Showing 1–50 of 68 results for author: Vo**ović, M

Using graph neural networks to reconstruct charged pion showers in the CMS High Granularity Calorimeter

Authors: M. Aamir, B. Acar, G. Adamov, T. Adams, C. Adloff, S. Afanasiev, C. Agrawal, C. Agrawal, A. Ahmad, H. A. Ahmed, S. Akbar, N. Akchurin, B. Akgul, B. Akgun, R. O. Akpinar, E. Aktas, A. AlKadhim, V. Alexakhin, J. Alimena, J. Alison, A. Alpana, W. Alshehri, P. Alvarez Dominguez, M. Alyari, C. Amendola , et al. (550 additional authors not shown)

Abstract: A novel method to reconstruct the energy of hadronic showers in the CMS High Granularity Calorimeter (HGCAL) is presented. The HGCAL is a sampling calorimeter with very fine transverse and longitudinal granularity. The active media are silicon sensors and scintillator tiles readout by SiPMs and the absorbers are a combination of lead and Cu/CuW in the electromagnetic section, and steel in the hadr… ▽ More A novel method to reconstruct the energy of hadronic showers in the CMS High Granularity Calorimeter (HGCAL) is presented. The HGCAL is a sampling calorimeter with very fine transverse and longitudinal granularity. The active media are silicon sensors and scintillator tiles readout by SiPMs and the absorbers are a combination of lead and Cu/CuW in the electromagnetic section, and steel in the hadronic section. The shower reconstruction method is based on graph neural networks and it makes use of a dynamic reduction network architecture. It is shown that the algorithm is able to capture and mitigate the main effects that normally hinder the reconstruction of hadronic showers using classical reconstruction methods, by compensating for fluctuations in the multiplicity, energy, and spatial distributions of the shower's constituents. The performance of the algorithm is evaluated using test beam data collected in 2018 prototype of the CMS HGCAL accompanied by a section of the CALICE AHCAL prototype. The capability of the method to mitigate the impact of energy leakage from the calorimeter is also demonstrated. △ Less

Submitted 30 June, 2024; v1 submitted 17 June, 2024; originally announced June 2024.

Comments: Prepared for submission to JINST

An Adaptive Approach for Infinitely Many-armed Bandits under Generalized Rotting Constraints

Authors: Jung-hun Kim, Milan Vojnovic, Se-Young Yun

Abstract: In this study, we consider the infinitely many-armed bandit problems in a rested rotting setting, where the mean reward of an arm may decrease with each pull, while otherwise, it remains unchanged. We explore two scenarios regarding the rotting of rewards: one in which the cumulative amount of rotting is bounded by $V_T$, referred to as the slow-rotting case, and the other in which the cumulative… ▽ More In this study, we consider the infinitely many-armed bandit problems in a rested rotting setting, where the mean reward of an arm may decrease with each pull, while otherwise, it remains unchanged. We explore two scenarios regarding the rotting of rewards: one in which the cumulative amount of rotting is bounded by $V_T$, referred to as the slow-rotting case, and the other in which the cumulative number of rotting instances is bounded by $S_T$, referred to as the abrupt-rotting case. To address the challenge posed by rotting rewards, we introduce an algorithm that utilizes UCB with an adaptive sliding window, designed to manage the bias and variance trade-off arising due to rotting rewards. Our proposed algorithm achieves tight regret bounds for both slow and abrupt rotting scenarios. Lastly, we demonstrate the performance of our algorithm using numerical experiments. △ Less

Submitted 24 May, 2024; v1 submitted 22 April, 2024; originally announced April 2024.

Symmetry breaking mechanisms of the 3BF action for the Standard Model coupled to gravity

Authors: Pavle Stipsic, Marko Vo**ovic

Abstract: We study the details of the explicit and spontaneous symmetry breaking of the constrained 3BF action representing the Standard Model coupled to Einstein-Cartan gravity. First we discuss how each particular constraint breaks the original symmetry of the topological 3BF action. Then we investigate the spontaneous symmetry breaking and the Higgs mechanism for the electroweak theory in the constrained… ▽ More We study the details of the explicit and spontaneous symmetry breaking of the constrained 3BF action representing the Standard Model coupled to Einstein-Cartan gravity. First we discuss how each particular constraint breaks the original symmetry of the topological 3BF action. Then we investigate the spontaneous symmetry breaking and the Higgs mechanism for the electroweak theory in the constrained 3BF form, in order to demonstrate that they can indeed be performed in the framework of higher gauge theory. A formulation of the Proca action as a constrained 3BF theory is also studied in detail. △ Less

Submitted 27 February, 2024; originally announced February 2024.

Comments: 29 pages

On the Convergence of Loss and Uncertainty-based Active Learning Algorithms

Authors: Daniel Haimovich, Dima Karamshuk, Fridolin Linder, Niek Tax, Milan Vojnovic

Abstract: We investigate the convergence rates and data sample sizes required for training a machine learning model using a stochastic gradient descent (SGD) algorithm, where data points are sampled based on either their loss value or uncertainty value. These training methods are particularly relevant for active learning and data subset selection problems. For SGD with a constant step size update, we presen… ▽ More We investigate the convergence rates and data sample sizes required for training a machine learning model using a stochastic gradient descent (SGD) algorithm, where data points are sampled based on either their loss value or uncertainty value. These training methods are particularly relevant for active learning and data subset selection problems. For SGD with a constant step size update, we present convergence results for linear classifiers and linearly separable datasets using squared hinge loss and similar training loss functions. Additionally, we extend our analysis to more general classifiers and datasets, considering a wide range of loss-based sampling strategies and smooth convex training loss functions. We propose a novel algorithm called Adaptive-Weight Sampling (AWS) that utilizes SGD with an adaptive step size that achieves stochastic Polyak's step size in expectation. We establish convergence rate results for AWS for smooth convex training loss functions. Our numerical experiments demonstrate the efficiency of AWS on various datasets by using either exact or estimated loss values. △ Less

Submitted 11 June, 2024; v1 submitted 21 December, 2023; originally announced December 2023.

Challenges for extensions of the process matrix formalism to quantum field theory

Authors: Nikola Paunkovic, Marko Vo**ovic

Abstract: We discuss the issues with tentative generalisations of the process matrix formalism from finite-dimensional mechanical systems all the way to quantum field theory. We present a detailed overview of possible open problems that arise when one attempts to move from particle ontology into the realm of field ontology, i.e., when one transitions from mechanics to field theory framework. These issues ne… ▽ More We discuss the issues with tentative generalisations of the process matrix formalism from finite-dimensional mechanical systems all the way to quantum field theory. We present a detailed overview of possible open problems that arise when one attempts to move from particle ontology into the realm of field ontology, i.e., when one transitions from mechanics to field theory framework. These issues need to be addressed, and problems solved, if one aims to expand the scope of applicability of the process matrix formalism, and therefore its usefulness. This is far from a trivial and straightforward endeavour, but rather a task for a whole future research programme. △ Less

Submitted 6 October, 2023; originally announced October 2023.

Comments: 6 pages, 4 figures

Combinatorial Bandits for Maximum Value Reward Function under Max Value-Index Feedback

Authors: Yiliu Wang, Wei Chen, Milan Vojnović

Abstract: We consider a combinatorial multi-armed bandit problem for maximum value reward function under maximum value and index feedback. This is a new feedback structure that lies in between commonly studied semi-bandit and full-bandit feedback structures. We propose an algorithm and provide a regret bound for problem instances with stochastic arm outcomes according to arbitrary distributions with finite… ▽ More We consider a combinatorial multi-armed bandit problem for maximum value reward function under maximum value and index feedback. This is a new feedback structure that lies in between commonly studied semi-bandit and full-bandit feedback structures. We propose an algorithm and provide a regret bound for problem instances with stochastic arm outcomes according to arbitrary distributions with finite supports. The regret analysis rests on considering an extended set of arms, associated with values and probabilities of arm outcomes, and applying a smoothness condition. Our algorithm achieves a $O((k/Δ)\log(T))$ distribution-dependent and a $\tilde{O}(\sqrt{T})$ distribution-independent regret where $k$ is the number of arms selected in each round, $Δ$ is a distribution-dependent reward gap and $T$ is the horizon time. Perhaps surprisingly, the regret bound is comparable to previously-known bound under more informative semi-bandit feedback. We demonstrate the effectiveness of our algorithm through experimental results. △ Less

Submitted 25 May, 2023; originally announced May 2023.

Henneaux-Teitelboim gauge symmetry and its applications to higher gauge theories

Authors: Mihailo Djordjevic, Tijana Radenkovic, Pavle Stipsic, Marko Vo**ovic

Abstract: When discussing the gauge symmetries of any theory, the Henneaux-Teitelboim transformations are often underappreciated or even completely ignored, due to their on-shell triviality. Nevertheless, these gauge transformations play an important role in understanding the structure of the full gauge symmetry group of any theory, especially regarding the subgroup of diffeomorphisms. We give a review of t… ▽ More When discussing the gauge symmetries of any theory, the Henneaux-Teitelboim transformations are often underappreciated or even completely ignored, due to their on-shell triviality. Nevertheless, these gauge transformations play an important role in understanding the structure of the full gauge symmetry group of any theory, especially regarding the subgroup of diffeomorphisms. We give a review of the Henneaux-Teitelboim transformations and the resulting gauge group in the general case, and then discuss its role in the applications to the class of topological theories called nBF models, relevant for the constructions of higher gauge theories and quantum gravity. △ Less

Submitted 10 June, 2023; v1 submitted 28 April, 2023; originally announced May 2023.

Comments: v3: published version, 18 pages

Journal ref: Universe 9, 281 (2023)

Doubly Adversarial Federated Bandits

Authors: Jialin Yi, Milan Vojnović

Abstract: We study a new non-stochastic federated multi-armed bandit problem with multiple agents collaborating via a communication network. The losses of the arms are assigned by an oblivious adversary that specifies the loss of each arm not only for each time step but also for each agent, which we call ``doubly adversarial". In this setting, different agents may choose the same arm in the same time step b… ▽ More We study a new non-stochastic federated multi-armed bandit problem with multiple agents collaborating via a communication network. The losses of the arms are assigned by an oblivious adversary that specifies the loss of each arm not only for each time step but also for each agent, which we call ``doubly adversarial". In this setting, different agents may choose the same arm in the same time step but observe different feedback. The goal of each agent is to find a globally best arm in hindsight that has the lowest cumulative loss averaged over all agents, which necessities the communication among agents. We provide regret lower bounds for any federated bandit algorithm under different settings, when agents have access to full-information feedback, or the bandit feedback. For the bandit feedback setting, we propose a near-optimal federated bandit algorithm called FEDEXP3. Our algorithm gives a positive answer to an open question proposed in Cesa-Bianchi et al. (2016): FEDEXP3 can guarantee a sub-linear regret without exchanging sequences of selected arm identities or loss sequences among agents. We also provide numerical evaluations of our algorithm to validate our theoretical results and demonstrate its effectiveness on synthetic and real-world datasets △ Less

Submitted 21 October, 2023; v1 submitted 22 January, 2023; originally announced January 2023.

Comments: Published in ICML 2023 https://proceedings.mlr.press/v202/yi23a.html

Journal ref: Proceedings of the 40th International Conference on Machine Learning 2023

On Regret-optimal Cooperative Nonstochastic Multi-armed Bandits

Authors: Jialin Yi, Milan Vojnović

Abstract: We consider the nonstochastic multi-agent multi-armed bandit problem with agents collaborating via a communication network with delays. We show a lower bound for individual regret of all agents. We show that with suitable regularizers and communication protocols, a collaborative multi-agent \emph{follow-the-regularized-leader} (FTRL) algorithm has an individual regret upper bound that matches the… ▽ More We consider the nonstochastic multi-agent multi-armed bandit problem with agents collaborating via a communication network with delays. We show a lower bound for individual regret of all agents. We show that with suitable regularizers and communication protocols, a collaborative multi-agent \emph{follow-the-regularized-leader} (FTRL) algorithm has an individual regret upper bound that matches the lower bound up to a constant factor when the number of arms is large enough relative to degrees of agents in the communication graph. We also show that an FTRL algorithm with a suitable regularizer is regret optimal with respect to the scaling with the edge-delay parameter. We present numerical experiments validating our theoretical results and demonstrate cases when our algorithms outperform previously proposed algorithms. △ Less

Submitted 21 October, 2023; v1 submitted 30 November, 2022; originally announced November 2022.

Comments: Published in AAMAS 2023

Journal ref: Proceedings of the 2023 International Conference on Autonomous Agents and Multiagent Systems 1329 1335

Dynamics and Inference for Voter Model Processes

Authors: Milan Vojnovic, Kaifang Zhou

Abstract: We consider a discrete-time voter model process on a set of nodes, each being in one of two states, either 0 or 1. In each time step, each node adopts the state of a randomly sampled neighbor according to sampling probabilities, referred to as node interaction parameters. We study the maximum likelihood estimation of the node interaction parameters from observed node states for a given number of r… ▽ More We consider a discrete-time voter model process on a set of nodes, each being in one of two states, either 0 or 1. In each time step, each node adopts the state of a randomly sampled neighbor according to sampling probabilities, referred to as node interaction parameters. We study the maximum likelihood estimation of the node interaction parameters from observed node states for a given number of realizations of the voter model process. In contrast to previous work on parameter estimation of network autoregressive processes, whose long-run behavior is according to a stationary stochastic process, the voter model is an absorbing stochastic process that eventually reaches a consensus state. This requires develo** a framework for deriving parameter estimation error bounds from observations consisting of several realizations of a voter model process. We present parameter estimation error bounds by interpreting the observation data as being generated according to an extended voter process that consists of cycles, each corresponding to a realization of the voter model process until absorption to a consensus state. In order to obtain these results, consensus time of a voter model process plays an important role. We present new bounds for all moments and a bound that holds with any given probability for consensus time, which may be of independent interest. In contrast to most existing work, our results yield a consensus time bound that holds with high probability. We also present a sampling complexity lower bound for parameter estimation within a prescribed error tolerance for the class of locally stable estimators. △ Less

Submitted 24 November, 2022; originally announced November 2022.

Performance of the CMS High Granularity Calorimeter prototype to charged pion beams of 20$-$300 GeV/c

Authors: B. Acar, G. Adamov, C. Adloff, S. Afanasiev, N. Akchurin, B. Akgün, M. Alhusseini, J. Alison, J. P. Figueiredo de sa Sousa de Almeida, P. G. Dias de Almeida, A. Alpana, M. Alyari, I. Andreev, U. Aras, P. Aspell, I. O. Atakisi, O. Bach, A. Baden, G. Bakas, A. Bakshi, S. Banerjee, P. DeBarbaro, P. Bargassa, D. Barney, F. Beaudette , et al. (435 additional authors not shown)

Abstract: The upgrade of the CMS experiment for the high luminosity operation of the LHC comprises the replacement of the current endcap calorimeter by a high granularity sampling calorimeter (HGCAL). The electromagnetic section of the HGCAL is based on silicon sensors interspersed between lead and copper (or copper tungsten) absorbers. The hadronic section uses layers of stainless steel as an absorbing med… ▽ More The upgrade of the CMS experiment for the high luminosity operation of the LHC comprises the replacement of the current endcap calorimeter by a high granularity sampling calorimeter (HGCAL). The electromagnetic section of the HGCAL is based on silicon sensors interspersed between lead and copper (or copper tungsten) absorbers. The hadronic section uses layers of stainless steel as an absorbing medium and silicon sensors as an active medium in the regions of high radiation exposure, and scintillator tiles directly readout by silicon photomultipliers in the remaining regions. As part of the development of the detector and its readout electronic components, a section of a silicon-based HGCAL prototype detector along with a section of the CALICE AHCAL prototype was exposed to muons, electrons and charged pions in beam test experiments at the H2 beamline at the CERN SPS in October 2018. The AHCAL uses the same technology as foreseen for the HGCAL but with much finer longitudinal segmentation. The performance of the calorimeters in terms of energy response and resolution, longitudinal and transverse shower profiles is studied using negatively charged pions, and is compared to GEANT4 predictions. This is the first report summarizing results of hadronic showers measured by the HGCAL prototype using beam test data. △ Less

Submitted 27 May, 2023; v1 submitted 9 November, 2022; originally announced November 2022.

Comments: Accepted for publication by JINST

Equivalence Principle in Classical and Quantum Gravity

Authors: Nikola Paunkovic, Marko Vo**ovic

Abstract: We give a general overview of various flavors of the equivalence principle in classical and quantum physics, with special emphasis on the so-called weak equivalence principle, and contrast its validity in mechanics versus field theory. We also discuss its generalisation to a theory of quantum gravity. Our analysis suggests that only the strong equivalence principle can be considered fundamental en… ▽ More We give a general overview of various flavors of the equivalence principle in classical and quantum physics, with special emphasis on the so-called weak equivalence principle, and contrast its validity in mechanics versus field theory. We also discuss its generalisation to a theory of quantum gravity. Our analysis suggests that only the strong equivalence principle can be considered fundamental enough to be generalised to a quantum gravity context, since all other flavors of equivalence principle hold only approximately already at the classical level. △ Less

Submitted 12 November, 2022; v1 submitted 30 September, 2022; originally announced October 2022.

Comments: v2: published version, 7+6 pages

Journal ref: Universe 8, 598 (2022)

Operational verification of the existence of a spacetime manifold

Authors: Nikola Paunkovic, Marko Vo**ovic

Abstract: We argue that there exists an operational way to establish the objective reality of the notions of space and time. Specifically, we propose a theory-independent protocol for a gedanken-experiment, whose outcome is a signal establishing the observability of the spacetime manifold, without a priori assuming its existence. The experimental signal contains the information about the dimension and the t… ▽ More We argue that there exists an operational way to establish the objective reality of the notions of space and time. Specifically, we propose a theory-independent protocol for a gedanken-experiment, whose outcome is a signal establishing the observability of the spacetime manifold, without a priori assuming its existence. The experimental signal contains the information about the dimension and the topology of spacetime (with the currently achievable precision), and establishes its manifold structure, while respecting its underlying diffeomorphism symmetry. We also introduce and discuss appropriate criteria for the concept of emergence of spacetime, which any tentative theoretical model of physics must satisfy in order to claim that spacetime does emerge from some more fundamental concepts. △ Less

Submitted 9 February, 2024; v1 submitted 11 September, 2022; originally announced September 2022.

Comments: v2: various text improvements, several typos fixed. 14+4 pages, 7 figures

Rotting Infinitely Many-armed Bandits

Authors: Jung-hun Kim, Milan Vojnovic, Se-Young Yun

Abstract: We consider the infinitely many-armed bandit problem with rotting rewards, where the mean reward of an arm decreases at each pull of the arm according to an arbitrary trend with maximum rotting rate $\varrho=o(1)$. We show that this learning problem has an $Ω(\max\{\varrho^{1/3}T,\sqrt{T}\})$ worst-case regret lower bound where $T$ is the horizon time. We show that a matching upper bound… ▽ More We consider the infinitely many-armed bandit problem with rotting rewards, where the mean reward of an arm decreases at each pull of the arm according to an arbitrary trend with maximum rotting rate $\varrho=o(1)$. We show that this learning problem has an $Ω(\max\{\varrho^{1/3}T,\sqrt{T}\})$ worst-case regret lower bound where $T$ is the horizon time. We show that a matching upper bound $\tilde{O}(\max\{\varrho^{1/3}T,\sqrt{T}\})$, up to a poly-logarithmic factor, can be achieved by an algorithm that uses a UCB index for each arm and a threshold value to decide whether to continue pulling an arm or remove the arm from further consideration, when the algorithm knows the value of the maximum rotting rate $\varrho$. We also show that an $\tilde{O}(\max\{\varrho^{1/3}T,T^{3/4}\})$ regret upper bound can be achieved by an algorithm that does not know the value of $\varrho$, by using an adaptive UCB index along with an adaptive threshold value. △ Less

Submitted 17 December, 2023; v1 submitted 30 January, 2022; originally announced January 2022.

Comments: ICML2022

Topological invariant of 4-manifolds based on a 3-group

Authors: Tijana Radenkovic, Marko Vo**ovic

Abstract: We study a generalization of a 4-dimensional BF-theory in the context of higher gauge theory. We construct a triangulation independent topological state sum Z, based on the classical 3BF action for a general 3-group and a 4-dimensional spacetime manifold. This state sum coincides with Porter's TQFT for d=4 and n=3. In order to verify that the constructed state sum is a topological invariant of the… ▽ More We study a generalization of a 4-dimensional BF-theory in the context of higher gauge theory. We construct a triangulation independent topological state sum Z, based on the classical 3BF action for a general 3-group and a 4-dimensional spacetime manifold. This state sum coincides with Porter's TQFT for d=4 and n=3. In order to verify that the constructed state sum is a topological invariant of the underlying 4-dimensional manifold, its behavior under Pachner moves is analyzed, and it is obtained that the state sum Z remains the same. This paper is a generalization of the work done by Girelli, Pfeiffer, and Popescu for the case of state sum based on the classical 2BF action with the underlying 2-group structure. △ Less

Submitted 7 January, 2022; originally announced January 2022.

Comments: 21+12 pages. arXiv admin note: text overlap with arXiv:0708.3051 by other authors

Scheduling Servers with Stochastic Bilinear Rewards

Authors: Jung-hun Kim, Milan Vojnovic

Abstract: In this paper, we study scheduling in multi-class, multi-server queueing systems with stochastic rewards of job-server assignments following a bilinear model in feature vectors characterizing jobs and servers. A bilinear model allows capturing pairwise interactions of features of jobs and servers. Our goal is regret minimization for the objective of maximizing cumulative reward of job-server assig… ▽ More In this paper, we study scheduling in multi-class, multi-server queueing systems with stochastic rewards of job-server assignments following a bilinear model in feature vectors characterizing jobs and servers. A bilinear model allows capturing pairwise interactions of features of jobs and servers. Our goal is regret minimization for the objective of maximizing cumulative reward of job-server assignments over a time horizon against an oracle policy that has complete information about system parameters, while maintaining queueing system stable and allowing for different job priorities. The scheduling problem we study is motivated by various applications including matching in online platforms, such as crowdsourcing and labour platforms, and cluster computing systems. We study a scheduling algorithm based on weighted proportionally fair allocation criteria augmented with marginal costs for reward maximization, along with a linear bandit algorithm for estimating rewards of job-server assignments. For a baseline setting, in which jobs have identical mean service times, we show that our algorithm has a sub-linear regret, as well as a sub-linear bound on the mean queue length, in the time horizon. We show that similar bounds hold under more general assumptions, allowing for mean service times to be different across job classes and a time-varying set of server classes. We also show stability conditions for distributed iterative algorithms for computing allocations, which is of interest in large-scale system applications. We demonstrate the efficiency of our algorithms by numerical experiments using both synthetic randomly generated data and a real-world cluster computing data trace. △ Less

Submitted 16 June, 2022; v1 submitted 12 December, 2021; originally announced December 2021.

Response of a CMS HGCAL silicon-pad electromagnetic calorimeter prototype to 20-300 GeV positrons

Authors: B. Acar, G. Adamov, C. Adloff, S. Afanasiev, N. Akchurin, B. Akgün, F. Alam Khan, M. Alhusseini, J. Alison, A. Alpana, G. Altopp, M. Alyari, S. An, S. Anagul, I. Andreev, P. Aspell, I. O. Atakisi, O. Bach, A. Baden, G. Bakas, A. Bakshi, S. Bannerjee, P. Bargassa, D. Barney, F. Beaudette , et al. (364 additional authors not shown)

Abstract: The Compact Muon Solenoid Collaboration is designing a new high-granularity endcap calorimeter, HGCAL, to be installed later this decade. As part of this development work, a prototype system was built, with an electromagnetic section consisting of 14 double-sided structures, providing 28 sampling layers. Each sampling layer has an hexagonal module, where a multipad large-area silicon sensor is glu… ▽ More The Compact Muon Solenoid Collaboration is designing a new high-granularity endcap calorimeter, HGCAL, to be installed later this decade. As part of this development work, a prototype system was built, with an electromagnetic section consisting of 14 double-sided structures, providing 28 sampling layers. Each sampling layer has an hexagonal module, where a multipad large-area silicon sensor is glued between an electronics circuit board and a metal baseplate. The sensor pads of approximately 1 cm$^2$ are wire-bonded to the circuit board and are readout by custom integrated circuits. The prototype was extensively tested with beams at CERN's Super Proton Synchrotron in 2018. Based on the data collected with beams of positrons, with energies ranging from 20 to 300 GeV, measurements of the energy resolution and linearity, the position and angular resolutions, and the shower shapes are presented and compared to a detailed Geant4 simulation. △ Less

Submitted 31 March, 2022; v1 submitted 12 November, 2021; originally announced November 2021.

Pure Exploration and Regret Minimization in Matching Bandits

Authors: Flore Sentenac, Jialin Yi, Clément Calauzènes, Vianney Perchet, Milan Vojnovic

Abstract: Finding an optimal matching in a weighted graph is a standard combinatorial problem. We consider its semi-bandit version where either a pair or a full matching is sampled sequentially. We prove that it is possible to leverage a rank-1 assumption on the adjacency matrix to reduce the sample complexity and the regret of off-the-shelf algorithms up to reaching a linear dependency in the number of ver… ▽ More Finding an optimal matching in a weighted graph is a standard combinatorial problem. We consider its semi-bandit version where either a pair or a full matching is sampled sequentially. We prove that it is possible to leverage a rank-1 assumption on the adjacency matrix to reduce the sample complexity and the regret of off-the-shelf algorithms up to reaching a linear dependency in the number of vertices (up to poly log terms). △ Less

Submitted 31 July, 2021; originally announced August 2021.

Journal ref: Proceedings of the 38th International Conference on Machine Learning, PMLR 139, 2021

Scheduling Jobs with Stochastic Holding Costs

Authors: Dabeen Lee, Milan Vojnovic

Abstract: We study a single-server scheduling problem for the objective of minimizing the expected cumulative holding cost incurred by jobs, where parameters defining stochastic job holding costs are unknown to the scheduler. We consider a general setting allowing for different job classes, where jobs of the same class have statistically identical holding costs and service times, with an arbitrary number of… ▽ More We study a single-server scheduling problem for the objective of minimizing the expected cumulative holding cost incurred by jobs, where parameters defining stochastic job holding costs are unknown to the scheduler. We consider a general setting allowing for different job classes, where jobs of the same class have statistically identical holding costs and service times, with an arbitrary number of jobs across classes. In each time step, the server can process a job and observes random holding costs of the jobs that are yet to be completed. We consider a learning-based $cμ$ rule scheduling which starts with a preemption period of fixed duration, serving as a learning phase, and having gathered data about jobs, it switches to nonpreemptive scheduling. Our algorithms are designed to handle instances with large and small gaps in mean job holding costs and achieve near-optimal performance guarantees. The performance of algorithms is evaluated by regret, where the benchmark is the minimum possible total holding cost attained by the $cμ$ rule scheduling policy when the parameters of jobs are known. We show regret lower bounds and algorithms that achieve nearly matching regret upper bounds. Our numerical results demonstrate the efficacy of our algorithms and show that our regret analysis is nearly tight. △ Less

Submitted 21 September, 2022; v1 submitted 28 May, 2021; originally announced May 2021.

Comments: Extended abstract appeared in NeurIPS 2021

Quantum gravity and elementary particles from higher gauge theory

Authors: Tijana Radenkovic, Marko Vo**ovic

Abstract: We give a brief overview how to couple general relativity to the Standard Model of elementary particles, within the higher gauge theory framework, suitable for the spinfoam quantization procedure. We begin by providing a short review of all relevant mathematical concepts, most notably the idea of a categorical ladder, 3-groups and generalized parallel transport. Then, we give an explicit construct… ▽ More We give a brief overview how to couple general relativity to the Standard Model of elementary particles, within the higher gauge theory framework, suitable for the spinfoam quantization procedure. We begin by providing a short review of all relevant mathematical concepts, most notably the idea of a categorical ladder, 3-groups and generalized parallel transport. Then, we give an explicit construction of the algebraic structure which describes the full Standard Model coupled to Einstein-Cartan gravity, along with the classical action, written in the form suitable for the spinfoam quantization procedure. We emphasize the usefulness of the 3-group concept as a superior tool to describe gauge symmetry, compared to an ordinary Lie group, as well as the possibility to employ this new structure to classify matter fields and study their spectrum, including the origin of fermion families. △ Less

Submitted 14 March, 2021; originally announced March 2021.

Comments: Proceedings of the Workshop on Quantum Fields and Nonlinear Phenomena, Craiova, Romania, 2020. 14 pages

Journal ref: Ann. Univ. Craiova Phys. 30, 74 (2020)

Gauge symmetry of the 3BF theory for a generic Lie 3-group

Authors: Tijana Radenkovic, Marko Vo**ovic

Abstract: The higher category theory can be employed to generalize the BF action to the so-called 3BF action, by passing from the notion of a gauge group to the notion of a gauge 3-group. In this work we determine the full gauge symmetry of the 3BF action. To that end, the complete Hamiltonian analysis of the 3BF action for an arbitrary semistrict Lie 3-group is performed, by using the Dirac procedure. This… ▽ More The higher category theory can be employed to generalize the BF action to the so-called 3BF action, by passing from the notion of a gauge group to the notion of a gauge 3-group. In this work we determine the full gauge symmetry of the 3BF action. To that end, the complete Hamiltonian analysis of the 3BF action for an arbitrary semistrict Lie 3-group is performed, by using the Dirac procedure. This analysis is the first step towards a canonical quantization of a 3BF theory. This is an important step**-stone for the quantization of the complete Standard Model of elementary particles coupled to Einstein-Cartan gravity, formulated as a 3BF action with suitable simplicity constraints. We show that the resulting gauge symmetry group consists of the already familiar G-, H-, and L-gauge transformations, as well as additional M- and N-gauge transformations, which have not been discussed in the existing literature. △ Less

Submitted 12 April, 2022; v1 submitted 11 January, 2021; originally announced January 2021.

Comments: v4: new references added, minor corrections made; 22+14 pages, 2 figures

Test Score Algorithms for Budgeted Stochastic Utility Maximization

Authors: Dabeen Lee, Milan Vojnovic, Se-Young Yun

Abstract: Motivated by recent developments in designing algorithms based on individual item scores for solving utility maximization problems, we study the framework of using test scores, defined as a statistic of observed individual item performance data, for solving the budgeted stochastic utility maximization problem. We extend an existing scoring mechanism, namely the replication test scores, to incorpor… ▽ More Motivated by recent developments in designing algorithms based on individual item scores for solving utility maximization problems, we study the framework of using test scores, defined as a statistic of observed individual item performance data, for solving the budgeted stochastic utility maximization problem. We extend an existing scoring mechanism, namely the replication test scores, to incorporate heterogeneous item costs as well as item values. We show that a natural greedy algorithm that selects items solely based on their replication test scores outputs solutions within a constant factor of the optimum for a broad class of utility functions. Our algorithms and approximation guarantees assume that test scores are noisy estimates of certain expected values with respect to marginal distributions of individual item values, thus making our algorithms practical and extending previous work that assumes noiseless estimates. Moreover, we show how our algorithm can be adapted to the setting where items arrive in a streaming fashion while maintaining the same approximation guarantee. We present numerical results, using synthetic data and data sets from the Academia.StackExchange Q&A forum, which show that our test score algorithm can achieve competitiveness, and in some cases better performance than a benchmark algorithm that requires access to a value oracle to evaluate function values. △ Less

Submitted 24 February, 2022; v1 submitted 30 December, 2020; originally announced December 2020.

Operational interpretation of the vacuum and process matrices for identical particles

Authors: Ricardo Faleiro, Nikola Paunkovic, Marko Vo**ovic

Abstract: This work overviews the single-particle two-way communication protocol recently introduced by del Santo and Dakić (dSD), and analyses it using the process matrix formalism. We give a detailed account of the importance and the operational meaning of the interaction of an agent with the vacuum -- in particular its role in the process matrix description. Our analysis shows that the interaction with t… ▽ More This work overviews the single-particle two-way communication protocol recently introduced by del Santo and Dakić (dSD), and analyses it using the process matrix formalism. We give a detailed account of the importance and the operational meaning of the interaction of an agent with the vacuum -- in particular its role in the process matrix description. Our analysis shows that the interaction with the vacuum should be treated as an operation, on equal footing with all other interactions. This raises the issue of counting such operations in an operational manner. Motivated by this analysis, we apply the process matrix formalism to capped Fock spaces using the framework of second quantisation, in order to characterise protocols with an indefinite number of identical particles. △ Less

Submitted 14 April, 2023; v1 submitted 29 October, 2020; originally announced October 2020.

Comments: final version, 12+3 pages (double column), 3 figures

Journal ref: Quantum 7, 986 (2023)

Popularity Prediction for Social Media over Arbitrary Time Horizons

Authors: Daniel Haimovich, Dima Karamshuk, Thomas J. Leeper, Evgeniy Riabenko, Milan Vojnovic

Abstract: Predicting the popularity of social media content in real time requires approaches that efficiently operate at global scale. Popularity prediction is important for many applications, including detection of harmful viral content to enable timely content moderation. The prediction task is difficult because views result from interactions between user interests, content features, resharing, feed ranki… ▽ More Predicting the popularity of social media content in real time requires approaches that efficiently operate at global scale. Popularity prediction is important for many applications, including detection of harmful viral content to enable timely content moderation. The prediction task is difficult because views result from interactions between user interests, content features, resharing, feed ranking, and network structure. We consider the problem of accurately predicting popularity both at any given prediction time since a content item's creation and for arbitrary time horizons into the future. In order to achieve high accuracy for different prediction time horizons, it is essential for models to use static features (of content and user) as well as observed popularity growth up to prediction time. We propose a feature-based approach based on a self-excited Hawkes point process model, which involves prediction of the content's popularity at one or more reference horizons in tandem with a point predictor of an effective growth parameter that reflects the timescale of popularity growth. This results in a highly scalable method for popularity prediction over arbitrary prediction time horizons that also achieves a high degree of accuracy, compared to several leading baselines, on a dataset of public page content on Facebook over a two-month period, covering billions of content views and hundreds of thousands of distinct content items. The model has shown competitive prediction accuracy against a strong baseline that consists of separately trained models for specific prediction time horizons. △ Less

Submitted 22 December, 2021; v1 submitted 4 September, 2020; originally announced September 2020.

Comments: International Conference on Very Large Data Bases (VLDB'2022)

Standard Model and 4-groups

Authors: Aleksandar Mikovic, Marko Vo**ovic

Abstract: We show that a categorical generalization of the the Poincaré symmetry which is based on the n-crossed modules becomes natural and simple when n=3 and that the corresponding 3-form and 4-form gauge fields have to be a Dirac spinor and a Lorentz scalar, respectively. Hence by using a Poincaré 4-group we naturally incorporate fermionic and scalar matter into the corresponding 4-connection. The inter… ▽ More We show that a categorical generalization of the the Poincaré symmetry which is based on the n-crossed modules becomes natural and simple when n=3 and that the corresponding 3-form and 4-form gauge fields have to be a Dirac spinor and a Lorentz scalar, respectively. Hence by using a Poincaré 4-group we naturally incorporate fermionic and scalar matter into the corresponding 4-connection. The internal symmetries can be included into the 4-group structure by using a 3-crossed module based on the $SL(2,\mathbb{C}) \times K$ group, so that for $K=U(1)\times SU(2) \times SU(3)$ we can include the Standard Model into this categorification scheme. △ Less

Submitted 13 May, 2021; v1 submitted 9 August, 2020; originally announced August 2020.

Comments: v2: published version, 4 pages

Journal ref: Europhys. Lett. 133, 61001 (2021)

Construction and examples of higher gauge theories

Authors: Tijana Radenkovic, Marko Vo**ovic

Abstract: We provide several examples of higher gauge theories, constructed as generalizations of a BF model to 2BF and 3BF models with constraints. Using the framework of higher category theory, we introduce appropriate 2-groups and 3-groups, and construct the actions for the corresponding constrained 2BF and 3BF theories. In this way, we can construct actions which describe the correct dynamics of Yang-Mi… ▽ More We provide several examples of higher gauge theories, constructed as generalizations of a BF model to 2BF and 3BF models with constraints. Using the framework of higher category theory, we introduce appropriate 2-groups and 3-groups, and construct the actions for the corresponding constrained 2BF and 3BF theories. In this way, we can construct actions which describe the correct dynamics of Yang-Mills, Klein-Gordon, Dirac, Weyl, and Majorana fields coupled to Einstein-Cartan gravity. Each action is naturally split into a topological sector and a sector with simplicity constraints. The properties of the higher gauge group structure opens up a possibility of a nontrivial unification of all fields. △ Less

Submitted 19 May, 2020; originally announced May 2020.

Comments: 26 pages, contribution to the MPHYS10 2019 conference proceedings

Journal ref: SFIN XXXIII, 251 (2020)

Hamiltonian Analysis for the Scalar Electrodynamics as 3BF Theory

Authors: Tijana Radenkovic, Marko Vo**ovic

Abstract: The higher category theory can be employed to generalize the BF action to the so-called 3BF action, by passing from the notion of a gauge group to the notion of a gauge 3-group. The theory of scalar electrodynamics coupled to Einstein-Cartan gravity can be formulated as a constrained 3BF theory for a specific choice of the gauge 3-group. The complete Hamiltonian analysis of the 3BF action for the… ▽ More The higher category theory can be employed to generalize the BF action to the so-called 3BF action, by passing from the notion of a gauge group to the notion of a gauge 3-group. The theory of scalar electrodynamics coupled to Einstein-Cartan gravity can be formulated as a constrained 3BF theory for a specific choice of the gauge 3-group. The complete Hamiltonian analysis of the 3BF action for the choice of a Lie 3-group corresponding to scalar electrodynamics is performed. This analysis is the first step towards a canonical quantization of a 3BF theory, an important step** stone for the quantization of the complete scalar electrodynamics coupled to Einstein-Cartan gravity formulated as a 3BF action with suitable simplicity constraints. It is shown that the resulting dynamic constraints eliminate all propagating degrees of freedom, i.e., the 3BF theory for this choice of a 3-group is a topological field theory, as expected. △ Less

Submitted 15 April, 2020; originally announced April 2020.

Comments: published version, 21 pages

Journal ref: Symmetry 12, 620 (2020)

Causal orders, quantum circuits and spacetime: distinguishing between definite and superposed causal orders

Authors: Nikola Paunkovic, Marko Vo**ovic

Abstract: We study the notion of causal orders for the cases of (classical and quantum) circuits and spacetime events. We show that every circuit can be immersed into a classical spacetime, preserving the compatibility between the two causal structures. Using the process matrix formalism, we analyse the realisations of the quantum switch using 4 and 3 spacetime events in classical spacetimes with fixed caus… ▽ More We study the notion of causal orders for the cases of (classical and quantum) circuits and spacetime events. We show that every circuit can be immersed into a classical spacetime, preserving the compatibility between the two causal structures. Using the process matrix formalism, we analyse the realisations of the quantum switch using 4 and 3 spacetime events in classical spacetimes with fixed causal orders, and the realisation of a gravitational switch with only 2 spacetime events that features superpositions of different gravitational field configurations and their respective causal orders. We show that the current quantum switch experimental implementations do not feature superpositions of causal orders between spacetime events, and that these superpositions can only occur in the context of superposed gravitational fields. We also discuss a recently introduced operational notion of an event, which does allow for superpositions of respective causal orders in flat spacetime quantum switch implementations. We construct two observables that can distinguish between the quantum switch realisations in classical spacetimes, and gravitational switch implementations in superposed spacetimes. Finally, we discuss our results in the light of the modern relational approach to physics. △ Less

Submitted 23 May, 2020; v1 submitted 23 May, 2019; originally announced May 2019.

Comments: 18 pages + Appendices

Journal ref: Quantum 4, 275 (2020)

Higher Gauge Theories Based on 3-groups

Authors: Tijana Radenkovic, Marko Vo**ovic

Abstract: We study the categorical generalizations of a $BF$ theory to $2BF$ and $3BF$ theories, corresponding to $2$-groups and $3$-groups, in the framework of higher gauge theory. In particular, we construct the constrained $3BF$ actions describing the correct dynamics of Yang-Mills, Klein-Gordon, Dirac, Weyl and Majorana fields coupled to Einstein-Cartan gravity. The action is naturally split into a topo… ▽ More We study the categorical generalizations of a $BF$ theory to $2BF$ and $3BF$ theories, corresponding to $2$-groups and $3$-groups, in the framework of higher gauge theory. In particular, we construct the constrained $3BF$ actions describing the correct dynamics of Yang-Mills, Klein-Gordon, Dirac, Weyl and Majorana fields coupled to Einstein-Cartan gravity. The action is naturally split into a topological sector and a sector with simplicity constraints, adapted to the spinfoam quantization programme. In addition, the structure of the $3$-group gives rise to a novel gauge group which specifies the spectrum of matter fields present in the theory, just like the ordinary gauge group specifies the spectrum of gauge bosons in the Yang-Mills theory. This allows us to rewrite the whole Standard Model coupled to gravity as a constrained $3BF$ action, facilitating the nonperturbative quantization of both gravity and matter fields. Moreover, the presence and the properties of this new gauge group open up a possibility of a nontrivial unification of all fields and a possible explanation of fermion families and all other structure in the matter spectrum of the theory. △ Less

Submitted 4 November, 2019; v1 submitted 16 April, 2019; originally announced April 2019.

Comments: v4: published version

Journal ref: JHEP10(2019)222

Accelerated MM Algorithms for Ranking Scores Inference from Comparison Data

Authors: Milan Vojnovic, Seyoung Yun, Kaifang Zhou

Abstract: In this paper, we study a popular method for inference of the Bradley-Terry model parameters, namely the MM algorithm, for maximum likelihood estimation and maximum a posteriori probability estimation. This class of models includes the Bradley-Terry model of paired comparisons, the Rao-Kupper model of paired comparisons allowing for tie outcomes, the Luce choice model, and the Plackett-Luce rankin… ▽ More In this paper, we study a popular method for inference of the Bradley-Terry model parameters, namely the MM algorithm, for maximum likelihood estimation and maximum a posteriori probability estimation. This class of models includes the Bradley-Terry model of paired comparisons, the Rao-Kupper model of paired comparisons allowing for tie outcomes, the Luce choice model, and the Plackett-Luce ranking model. We establish tight characterizations of the convergence rate for the MM algorithm, and show that it is essentially equivalent to that of a gradient descent algorithm. For the maximum likelihood estimation, the convergence is shown to be linear with the rate crucially determined by the algebraic connectivity of the matrix of item pair co-occurrences in observed comparison data. For the Bayesian inference, the convergence rate is also shown to be linear, with the rate determined by a parameter of the prior distribution in a way that can make the convergence arbitrarily slow for small values of this parameter. We propose a simple modification of the classical MM algorithm that avoids the observed slow convergence issue and accelerates the convergence. The key component of the accelerated MM algorithm is a parameter rescaling performed at each iteration step that is carefully chosen based on theoretical analysis and characterisation of the convergence rate. Our experimental results, performed on both synthetic and real-world data, demonstrate the identified slow convergence issue of the classic MM algorithm, and show that significant efficiency gains can be obtained by our new proposed method. △ Less

Submitted 26 December, 2020; v1 submitted 1 January, 2019; originally announced January 2019.

Hamiltonian analysis of the BFCG formulation of General Relativity

Authors: Aleksandar Mikovic, Miguel A. Oliveira, Marko Vo**ovic

Abstract: We perform the complete Hamiltonian analysis of the BFCG action for General Relativity. We determine all the constraints of the theory and classify them into the first-class and the second-class constraints. We also show how the canonical formulation of BFCG General Relativity reduces to the Einstein-Cartan and triad canonical formulations. The reduced phase space analysis also gives a 2-connectio… ▽ More We perform the complete Hamiltonian analysis of the BFCG action for General Relativity. We determine all the constraints of the theory and classify them into the first-class and the second-class constraints. We also show how the canonical formulation of BFCG General Relativity reduces to the Einstein-Cartan and triad canonical formulations. The reduced phase space analysis also gives a 2-connection which is suitable for the construction of a spin-foam basis which will be a categorical generalization of the spin-network basis from Loop Quantum Gravity. △ Less

Submitted 3 December, 2018; v1 submitted 17 July, 2018; originally announced July 2018.

Comments: v2: published version, 21 pages

Journal ref: Class. Quant. Grav. 36, 015005 (2019)

KONG: Kernels for ordered-neighborhood graphs

Authors: Moez Draief, Konstantin Kutzkov, Kevin Scaman, Milan Vojnovic

Abstract: We present novel graph kernels for graphs with node and edge labels that have ordered neighborhoods, i.e. when neighbor nodes follow an order. Graphs with ordered neighborhoods are a natural data representation for evolving graphs where edges are created over time, which induces an order. Combining convolutional subgraph kernels and string kernels, we design new scalable algorithms for generation… ▽ More We present novel graph kernels for graphs with node and edge labels that have ordered neighborhoods, i.e. when neighbor nodes follow an order. Graphs with ordered neighborhoods are a natural data representation for evolving graphs where edges are created over time, which induces an order. Combining convolutional subgraph kernels and string kernels, we design new scalable algorithms for generation of explicit graph feature maps using sketching techniques. We obtain precise bounds for the approximation accuracy and computational complexity of the proposed approaches and demonstrate their applicability on real datasets. In particular, our experiments demonstrate that neighborhood ordering results in more informative features. For the special case of general graphs, i.e. graphs without ordered neighborhoods, the new graph kernels yield efficient and simple algorithms for the comparison of label distributions between graphs. △ Less

Submitted 29 May, 2018; v1 submitted 25 May, 2018; originally announced May 2018.

Quantum gravity for piecewise flat spacetimes

Authors: Aleksandar Mikovic, Marko Vo**ovic

Abstract: We describe a theory of quantum gravity which is based on the assumption that the spacetime structure at small distances is given by a piecewise linear (PL) 4-manifold corresponding to a triangulation of a smooth 4-manifold. The fundamental degrees of freedom are the edge lengths of the triangulation. One can work with finitely many edge lengths, so that the corresponding Regge path integral can b… ▽ More We describe a theory of quantum gravity which is based on the assumption that the spacetime structure at small distances is given by a piecewise linear (PL) 4-manifold corresponding to a triangulation of a smooth 4-manifold. The fundamental degrees of freedom are the edge lengths of the triangulation. One can work with finitely many edge lengths, so that the corresponding Regge path integral can be made finite by using an appropriate path-integral measure. The semi-classical limit is computed by using the effective action formalism, and the existence of a semi-classical effective action restricts the choice of the path-integral measure. The classical limit is given by the Regge action, so that one has a quantum gravity theory for a piecewise-flat general relativity. By using the effective action formalism we show that the observed value of the cosmological constant can be recovered from the effective cosmological constant. When the number of 4-simplices in the spacetime triangulation is large, then the PL effective action is well approximated by a quantum field theory effective action with a physical cutoff determined by the smallest edge length. △ Less

Submitted 7 April, 2018; originally announced April 2018.

Comments: 14 pages. To appear as a contribution to the MPHYS9 2017 conference proceedings

Journal ref: SFIN XXXI, 267 (2018)

Entanglement-induced deviation from the geodesic motion in quantum gravity

Authors: Francisco Pipa, Nikola Paunkovic, Marko Vo**ovic

Abstract: We study the derivation of the effective equation of motion for a pointlike particle in the framework of quantum gravity. Just like the geodesic motion of a classical particle is a consequence of classical field theory coupled to general relativity, we introduce the similar notion of an effective equation of motion, but starting from an abstract quantum gravity description. In the presence of enta… ▽ More We study the derivation of the effective equation of motion for a pointlike particle in the framework of quantum gravity. Just like the geodesic motion of a classical particle is a consequence of classical field theory coupled to general relativity, we introduce the similar notion of an effective equation of motion, but starting from an abstract quantum gravity description. In the presence of entanglement between gravity and matter, quantum effects give rise to modifications of the geodesic trajectory, primarily as a consequence of the nonzero overlap between various coherent states of the gravity-matter system. Finally, we discuss the status of the weak equivalence principle in quantum gravity and its possible violation due to the nongeodesic motion. △ Less

Submitted 27 September, 2019; v1 submitted 9 January, 2018; originally announced January 2018.

Comments: v4: published version, 37 pages

Journal ref: Jour. Cosmol. Astropart. Phys. 09, 057 (2019)

Parameter Estimation for Thurstone Choice Models

Authors: Milan Vojnovic, Se-Young Yun

Abstract: We consider the estimation accuracy of individual strength parameters of a Thurstone choice model when each input observation consists of a choice of one item from a set of two or more items (so called top-1 lists). This model accommodates the well-known choice models such as the Luce choice model for comparison sets of two or more items and the Bradley-Terry model for pair comparisons. We provi… ▽ More We consider the estimation accuracy of individual strength parameters of a Thurstone choice model when each input observation consists of a choice of one item from a set of two or more items (so called top-1 lists). This model accommodates the well-known choice models such as the Luce choice model for comparison sets of two or more items and the Bradley-Terry model for pair comparisons. We provide a tight characterization of the mean squared error of the maximum likelihood parameter estimator. We also provide similar characterizations for parameter estimators defined by a rank-breaking method, which amounts to deducing one or more pair comparisons from a comparison of two or more items, assuming independence of these pair comparisons, and maximizing a likelihood function derived under these assumptions. We also consider a related binary classification problem where each individual parameter takes value from a set of two possible values and the goal is to correctly classify all items within a prescribed classification error. △ Less

Submitted 29 April, 2017; originally announced May 2017.

Comments: 55 pages

Communication complexity of approximate maximum matching in the message-passing model

Authors: Zengfeng Huang, Bozidar Radunovic, Milan Vojnovic, Qin Zhang

Abstract: We consider the communication complexity of finding an approximate maximum matching in a graph in a multi-party message-passing communication model. The maximum matching problem is one of the most fundamental graph combinatorial problems, with a variety of applications. The input to the problem is a graph $G$ that has $n$ vertices and the set of edges partitioned over $k$ sites, and an approxima… ▽ More We consider the communication complexity of finding an approximate maximum matching in a graph in a multi-party message-passing communication model. The maximum matching problem is one of the most fundamental graph combinatorial problems, with a variety of applications. The input to the problem is a graph $G$ that has $n$ vertices and the set of edges partitioned over $k$ sites, and an approximation ratio parameter $α$. The output is required to be a matching in $G$ that has to be reported by one of the sites, whose size is at least factor $α$ of the size of a maximum matching in $G$. We show that the communication complexity of this problem is $Ω(α^2 k n)$ information bits. This bound is shown to be tight up to a $\log n$ factor, by constructing an algorithm, establishing its correctness, and an upper bound on the communication cost. The lower bound also applies to other graph combinatorial problems in the message-passing communication model, including max-flow and graph sparsification. △ Less

Submitted 27 April, 2017; originally announced April 2017.

Adaptive Matching for Expert Systems with Uncertain Task Types

Authors: Virag Shah, Lennart Gulikers, Laurent Massoulie, Milan Vojnovic

Abstract: A matching in a two-sided market often incurs an externality: a matched resource may become unavailable to the other side of the market, at least for a while. This is especially an issue in online platforms involving human experts as the expert resources are often scarce. The efficient utilization of experts in these platforms is made challenging by the fact that the information available about th… ▽ More A matching in a two-sided market often incurs an externality: a matched resource may become unavailable to the other side of the market, at least for a while. This is especially an issue in online platforms involving human experts as the expert resources are often scarce. The efficient utilization of experts in these platforms is made challenging by the fact that the information available about the parties involved is usually limited. To address this challenge, we develop a model of a task-expert matching system where a task is matched to an expert using not only the prior information about the task but also the feedback obtained from the past matches. In our model the tasks arrive online while the experts are fixed and constrained by a finite service capacity. For this model, we characterize the maximum task resolution throughput a platform can achieve. We show that the natural greedy approaches where each expert is assigned a task most suitable to her skill is suboptimal, as it does not internalize the above externality. We develop a throughput optimal backpressure algorithm which does so by accounting for the `congestion' among different task types. Finally, we validate our model and confirm our theoretical findings with data-driven simulations via logs of Math.StackExchange, a StackOverflow forum dedicated to mathematics. △ Less

Submitted 26 October, 2018; v1 submitted 2 March, 2017; originally announced March 2017.

Comments: A part of it presented at Allerton Conference 2017, 18 pages

Gauge protected entanglement between gravity and matter

Authors: Nikola Paunkovic, Marko Vo**ovic

Abstract: We show that gravity and matter fields are generically entangled, as a consequence of the local Poincaré symmetry. First, we present a general argument, applicable to any particular theory of quantum gravity with matter, by performing the analysis in the abstract nonperturbative canonical framework, demonstrating the nonseparability of the scalar constraint, thus promoting the entangled states as… ▽ More We show that gravity and matter fields are generically entangled, as a consequence of the local Poincaré symmetry. First, we present a general argument, applicable to any particular theory of quantum gravity with matter, by performing the analysis in the abstract nonperturbative canonical framework, demonstrating the nonseparability of the scalar constraint, thus promoting the entangled states as the physical ones. Also, within the covariant framework, we show explicitly that the Hartle-Hawking state in the Regge model of quantum gravity is necessarily entangled. Our result is potentially relevant for the quantum-to-classical transition, taken within the framework of the decoherence programme: due to the gauge symmetry requirements, the matter does not decohere, it is by default decohered by gravity. Generically, entanglement is a consequence of interaction. This new entanglement could potentially, in form of an "effective interaction", bring about corrections to the weak equivalence principle, further confirming that spacetime as a smooth four-dimensional manifold is an emergent phenomenon. Finally, the existence of the gauge-protected entanglement between gravity and matter could be seen as a criterion for a plausible theory of quantum gravity, and in the case of perturbative quantisation approaches, a confirmation of the persistence of the manifestly broken gauge symmetry. △ Less

Submitted 29 August, 2018; v1 submitted 24 February, 2017; originally announced February 2017.

Comments: v4: published version, 30 pages

Journal ref: Class. Quant. Grav. 35, 185015 (2018)

Hamiltonian analysis of the BFCG theory for a strict Lie 2-group

Authors: Aleksandar Mikovic, Miguel Angelo Oliveira, Marko Vo**ovic

Abstract: We perform a complete Hamiltonian analysis of the BFCG action for a general Lie 2-group by using the Dirac procedure. We show that the resulting dynamical constraints eliminate all local degrees of freedom which implies that the BFCG theory is a topological field theory. We perform a complete Hamiltonian analysis of the BFCG action for a general Lie 2-group by using the Dirac procedure. We show that the resulting dynamical constraints eliminate all local degrees of freedom which implies that the BFCG theory is a topological field theory. △ Less

Submitted 25 March, 2024; v1 submitted 30 October, 2016; originally announced October 2016.

Comments: v3: 18 pages, published version

Journal ref: Adv. Theor. Math. Phys. 26, 3783 (2022)

QSGD: Communication-Efficient SGD via Gradient Quantization and Encoding

Authors: Dan Alistarh, Demjan Grubic, Jerry Li, Ryota Tomioka, Milan Vojnovic

Abstract: Parallel implementations of stochastic gradient descent (SGD) have received significant research attention, thanks to excellent scalability properties of this algorithm, and to its efficiency in the context of training deep neural networks. A fundamental barrier for parallelizing large-scale SGD is the fact that the cost of communicating the gradient updates between nodes can be very large. Conseq… ▽ More Parallel implementations of stochastic gradient descent (SGD) have received significant research attention, thanks to excellent scalability properties of this algorithm, and to its efficiency in the context of training deep neural networks. A fundamental barrier for parallelizing large-scale SGD is the fact that the cost of communicating the gradient updates between nodes can be very large. Consequently, lossy compression heuristics have been proposed, by which nodes only communicate quantized gradients. Although effective in practice, these heuristics do not always provably converge, and it is not clear whether they are optimal. In this paper, we propose Quantized SGD (QSGD), a family of compression schemes which allow the compression of gradient updates at each node, while guaranteeing convergence under standard assumptions. QSGD allows the user to trade off compression and convergence time: it can communicate a sublinear number of bits per iteration in the model dimension, and can achieve asymptotically optimal communication cost. We complement our theoretical results with empirical data, showing that QSGD can significantly reduce communication cost, while being competitive with standard uncompressed techniques on a variety of real tasks. In particular, experiments show that gradient quantization applied to training of deep neural networks for image classification and automated speech recognition can lead to significant reductions in communication cost, and end-to-end training time. For instance, on 16 GPUs, we are able to train a ResNet-152 network on ImageNet 1.8x faster to full accuracy. Of note, we show that there exist generic parameter settings under which all known network architectures preserve or slightly improve their full accuracy when using quantization. △ Less

Submitted 6 December, 2017; v1 submitted 6 October, 2016; originally announced October 2016.

Submodular Maximization using Test Scores

Authors: Shreyas Sekar, Milan Vojnovic, Se-Young Yun

Abstract: We study the canonical problem of maximizing a stochastic submodular function subject to a cardinality constraint, where the goal is to select a subset from a ground set of items with uncertain individual performances to maximize their expected group value. Although near-optimal algorithms have been proposed for this problem, practical concerns regarding scalability, compatibility with distributed… ▽ More We study the canonical problem of maximizing a stochastic submodular function subject to a cardinality constraint, where the goal is to select a subset from a ground set of items with uncertain individual performances to maximize their expected group value. Although near-optimal algorithms have been proposed for this problem, practical concerns regarding scalability, compatibility with distributed implementation, and expensive oracle queries persist in large-scale applications. Motivated by online platforms that rely on individual item scores for content recommendation and team selection, we propose a special class of algorithms that select items based solely on individual performance measures known as test scores. The central contribution of this work is a novel and systematic framework for designing test score based algorithms for a broad class of naturally occurring utility functions. We introduce a new scoring mechanism that we refer to as replication test scores and prove that as long as the objective function satisfies a diminishing returns property, one can leverage these scores to compute solutions that are within a constant factor of the optimum. We then extend our results to the more general stochastic submodular welfare maximization problem, where the goal is to select items and assign them to multiple groups to maximize the sum of the expected group values. For this more difficult problem, we show that replication test scores can be used to develop an algorithm that approximates the optimum solution up to a logarithmic factor. The techniques presented in this work bridge the gap between the rigorous theoretical work on submodular optimization and simple, scalable heuristics that are useful in certain domains. △ Less

Submitted 9 May, 2019; v1 submitted 23 May, 2016; originally announced May 2016.

Comments: Under review

Gravity-Matter Entanglement in Regge Quantum Gravity

Authors: Nikola Paunković, Marko Vo**ović

Abstract: We argue that Hartle-Hawking states in the Regge quantum gravity model generically contain non-trivial entanglement between gravity and matter fields. Generic impossibility to talk about "matter in a point of space" is in line with the idea of an emergent spacetime, and as such could be taken as a possible candidate for a criterion for a plausible theory of quantum gravity. Finally, this new entan… ▽ More We argue that Hartle-Hawking states in the Regge quantum gravity model generically contain non-trivial entanglement between gravity and matter fields. Generic impossibility to talk about "matter in a point of space" is in line with the idea of an emergent spacetime, and as such could be taken as a possible candidate for a criterion for a plausible theory of quantum gravity. Finally, this new entanglement could be seen as an additional "effective interaction", which could possibly bring corrections to the weak equivalence principle. △ Less

Submitted 22 January, 2016; originally announced January 2016.

Comments: Proceedings of the EmQM15 conference, to appear in J. Phys. Conf. Ser. 2 pages

Journal ref: J. Phys.: Conf. Ser. 701, 012035 (2016)

Categorical generalization of spinfoam models

Authors: Aleksandar Mikovic, Marko Vo**ovic

Abstract: We give a brief review of the problem of quantum gravity. After the discussion of the nonrenormalizability of general relativity, we briefly mention the main research directions which aim to resolve this problem. Our attention then focuses on the approach of Loop Quantum Gravity, specifically spinfoam models. These models have some issues concerning the semiclassical limit and coupling of matter f… ▽ More We give a brief review of the problem of quantum gravity. After the discussion of the nonrenormalizability of general relativity, we briefly mention the main research directions which aim to resolve this problem. Our attention then focuses on the approach of Loop Quantum Gravity, specifically spinfoam models. These models have some issues concerning the semiclassical limit and coupling of matter fields. The recent developments in category theory provide us with the necessary formalism to introduce a new action for general relativity and perform covariant quantization so that the issues of spinfoam models are successfully resolved. △ Less

Submitted 19 December, 2015; originally announced December 2015.

Comments: Proceedings of the QQQ conference, Tallinn, Estonia, 2012. 10 pages

Journal ref: J. Phys.: Conf. Ser. 532, 012020 (2014)

Hamiltonian analysis of the BFCG theory for the Poincare 2-group

Authors: Aleksandar Mikovic, Miguel A. Oliveira, Marko Vo**ovic

Abstract: We perform the full Hamiltonian analysis of the topological BFCG action based on the Poincare 2-group. The Hamiltonian of the theory is constructed, and the algebra of constraints is computed. The Dirac brackets are evaluated, and the second class constraints are then eliminated from the theory. The results are contrasted to those of the topological Poincare gauge theory, which is equivalent to th… ▽ More We perform the full Hamiltonian analysis of the topological BFCG action based on the Poincare 2-group. The Hamiltonian of the theory is constructed, and the algebra of constraints is computed. The Dirac brackets are evaluated, and the second class constraints are then eliminated from the theory. The results are contrasted to those of the topological Poincare gauge theory, which is equivalent to the BFCG model at the level of the classical action, but has a very different Hamiltonian structure. △ Less

Submitted 25 February, 2016; v1 submitted 23 August, 2015; originally announced August 2015.

Comments: v3: Published version. 21 pages

Journal ref: Class. Quant. Grav. 33, 065007 (2016)

Causal Dynamical Triangulations in the Spincube Model of Quantum Gravity

Authors: Marko Vo**ovic

Abstract: We study the implications of the simplicity constraint in the spincube model of quantum gravity. By relating the edge-lengths to the integer areas of triangles, the simplicity constraint imposes very strong restrictions between them, ultimately leading to a requirement that all 4-simplices in the triangulation must be almost mutually identical. As a surprising and unexpected consequence of this pr… ▽ More We study the implications of the simplicity constraint in the spincube model of quantum gravity. By relating the edge-lengths to the integer areas of triangles, the simplicity constraint imposes very strong restrictions between them, ultimately leading to a requirement that all 4-simplices in the triangulation must be almost mutually identical. As a surprising and unexpected consequence of this property, one can obtain the CDT state sum as a special case of the spincube state sum. This relationship brings new insight into the long-standing problem of the relationship between the spinfoam approach and the CDT approach to quantum gravity. In particular, it turns out that the spincube model contains properties of both approaches, providing a single unifying framework for their analysis and comparison. In addition, the spincube state sum also contains some other special cases, very similar but not equivalent to the CDT state sum. △ Less

Submitted 31 July, 2016; v1 submitted 22 June, 2015; originally announced June 2015.

Comments: v4: published version; 18 pages

Journal ref: Phys. Rev. D 94, 024058 (2016)

Solution to the Cosmological Constant Problem in a Regge Quantum Gravity Model

Authors: Aleksandar Mikovic, Marko Vo**ovic

Abstract: We show that it is possible to solve the cosmological constant (CC) problem in a discrete quantum gravity theory based on Regge calculus by using the effective action approach and a special path-integral measure. The effective cosmological constant is given as a sum of 3 terms: the classical CC, the quantum gravity CC and the matter CC. Since the observations can only measure the sum of these 3 te… ▽ More We show that it is possible to solve the cosmological constant (CC) problem in a discrete quantum gravity theory based on Regge calculus by using the effective action approach and a special path-integral measure. The effective cosmological constant is given as a sum of 3 terms: the classical CC, the quantum gravity CC and the matter CC. Since the observations can only measure the sum of these 3 terms, we can choose the classical CC to be equal to the negative value of the matter CC. Hence the effective CC is given only by the quantum gravity CC, which is determined by the path-integral measure. Since the path-integral measure depends on a free parameter, this parameter can be chosen such that the effective CC gives the observed value. △ Less

Submitted 12 June, 2015; v1 submitted 5 July, 2014; originally announced July 2014.

Comments: v4: published version. arXiv admin note: substantial text overlap with arXiv:1407.1124

Journal ref: Europhys. Lett. 110, 40008 (2015)

Cosmological Constant in a Quantum Gravity Theory for a Piecewise-linear Spacetime

Authors: Aleksandar Mikovic, Marko Vo**ovic

Abstract: We study the quantum contributions to the classical cosmological constant in a quantum gravity theory for GR with matter on a piecewise linear spacetime corresponding to a triangulation of a smooth manifold. We use the effective action approach and a special path-integral (PI) measure which depends on a free parameter, while matter is modeled by a massive self-interacting scalar field. The effecti… ▽ More We study the quantum contributions to the classical cosmological constant in a quantum gravity theory for GR with matter on a piecewise linear spacetime corresponding to a triangulation of a smooth manifold. We use the effective action approach and a special path-integral (PI) measure which depends on a free parameter, while matter is modeled by a massive self-interacting scalar field. The effective cosmological constant (CC) is given as a sum of 3 terms: the classical CC, the quantum gravity CC and the matter CC. We show that the free parameters of the theory can be chosen such that the classical CC cancels the matter CC so that the effective CC is given by the QG CC. Since the value of the quantum gravity CC is determined by the PI measure only, the PI measure parameter can be chosen such that the effective CC gives the observed value. This is equivalent to the statement that the experimentally observed CC value belongs to the spectrum of the CC operator in this QG theory. △ Less

Submitted 7 April, 2017; v1 submitted 4 July, 2014; originally announced July 2014.

Comments: 15 pages. v11: small modification of the abstract

Spectral Ranking using Seriation

Authors: Fajwel Fogel, Alexandre d'Aspremont, Milan Vojnovic

Abstract: We describe a seriation algorithm for ranking a set of items given pairwise comparisons between these items. Intuitively, the algorithm assigns similar rankings to items that compare similarly with all others. It does so by constructing a similarity matrix from pairwise comparisons, using seriation methods to reorder this matrix and construct a ranking. We first show that this spectral seriation a… ▽ More We describe a seriation algorithm for ranking a set of items given pairwise comparisons between these items. Intuitively, the algorithm assigns similar rankings to items that compare similarly with all others. It does so by constructing a similarity matrix from pairwise comparisons, using seriation methods to reorder this matrix and construct a ranking. We first show that this spectral seriation algorithm recovers the true ranking when all pairwise comparisons are observed and consistent with a total order. We then show that ranking reconstruction is still exact when some pairwise comparisons are corrupted or missing, and that seriation based spectral ranking is more robust to noise than classical scoring methods. Finally, we bound the ranking error when only a random subset of the comparions are observed. An additional benefit of the seriation formulation is that it allows us to solve semi-supervised ranking problems. Experiments on both synthetic and real datasets demonstrate that seriation based spectral ranking achieves competitive and in some cases superior performance compared to classical ranking methods. △ Less

Submitted 10 March, 2016; v1 submitted 20 June, 2014; originally announced June 2014.

Comments: Substantially revised. Accepted by JMLR

MSC Class: 62F07; 06A07; 90C27

Incentives and Efficiency in Uncertain Collaborative Environments

Authors: Yoram Bachrach, Vasilis Syrgkanis, Milan Vojnovic

Abstract: We consider collaborative systems where users make contributions across multiple available projects and are rewarded for their contributions in individual projects according to a local sharing of the value produced. This serves as a model of online social computing systems such as online Q&A forums and of credit sharing in scientific co-authorship settings. We show that the maximum feasible produc… ▽ More We consider collaborative systems where users make contributions across multiple available projects and are rewarded for their contributions in individual projects according to a local sharing of the value produced. This serves as a model of online social computing systems such as online Q&A forums and of credit sharing in scientific co-authorship settings. We show that the maximum feasible produced value can be well approximated by simple local sharing rules where users are approximately rewarded in proportion to their marginal contributions and that this holds even under incomplete information about the player's abilities and effort constraints. For natural instances we show almost 95% optimality at equilibrium. When players incur a cost for their effort, we identify a threshold phenomenon: the efficiency is a constant fraction of the optimal when the cost is strictly convex and decreases with the number of players if the cost is linear. △ Less

Submitted 5 August, 2013; originally announced August 2013.

Cosine problem in EPRL/FK spinfoam model

Authors: Marko Vo**ovic

Abstract: We calculate the classical limit effective action of the EPRL/FK spinfoam model of quantum gravity coupled to matter fields. By employing the standard QFT background field method adapted to the spinfoam setting, we find that the model has many different classical effective actions. Most notably, these include the ordinary Einstein-Hilbert action coupled to matter, but also an action which describe… ▽ More We calculate the classical limit effective action of the EPRL/FK spinfoam model of quantum gravity coupled to matter fields. By employing the standard QFT background field method adapted to the spinfoam setting, we find that the model has many different classical effective actions. Most notably, these include the ordinary Einstein-Hilbert action coupled to matter, but also an action which describes antigravity. All those multiple classical limits appear as a consequence of of the fact that the EPRL/FK vertex amplitude has cosine-like large spin asymptotics. We discuss some possible ways to eliminate the unwanted classical limits. △ Less

Submitted 11 December, 2013; v1 submitted 19 July, 2013; originally announced July 2013.

Comments: v3: Published version. 23 pages

Journal ref: Gen. Relativ. Gravit. 46, 1616 (2014)