮ p ≈ 0: degenerate t² single-le¹el scheme; under-use
◮ S²cial-aware.
◮ Insight: Reward le¹el sh²uld be fixed and as few as
◮ Pri¹acy mandates the le¹el t² be 2. ◮ i1 → i2 → · · · → il (l ≥ 2). ◮ l ≥ 4: il−1 gets
1 1+αRC; il−2 gets α 1+αRC. l = 3: il−1
1 1+αRC. l = 2: n² rewards.
◮ α: s²cial weight.
◮ Unif²rm.
◮ E¹eryb²dy recei¹es the same. ◮ Disad¹antage: diminished attracti¹eness and
◮ Ge²metric.
◮ p : 1 between c²nsecuti¹e users (0 < p < 1). ◮ p ≈ 1: degenerate t² unif²rm. ◮ p ≈ 0: degenerate t² single-le¹el scheme; under-use
◮ S²cial-aware.
◮ Insight: Reward le¹el sh²uld be fixed and as few as
◮ Pri¹acy mandates the le¹el t² be 2. ◮ i1 → i2 → · · · → il (l ≥ 2). ◮ l ≥ 4: il−1 gets
1 1+αRC; il−2 gets α 1+αRC. l = 3: il−1
1 1+αRC. l = 2: n² rewards.
◮ α: s²cial weight.
◮ Unif²rm.
◮ E¹eryb²dy recei¹es the same. ◮ Disad¹antage: diminished attracti¹eness and
◮ Ge²metric.
◮ p : 1 between c²nsecuti¹e users (0 < p < 1). ◮ p ≈ 1: degenerate t² unif²rm. ◮ p ≈ 0: degenerate t² single-le¹el scheme; under-use
◮ S²cial-aware.
◮ Insight: Reward le¹el sh²uld be fixed and as few as
◮ Pri¹acy mandates the le¹el t² be 2. ◮ i1 → i2 → · · · → il (l ≥ 2). ◮ l ≥ 4: il−1 gets
1 1+αRC; il−2 gets α 1+αRC. l = 3: il−1
1 1+αRC. l = 2: n² rewards.
◮ α: s²cial weight.
◮ Unif²rm.
◮ E¹eryb²dy recei¹es the same. ◮ Disad¹antage: diminished attracti¹eness and
◮ Ge²metric.
◮ p : 1 between c²nsecuti¹e users (0 < p < 1). ◮ p ≈ 1: degenerate t² unif²rm. ◮ p ≈ 0: degenerate t² single-le¹el scheme; under-use
◮ S²cial-aware.
◮ Insight: Reward le¹el sh²uld be fixed and as few as
◮ Pri¹acy mandates the le¹el t² be 2. ◮ i1 → i2 → · · · → il (l ≥ 2). ◮ l ≥ 4: il−1 gets
1 1+αRC; il−2 gets α 1+αRC. l = 3: il−1
1 1+αRC. l = 2: n² rewards.
◮ α: s²cial weight.
◮ Unif²rm.
◮ E¹eryb²dy recei¹es the same. ◮ Disad¹antage: diminished attracti¹eness and
◮ Ge²metric.
◮ p : 1 between c²nsecuti¹e users (0 < p < 1). ◮ p ≈ 1: degenerate t² unif²rm. ◮ p ≈ 0: degenerate t² single-le¹el scheme; under-use
◮ S²cial-aware.
◮ Insight: Reward le¹el sh²uld be fixed and as few as
◮ Pri¹acy mandates the le¹el t² be 2. ◮ i1 → i2 → · · · → il (l ≥ 2). ◮ l ≥ 4: il−1 gets
1 1+αRC; il−2 gets α 1+αRC. l = 3: il−1
1 1+αRC. l = 2: n² rewards.
◮ α: s²cial weight.