kth.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Efficient Learning in Graphs and in Combinatorial Multi-Armed Bandits
KTH, School of Electrical Engineering and Computer Science (EECS), Computer Science, Theoretical Computer Science, TCS.ORCID iD: 0000-0002-4222-274x
2024 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]

Graphs have rich structures at both local and global scales. By exploiting structural properties in certain graph problems, it is possible to design computationally efficient algorithms or refine performance analysis. This thesis is divided into two parts: (i) designing new methods for discovering structure from graphs and (ii) studying the interplay between graphs and combinatorial multi-arm bandits. They differ in how structures are defined, discovered, and utilized. In part (i), we start with a graph-mining problem on signed networks [TOG20], where the graph patterns we aim to detect are groups with mostly positive intragroup edges and mostly negative inter-group edges. We design our objective such that, given a solution, it reflects how well that solution matches the desired pattern. Our proposed algorithm makes no assumptions about the graph. It demonstrates competitive empirical performance in real-world graphs and synthetic graphs. The performance evaluation is conducted through a worst-case analysis, approximating the optimal solution. Moreover, we extend our conflicting group detection [BGG+19, TOG20] as well as other graph-mining tasks (such as fair densest-subgraph detection [ABF+20], two-community detection [New06]) to a memory-limited and pass-limited setting. Under such a setting, Randomized SVD, which has been proposed by [HMT11], is the most preferable method in the memory-limited and pass-limited setting. However, for an input matrix of size n °ø n, it has no guarantee in the o(ln n)-pass regime, which is of most interest to practitioners. We hence derive a tighter analysis [TWA+22] for Randomized SVD for positive semi-definite matrices in any number of passes and for indefinite matrices under certain conditions. Furthermore, we initiate the study of a mixture of Johnson-Lindenstrauss distribution and the 0/1 Bernoulli distribution. We show that this mixture helps make the detection of 2-conflicting groups [BGG+19]

more efficient. In part (ii), we study combinatorial multi-arm bandits problems, where the graph properties play important roles but are somewhat hidden in the optimization problem. Instead, one derives the fundamental limit bounds on either the expected sample complexity or the expected cumulative regret, typically in the informationtheoretical sense, and then explores the abstract properties associated with those bounds to solve the problem satisfactorily, either statistically or computationally.

We focus on combiantorial semi-bandits where the learner observes individual feedbacks for each arm part of the selection. This formulation models many real-world problems, including online ranking [DKC21] in recommendation systems, network routing [CLK+14] in internet service providers, loan assignment [KWA+14], path planning [JMKK21], and influence marketing [Per22]. For best arm identification with fixed confidence, we propose the first polynomial-time algorithm whose sample complexity is instance-specifically optimal in high confidence regime and has polynomial dependency on the number of arms in moderate confidence regime. For regret minimization, we propose the first algorithm whose per-round time complexity is sublinear in the number of arms while matching the instance-specifically gap-dependent lower bound asymptotically.

Abstract [sv]

Grafer har rika strukturer både på lokal och global skala. Genom att utnyttja strukturella egenskaper i vissa grafproblem är det möjligt att designa beräkningsmässigt effektiva algoritmer eller förfina prestandaanalysen. Denna avhandling är uppdelad i två delar: (i) att designa nya metoder för att upptäcka struktur från grafer och (ii) att studera samspelet mellan grafer och kombinatoriska multi-armbanditer. De skiljer sig .t i hur strukturer definieras, upptäcks och utnyttjas.

I del (i) börjar vi med ett grafgruvningsproblem på signerade nätverk [TOG20], där de grafmönster vi siktar på att upptäcka är grupper med mestadels positiva intra-gruppkanter och mestadels negativa inter-gruppkanter. Vi utformar vårt mål så att, givet en lösning, det återspeglar hur väl den lösningen matchar det önskade mönstret. Vår föreslagna algoritm ger inga antaganden om grafen. Den visar konkurrenskraftig empirisk prestanda i verkliga grafer och syntetiska grafer. Prestandautvärderingen genomförs genom en värsta fall-analys, som approximativt finner den optimala lösningen. Dessutom utvidgar vi vår konfliktgruppsdetektion [BGG+19, TOG20] samt andra grafgruvningsuppgifter (såsom rättvis tätaste subgrafdetektion [ABF+20], två-samhällesdetektion [New06]) till en minnesbegränsad och passbegränsad inställning. Under en sådan inställning är Randomized SVD, som föreslagits av [HMT11], den mest föredragna metoden i en minnesbegränsad och passbegränsad inställning. Men för en ingångsmatris av storlek n × n har den ingen garanti i o(ln n)-passregimen, vilket är av största intresse för praktiker. Vi härleder därför en stramare analys [TWA+22] för Randomized SVD för positivt semi-definita matriser i vilket antal pass som helst och för indefinita matriser under vissa villkor. Vidare initierar vi studien av en blandning av Johnson-Lindenstrauss distribution och 0/1 Bernoulli-distributionen. Vi visar att denna blandning hjälper till att göra detektionen av 2-konfliktgrupper [BGG+19] mer effektiv.

I del (ii) studerar vi kombinatoriska multi-arm banditproblem, där grafer- nas egenskaper spelar viktiga roller men är något dolda i optimeringsproblemet. I stället härleder man de fundamentala gränserna för antingen den förväntade provtagningskomplexiteten eller den förväntade kumulativa ånger, vanligtvis i informations-teoretisk mening, och utforskar sedan de abstrakta egenskaperna associerade med dessa gränser för att lösa problemet på ett tillfredsställande sätt, antingen statistiskt eller beräkningsmässigt. Vi fokuserar på kombinatoriska semi-banditer där läraren observerar individuella återkopplingar för varje arm som är en del av urvalet. Denna formulering modellerar många verkliga problem, inklusive online ranking [ DKC21 ] i rekommendationssystem, nätverksrouting [ CLK+14] hos internetleverantörer, lånefördelning [ KWA+14 ], vägplanering [JMKK21 ], och influencer-marknadsföring [ Per22 ]. För bästa armidentifiering med fast förtroende föreslår vi den första polynomtidalgoritmen vars provtagningskomplexitet är instansspecifikt optimal i hög förtroenderegim och har polynomberoende på antalet armar i måttlig förtroenderegim. För ångerminimering föreslår vi den första algoritmen vars per-rundans tidskomplexitet är sublinjär i antalet armar samtidigt som den matchar den instansspecifikt gapberoende nedre gränsen asymptotiskt.

Place, publisher, year, edition, pages
KTH Royal Institute of Technology, 2024. , p. 172
Series
TRITA-EECS-AVL ; 2024:77
Keywords [en]
graph mining; combinatorial multi-armed bandits
National Category
Computer Systems
Research subject
Computer Science
Identifiers
URN: urn:nbn:se:kth:diva-355999ISBN: 978-91-8106-073-7 (print)OAI: oai:DiVA.org:kth-355999DiVA, id: diva2:1911388
Public defence
2024-12-06, https://kth-se.zoom.us/j/66180403376?pwd=elYurvM09QfBT7I5xvTPEmoPX3rbaF.1, D2, Lindstedtsvägen 9, Stockholm, 13:15 (English)
Opponent
Supervisors
Note

QC 20241107

Available from: 2024-11-07 Created: 2024-11-07 Last updated: 2025-01-27Bibliographically approved
List of papers
1. Discovering conflicting groups in signed networks
Open this publication in new window or tab >>Discovering conflicting groups in signed networks
2020 (English)In: Advances in Neural Information Processing Systems 33 (NeurIPS 2020) / [ed] H. Larochelle and M. Ranzato and R. Hadsell and M.F. Balcan and H. Lin, Neural information processing systems foundation , 2020Conference paper, Published paper (Refereed)
Abstract [en]

Signed networks are graphs where edges are annotated with a positive or negative sign, indicating whether an edge interaction is friendly or antagonistic. Signed networks can be used to study a variety of social phenomena, such as mining polarized discussions in social media, or modeling relations of trust and distrust in online review platforms. In this paper we study the problem of detecting k conflicting groups in a signed network. Our premise is that each group is positively connected internally and negatively connected with the other k − 1 groups. A distinguishing aspect of our formulation is that we are not searching for a complete partition of the signed network; instead, we allow a subset of nodes to be neutral with respect to the conflict structure we are searching. As a result, the problem we tackle differs from previously-studied problems, such as correlation clustering and k-way partitioning. To solve the conflicting-group discovery problem, we derive a novel formulation in which each conflicting group is naturally characterized by the solution to the maximum discrete Rayleigh’s quotient (MAX-DRQ) problem. We present two spectral methods for finding approximate solutions to the MAX-DRQ problem, which we analyze theoretically. Our experimental evaluation shows that, compared to state-of-the-art baselines, our methods find solutions of higher quality, are faster, and recover ground-truth conflicting groups with higher accuracy

Place, publisher, year, edition, pages
Neural information processing systems foundation, 2020
Keywords
Signed network; Spectral analysis; Social network; Polarization
National Category
Computer Sciences Telecommunications
Research subject
Computer Science
Identifiers
urn:nbn:se:kth:diva-291471 (URN)2-s2.0-85102152313 (Scopus ID)
Conference
34th Conference on Neural Information Processing Systems, NeurIPS 2020 Virtual, Online6 December 2020 through 12 December 2020
Projects
REBOUND
Note

QC 20210315

Available from: 2021-03-12 Created: 2021-03-12 Last updated: 2025-01-27Bibliographically approved
2. Improved analysis of randomized SVD for top-eigenvector approximation
Open this publication in new window or tab >>Improved analysis of randomized SVD for top-eigenvector approximation
Show others...
2022 (English)In: International Conference On Artificial Intelligence And Statistics, Vol 151 / [ed] Camps-Valls, G Ruiz, FJR Valera, I, ML Research Press , 2022, Vol. 151Conference paper, Published paper (Refereed)
Abstract [en]

Computing the top eigenvectors of a matrix is a problem of fundamental interest to various fields. While the majority of the literature has focused on analyzing the reconstruction error of low-rank matrices associated with the retrieved eigenvectors, in many applications one is interested in finding one vector with high Rayleigh quotient. In this paper we study the problem of approximating the top-eigenvector. Given a symmetric matrix A with largest eigenvalue lambda(1), our goal is to find a vector (u) over cap that approximates the leading eigenvector u(1) with high accuracy, as measured by the ratio R((u) over cap) = lambda(-1)(1) (u) over cap (T)A (u) over cap/(u) over cap (T)(u) over cap. We present a novel analysis of the randomized SVD algorithm of Halko et al. (2011b) and derive tight bounds in many cases of interest. Notably, this is the first work that provides non-trivial bounds for approximating the ratio R((u) over cap) using randomized SVD with any number of iterations. Our theoretical analysis is complemented with a thorough experimental study that confirms the efficiency and accuracy of the method.

Place, publisher, year, edition, pages
ML Research Press, 2022
Series
Proceedings of Machine Learning Research, ISSN 2640-3498 ; 151
National Category
Other Electrical Engineering, Electronic Engineering, Information Engineering Computational Mathematics Mathematical Analysis
Identifiers
urn:nbn:se:kth:diva-316704 (URN)000828072702005 ()2-s2.0-85163133169 (Scopus ID)
Conference
International Conference on Artificial Intelligence and Statistics, MAR 28-30, 2022, ELECTR NETWORK
Note

QC 20220905

Available from: 2022-09-05 Created: 2022-09-05 Last updated: 2025-01-27Bibliographically approved
3. Closing the Computational-Statistical Gap in Best Arm Identification for Combinatorial Semi-bandits
Open this publication in new window or tab >>Closing the Computational-Statistical Gap in Best Arm Identification for Combinatorial Semi-bandits
2023 (English)In: Advances in Neural Information Processing Systems 36 - 37th Conference on Neural Information Processing Systems, NeurIPS 2023, Neural Information Processing Systems Foundation , 2023Conference paper, Published paper (Refereed)
Abstract [en]

We study the best arm identification problem in combinatorial semi-bandits in the fixed confidence setting. We present Perturbed Frank-Wolfe Sampling (P-FWS), an algorithm that (i) runs in polynomial time, (ii) achieves the instance-specific minimal sample complexity in the high confidence regime, and (iii) enjoys polynomial sample complexity guarantees in the moderate confidence regime. To the best of our knowledge, even for the vanilla bandit problems, no algorithm was able to achieve (ii) and (iii) simultaneously. With P-FWS, we close the computational-statistical gap in best arm identification in combinatorial semi-bandits. The design of P-FWS starts from the optimization problem that defines the information-theoretical and instance-specific sample complexity lower bound. P-FWS solves this problem in an online manner using, in each round, a single iteration of the Frank-Wolfe algorithm. Structural properties of the problem are leveraged to make the P-FWS successive updates computationally efficient. In turn, P-FWS only relies on a simple linear maximization oracle.

Place, publisher, year, edition, pages
Neural Information Processing Systems Foundation, 2023
National Category
Computer Sciences
Identifiers
urn:nbn:se:kth:diva-350250 (URN)001220600003008 ()2-s2.0-85185584821 (Scopus ID)
Conference
37th Conference on Neural Information Processing Systems, NeurIPS 2023, New Orleans, United States of America, Dec 10 2023 - Dec 16 2023
Note

QC 20240710

Available from: 2024-07-10 Created: 2024-07-10 Last updated: 2024-11-07Bibliographically approved
4. Matroid Semi-Bandits in Sublinear Time
Open this publication in new window or tab >>Matroid Semi-Bandits in Sublinear Time
2024 (English)In: Proceedings of the 41 st International Conference on Machine Learning,, 2024Conference paper, Poster (with or without abstract) (Refereed)
Abstract [en]

We study the matroid semi-bandits problem, where at each round the learner plays a subset of K arms from a feasible set, and the goal is to maximize the expected cumulative linear rewards. Existing algorithms have per-round time complexity at least Ω(K), which becomes expensive when K is large. To address this computational issue, we propose FasterCUCB whose sampling rule takes time sublinear in K for common classes of matroids: O(D polylog(K) polylog(T)) for uniform matroids, partition matroids, and graphical matroids, and O(D√ Kpolylog(T)) for transversal matroids. Here, D is the maximum number of elements in any feasible subset of arms, and T is the horizon. Our technique is based on dynamic maintenance of an approximate maximum-weight basis over inner-product weights. Although the introduction of an approximate maximum-weight basis presents a challenge in regret analysis, we can still guarantee an upper bound on regret as tight as CUCB in the sense that it matches the gap-dependent lower bound by Kveton et al. (2014a) asymptotically.

National Category
Computer Sciences
Identifiers
urn:nbn:se:kth:diva-348901 (URN)2-s2.0-85203790933 (Scopus ID)
Conference
The 41 st International Conference on Machine Learning, Vienna, Austria, Sun Jul 21st through Sat Jul 27 th
Note

QC 20240628

Available from: 2024-06-27 Created: 2024-06-27 Last updated: 2024-11-07Bibliographically approved

Open Access in DiVA

fulltext(33028 kB)651 downloads
File information
File name FULLTEXT01.pdfFile size 33028 kBChecksum SHA-512
1497b42739f846547a94e4364107d8d5b2a1e754e88bfce764afcb60a51d4d30a1ddd8cce853b063866e9e90bb143eed398ae8d9108b164c5fca5e27fceef044
Type fulltextMimetype application/pdf

Authority records

Tzeng, Ruo-Chun

Search in DiVA

By author/editor
Tzeng, Ruo-Chun
By organisation
Theoretical Computer Science, TCS
Computer Systems

Search outside of DiVA

GoogleGoogle Scholar
Total: 653 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

isbn
urn-nbn

Altmetric score

isbn
urn-nbn
Total: 976 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf