arXiv subjects

Se-Young Yun

Publications and source records attributed to Se-Young Yun.

142 records · Page 8Linked to original sources

arXiv2019

Reinforcement with Fading Memories

We study the effect of imperfect memory on decision making in the context of a stochastic sequential action-reward problem. An agent chooses a sequence of actions which generate discrete rewards at different rates. She is allowed to make new choices at rate $β$, while past rewards disappear from her memory at rate $μ$. We focus on a family of decision rules where the agent makes a new choice by randomly selecting an action with a probability approximately proportional to the amount of past rewards associated with each action in her memory. We provide closed-form formulae for the agent's steady-state choice distribution in the regime where the memory span is large ($μ\to 0$), and show that the agent's success critically depends on how quickly she updates her choices relative to the speed of memory decay. If $β\gg μ$, the agent almost always chooses the best action, i.e., the one with the highest reward rate. Conversely, if $β\ll μ$, the agent chooses an action with a probability roughly proportional to its reward rate.

math.PR↗

arXiv2019

Clustering in Block Markov Chains

This paper considers cluster detection in Block Markov Chains (BMCs). These Markov chains are characterized by a block structure in their transition matrix. More precisely, the $n$ possible states are divided into a finite number of $K$ groups or clusters, such that states in the same cluster exhibit the same transition rates to other states. One observes a trajectory of the Markov chain, and the objective is to recover, from this observation only, the (initially unknown) clusters. In this paper we devise a clustering procedure that accurately, efficiently, and provably detects the clusters. We first derive a fundamental information-theoretical lower bound on the detection error rate satisfied under any clustering algorithm. This bound identifies the parameters of the BMC, and trajectory lengths, for which it is possible to accurately detect the clusters. We next develop two clustering algorithms that can together accurately recover the cluster structure from the shortest possible trajectories, whenever the parameters allow detection. These algorithms thus reach the fundamental detectability limit, and are optimal in that sense.

math.PR↗

arXiv2019

Spectral Approximate Inference

Given a graphical model (GM), computing its partition function is the most essential inference task, but it is computationally intractable in general. To address the issue, iterative approximation algorithms exploring certain local structure/consistency of GM have been investigated as popular choices in practice. However, due to their local/iterative nature, they often output poor approximations or even do not converge, e.g., in low-temperature regimes (hard instances of large parameters). To overcome the limitation, we propose a novel approach utilizing the global spectral feature of GM. Our contribution is two-fold: (a) we first propose a fully polynomial-time approximation scheme (FPTAS) for approximating the partition function of GM associating with a low-rank coupling matrix; (b) for general high-rank GMs, we design a spectral mean-field scheme utilizing (a) as a subroutine, where it approximates a high-rank GM into a product of rank-1 GMs for an efficient approximation of the partition function. The proposed algorithm is more robust in its running time and accuracy than prior methods, i.e., neither suffers from the convergence issue nor depends on hard local structures, as demonstrated in our experiments.

cs.LG↗

arXiv2019

Submodular Maximization using Test Scores

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.

cs.DS↗

arXiv2018

Spectrogram-channels u-net: a source separation model viewing each channel as the spectrogram of each source

Sound source separation has attracted attention from Music Information Retrieval(MIR) researchers, since it is related to many MIR tasks such as automatic lyric transcription, singer identification, and voice conversion. In this paper, we propose an intuitive spectrogram-based model for source separation by adapting U-Net. We call it Spectrogram-Channels U-Net, which means each channel of the output corresponds to the spectrogram of separated source itself. The proposed model can be used for not only singing voice separation but also multi-instrument separation by changing only the number of output channels. In addition, we propose a loss function that balances volumes between different sources. Finally, we yield performance that is state-of-the-art on both separation tasks.

cs.SD↗

arXiv2017

Parameter Estimation for Thurstone Choice Models

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.

math.ST↗

arXiv2016

Optimal Cluster Recovery in the Labeled Stochastic Block Model

We consider the problem of community detection or clustering in the labeled Stochastic Block Model (LSBM) with a finite number $K$ of clusters of sizes linearly growing with the global population of items $n$. Every pair of items is labeled independently at random, and label $\ell$ appears with probability $p(i,j,\ell)$ between two items in clusters indexed by $i$ and $j$, respectively. The objective is to reconstruct the clusters from the observation of these random labels. Clustering under the SBM and their extensions has attracted much attention recently. Most existing work aimed at characterizing the set of parameters such that it is possible to infer clusters either positively correlated with the true clusters, or with a vanishing proportion of misclassified items, or exactly matching the true clusters. We find the set of parameters such that there exists a clustering algorithm with at most $s$ misclassified items in average under the general LSBM and for any $s=o(n)$, which solves one open problem raised in \cite{abbe2015community}. We further develop an algorithm, based on simple spectral methods, that achieves this fundamental performance limit within $O(n \mbox{polylog}(n))$ computations and without the a-priori knowledge of the model parameters.

math.PR↗

arXiv2015

Streaming, Memory Limited Matrix Completion with Noise

In this paper, we consider the streaming memory-limited matrix completion problem when the observed entries are noisy versions of a small random fraction of the original entries. We are interested in scenarios where the matrix size is very large so the matrix is very hard to store and manipulate. Here, columns of the observed matrix are presented sequentially and the goal is to complete the missing entries after one pass on the data with limited memory space and limited computational complexity. We propose a streaming algorithm which produces an estimate of the original matrix with a vanishing mean square error, uses memory space scaling linearly with the ambient dimension of the matrix, i.e. the memory required to store the output alone, and spends computations as much as the number of non-zero entries of the input matrix.

math.SP↗

arXiv2015

CSMA using the Bethe Approximation: Scheduling and Utility Maximization

CSMA (Carrier Sense Multiple Access), which resolves contentions over wireless networks in a fully distributed fashion, has recently gained a lot of attentions since it has been proved that appropriate control of CSMA parameters guarantees optimality in terms of stability (i.e., scheduling) and system- wide utility (i.e., scheduling and congestion control). Most CSMA-based algorithms rely on the popular MCMC (Markov Chain Monte Carlo) technique, which enables one to find optimal CSMA parameters through iterative loops of `simulation-and-update'. However, such a simulation-based approach often becomes a major cause of exponentially slow convergence, being poorly adaptive to flow/topology changes. In this paper, we develop distributed iterative algorithms which produce approximate solutions with convergence in polynomial time for both stability and utility maximization problems. In particular, for the stability problem, the proposed distributed algorithm requires, somewhat surprisingly, only one iteration among links. Our approach is motivated by the Bethe approximation (introduced by Yedidia, Freeman and Weiss in 2005) allowing us to express approximate solutions via a certain non-linear system with polynomial size. Our polynomial convergence guarantee comes from directly solving the non-linear system in a distributed manner, rather than multiple simulation-and-update loops in existing algorithms. We provide numerical results to show that the algorithm produces highly accurate solutions and converges much faster than the prior ones.

cs.NI↗

arXiv2014

Accurate Community Detection in the Stochastic Block Model via Spectral Algorithms

We consider the problem of community detection in the Stochastic Block Model with a finite number $K$ of communities of sizes linearly growing with the network size $n$. This model consists in a random graph such that each pair of vertices is connected independently with probability $p$ within communities and $q$ across communities. One observes a realization of this random graph, and the objective is to reconstruct the communities from this observation. We show that under spectral algorithms, the number of misclassified vertices does not exceed $s$ with high probability as $n$ grows large, whenever $pn=ω(1)$, $s=o(n)$ and \begin{equation*} \lim\inf_{n\to\infty} {n(α_1 p+α_2 q-(α_1 + α_2)p^{\frac{α_1}{α_1 + α_2}}q^{\frac{α_2}{α_1 + α_2}})\over \log (\frac{n}{s})} >1,\quad\quad(1) \end{equation*} where $α_1$ and $α_2$ denote the (fixed) proportions of vertices in the two smallest communities. In view of recent work by Abbe et al. and Mossel et al., this establishes that the proposed spectral algorithms are able to exactly recover communities whenever this is at all possible in the case of networks with two communities with equal sizes. We conjecture that condition (1) is actually necessary to obtain less than $s$ misclassified vertices asymptotically, which would establish the optimality of spectral method in more general scenarios.

cs.SI↗

arXiv2014

Distributed Load Balancing in Heterogeneous Systems

We consider the problem of distributed load balancing in heterogenous parallel server systems, where the service rate achieved by a user at a server depends on both the user and the server. Such heterogeneity typically arises in wireless networks (e.g., servers may represent frequency bands, and the service rate of a user varies across bands). Users select servers in a distributed manner. They initially attach to an arbitrary server. However, at random instants of time, they may probe the load at a new server and migrate there to improve their service rate. We analyze the system dynamics under the natural Random Local Search (RLS) migration scheme, introduced in \cite{sig10}. Under this scheme, when a user has the opportunity to switch servers, she does it only if this improves her service rate. The dynamics under RLS may be interpreted as those generated by strategic players updating their strategy in a load balancing game. In closed systems, where the user population is fixed, we show that this game has pure Nash Equilibriums (NEs), and we analyze their efficiency. We further prove that when the user population grows large, pure NEs get closer to a Proportionally Fair (PF) allocation of users to servers, and we characterize the gap between equilibriums and this ideal allocation depending on user population. Under the RLS algorithm, the system converges to pure NEs: we study the time it takes for the system to reach the PF allocation within a certain margin. In open systems, where users randomly enter the system and leave upon service completion, we establish that the RLS algorithm stabilizes the system whenever this it at all possible, i.e., it is throughput-optimal.

cs.GT↗

arXiv2014

Streaming, Memory Limited Algorithms for Community Detection

In this paper, we consider sparse networks consisting of a finite number of non-overlapping communities, i.e. disjoint clusters, so that there is higher density within clusters than across clusters. Both the intra- and inter-cluster edge densities vanish when the size of the graph grows large, making the cluster reconstruction problem nosier and hence difficult to solve. We are interested in scenarios where the network size is very large, so that the adjacency matrix of the graph is hard to manipulate and store. The data stream model in which columns of the adjacency matrix are revealed sequentially constitutes a natural framework in this setting. For this model, we develop two novel clustering algorithms that extract the clusters asymptotically accurately. The first algorithm is {\it offline}, as it needs to store and keep the assignments of nodes to clusters, and requires a memory that scales linearly with the network size. The second algorithm is {\it online}, as it may classify a node when the corresponding column is revealed and then discard this information. This algorithm requires a memory growing sub-linearly with the network size. To construct these efficient streaming memory-limited clustering algorithms, we first address the problem of clustering with partial information, where only a small proportion of the columns of the adjacency matrix is observed and develop, for this setting, a new spectral algorithm which is of independent interest.

cs.SI↗

arXiv2014

Community Detection via Random and Adaptive Sampling

In this paper, we consider networks consisting of a finite number of non-overlapping communities. To extract these communities, the interaction between pairs of nodes may be sampled from a large available data set, which allows a given node pair to be sampled several times. When a node pair is sampled, the observed outcome is a binary random variable, equal to 1 if nodes interact and to 0 otherwise. The outcome is more likely to be positive if nodes belong to the same communities. For a given budget of node pair samples or observations, we wish to jointly design a sampling strategy (the sequence of sampled node pairs) and a clustering algorithm that recover the hidden communities with the highest possible accuracy. We consider both non-adaptive and adaptive sampling strategies, and for both classes of strategies, we derive fundamental performance limits satisfied by any sampling and clustering algorithm. In particular, we provide necessary conditions for the existence of algorithms recovering the communities accurately as the network size grows large. We also devise simple algorithms that accurately reconstruct the communities when this is at all possible, hence proving that the proposed necessary conditions for accurate community detection are also sufficient. The classical problem of community detection in the stochastic block model can be seen as a particular instance of the problems consider here. But our framework covers more general scenarios where the sequence of sampled node pairs can be designed in an adaptive manner. The paper provides new results for the stochastic block model, and extends the analysis to the case of adaptive sampling.

cs.SI↗

arXiv2013

On the Delay Scaling Laws of Cache Networks

The Internet is becoming more and more content-oriented, where one of main components in content-oriented Internet architectures is network caching. Despite a surge of extensive use of network cashing in the current and future Internet architectures, analysis on the performance of general cache networks are still quite limited due to complex inter-plays among various components and thus analytical intractability. We study asymptotic delay performance of cache networks, in particular, focusing on the impact of heterogeneous content popularities and nodes' geometric `importances' in caching policies. Our theoretical findings provide useful engineering implications such as when and how various factors have impact on caching performance, and we provide extensive simulation results on the real Internet topology.

cs.NI↗

arXiv2013

CSMA over Time-varying Channels: Optimality, Uniqueness and Limited Backoff Rate

Recent studies on MAC scheduling have shown that carrier sense multiple access (CSMA) algo- rithms can be throughput optimal for arbitrary wireless network topology. However, these results are highly sensitive to the underlying assumption on 'static' or 'fixed' system conditions. For example, if channel conditions are time-varying, it is unclear how each node can adjust its CSMA parameters, so-called backoff and channel holding times, using its local channel information for the desired high performance. In this paper, we study 'channel-aware' CSMA (A-CSMA) algorithms in time-varying channels, where they adjust their parameters as some function of the current channel capacity. First, we show that the achievable rate region of A-CSMA equals to the maximum rate region if and only if the function is exponential. Furthermore, given an exponential function in A-CSMA, we design updating rules for their parameters, which achieve throughput optimality for an arbitrary wireless network topology. They are the first CSMA algorithms in the literature which are proved to be throughput optimal under time-varying channels. Moreover, we also consider the case when back-off rates of A- CSMA are highly restricted compared to the speed of channel variations, and characterize the throughput performance of A-CSMA in terms of the underlying wireless network topology. Our results not only guide a high-performance design on MAC scheduling under highly time-varying scenarios, but also provide new insights on the performance of CSMA algorithms in relation to their backoff rates and the network topology.

cs.NI↗

arXiv2012

Medium Access over Time-varying Channels with Limited Sensing Cost

This paper has been withdrawn because of some paper issues. Recent studies on MAC scheduling have shown that carrier sense multiple access (CSMA) can be controlled to be achieve optimality in terms of throughput or utility. These results imply that just a simple MAC algorithm without message passing is possible to achieve high performance guarantee. However, such studies are conducted only on the assumption that channel conditions are static. Noting that the main drive for achieving optimality in optimal CSMA is to let it run a good schedule for some time, formally referred to as the mixing time, it is under-explored how such optimal CSMA performs for time-varying channel conditions. In this paper, under the practical constraint of restricted back-off rates (i.e., limited sensing speed), we consider two versions of CSMAs: (i) channel-unaware CSMA (U-CSMA) and (ii) channel-aware CSMA (A-CSMA), each of which is characterized as its ability of tracking channel conditions. We first show that for fast channel variations, A-CSMA achieves almost zero throughput, implying that incomplete tracking of channel conditions may seriously degrade performance, whereas U-CSMA, accessing the media without explicit consideration of channel conditions, has positive worst-case guarantee in throughput, where the ratio of guarantee depends on network topology. On the other hand, for slow channel variations, we prove that A-CSMA is throughput-optimal for any network topology. Our results provide the precise trade-off between sensing costs and performances of CSMA algorithms, which guides a robust design on MAC scheduling under highly time-varying scenarios.

cs.NI↗