arXiv subjects

Jiaming Xu

Publications and source records attributed to Jiaming Xu.

118 records · Page 7Linked to original sources

arXiv2015

Statistical-Computational Tradeoffs in Planted Problems and Submatrix Localization with a Growing Number of Clusters and Submatrices

We consider two closely related problems: planted clustering and submatrix localization. The planted clustering problem assumes that a random graph is generated based on some underlying clusters of the nodes; the task is to recover these clusters given the graph. The submatrix localization problem concerns locating hidden submatrices with elevated means inside a large real-valued random matrix. Of particular interest is the setting where the number of clusters/submatrices is allowed to grow unbounded with the problem size. These formulations cover several classical models such as planted clique, planted densest subgraph, planted partition, planted coloring, and stochastic block model, which are widely used for studying community detection and clustering/bi-clustering. For both problems, we show that the space of the model parameters (cluster/submatrix size, cluster density, and submatrix mean) can be partitioned into four disjoint regions corresponding to decreasing statistical and computational complexities: (1) the \emph{impossible} regime, where all algorithms fail; (2) the \emph{hard} regime, where the computationally expensive Maximum Likelihood Estimator (MLE) succeeds; (3) the \emph{easy} regime, where the polynomial-time convexified MLE succeeds; (4) the \emph{simple} regime, where a simple counting/thresholding procedure succeeds. Moreover, we show that each of these algorithms provably fails in the previous harder regimes. Our theorems establish the minimax recovery limit, which are tight up to constants and hold with a growing number of clusters/submatrices, and provide a stronger performance guarantee than previously known for polynomial-time algorithms. Our study demonstrates the tradeoffs between statistical and computational considerations, and suggests that the minimax recovery limit may not be achievable by polynomial-time algorithms.

stat.ML↗

arXiv2015

Computational Lower Bounds for Community Detection on Random Graphs

This paper studies the problem of detecting the presence of a small dense community planted in a large Erdős-Rényi random graph $\mathcal{G}(N,q)$, where the edge probability within the community exceeds $q$ by a constant factor. Assuming the hardness of the planted clique detection problem, we show that the computational complexity of detecting the community exhibits the following phase transition phenomenon: As the graph size $N$ grows and the graph becomes sparser according to $q=N^{-α}$, there exists a critical value of $α= \frac{2}{3}$, below which there exists a computationally intensive procedure that can detect far smaller communities than any computationally efficient procedure, and above which a linear-time procedure is statistically optimal. The results also lead to the average-case hardness results for recovering the dense community and approximating the densest $K$-subgraph.

math.ST↗

arXiv2015

Short Text Hashing Improved by Integrating Multi-Granularity Topics and Tags

Due to computational and storage efficiencies of compact binary codes, hashing has been widely used for large-scale similarity search. Unfortunately, many existing hashing methods based on observed keyword features are not effective for short texts due to the sparseness and shortness. Recently, some researchers try to utilize latent topics of certain granularity to preserve semantic similarity in hash codes beyond keyword matching. However, topics of certain granularity are not adequate to represent the intrinsic semantic information. In this paper, we present a novel unified approach for short text Hashing using Multi-granularity Topics and Tags, dubbed HMTT. In particular, we propose a selection method to choose the optimal multi-granularity topics depending on the type of dataset, and design two distinct hashing strategies to incorporate multi-granularity topics. We also propose a simple and effective method to exploit tags to enhance the similarity of related texts. We carry out extensive experiments on one short text dataset as well as on one normal text dataset. The results demonstrate that our approach is effective and significantly outperforms baselines on several evaluation metrics.

cs.IR↗

arXiv2015

Reconstruction in the Labeled Stochastic Block Model

The labeled stochastic block model is a random graph model representing networks with community structure and interactions of multiple types. In its simplest form, it consists of two communities of approximately equal size, and the edges are drawn and labeled at random with probability depending on whether their two endpoints belong to the same community or not. It has been conjectured in \cite{Heimlicher12} that correlated reconstruction (i.e.\ identification of a partition correlated with the true partition into the underlying communities) would be feasible if and only if a model parameter exceeds a threshold. We prove one half of this conjecture, i.e., reconstruction is impossible when below the threshold. In the positive direction, we introduce a weighted graph to exploit the label information. With a suitable choice of weight function, we show that when above the threshold by a specific constant, reconstruction is achieved by (1) minimum bisection, (2) a semidefinite relaxation of minimum bisection, and (3) a spectral method combined with removal of edges incident to vertices of high degree. Furthermore, we show that hypothesis testing between the labeled stochastic block model and the labeled Erdős-Rényi random graph model exhibits a phase transition at the conjectured reconstruction threshold.

stat.ML↗

arXiv2014

Edge Label Inference in Generalized Stochastic Block Models: from Spectral Theory to Impossibility Results

The classical setting of community detection consists of networks exhibiting a clustered structure. To more accurately model real systems we consider a class of networks (i) whose edges may carry labels and (ii) which may lack a clustered structure. Specifically we assume that nodes possess latent attributes drawn from a general compact space and edges between two nodes are randomly generated and labeled according to some unknown distribution as a function of their latent attributes. Our goal is then to infer the edge label distributions from a partially observed network. We propose a computationally efficient spectral algorithm and show it allows for asymptotically correct inference when the average node degree could be as low as logarithmic in the total number of nodes. Conversely, if the average node degree is below a specific constant threshold, we show that no algorithm can achieve better inference than guessing without using the observations. As a byproduct of our analysis, we show that our model provides a general procedure to construct random graph models with a spectrum asymptotic to a pre-specified eigenvalue distribution such as a power-law distribution.

math.ST↗

arXiv2014

Minimax-optimal Inference from Partial Rankings

This paper studies the problem of inferring a global preference based on the partial rankings provided by many users over different subsets of items according to the Plackett-Luce model. A question of particular interest is how to optimally assign items to users for ranking and how many item assignments are needed to achieve a target estimation error. For a given assignment of items to users, we first derive an oracle lower bound of the estimation error that holds even for the more general Thurstone models. Then we show that the Cramér-Rao lower bound and our upper bounds inversely depend on the spectral gap of the Laplacian of an appropriately defined comparison graph. When the system is allowed to choose the item assignment, we propose a random assignment scheme. Our oracle lower bound and upper bounds imply that it is minimax-optimal up to a logarithmic factor among all assignment schemes and the lower bound can be achieved by the maximum likelihood estimator as well as popular rank-breaking schemes that decompose partial rankings into pairwise comparisons. The numerical experiments corroborate our theoretical findings.

stat.ML↗

arXiv2014

Jointly Clustering Rows and Columns of Binary Matrices: Algorithms and Trade-offs

In standard clustering problems, data points are represented by vectors, and by stacking them together, one forms a data matrix with row or column cluster structure. In this paper, we consider a class of binary matrices, arising in many applications, which exhibit both row and column cluster structure, and our goal is to exactly recover the underlying row and column clusters by observing only a small fraction of noisy entries. We first derive a lower bound on the minimum number of observations needed for exact cluster recovery. Then, we propose three algorithms with different running time and compare the number of observations needed by them for successful cluster recovery. Our analytical results show smooth time-data trade-offs: one can gradually reduce the computational complexity when increasingly more observations are available.

stat.ML↗

arXiv2012

The Supermarket Game

A supermarket game is considered with $N$ FCFS queues with unit exponential service rate and global Poisson arrival rate $N λ$. Upon arrival each customer chooses a number of queues to be sampled uniformly at random and joins the least loaded sampled queue. Customers are assumed to have cost for both waiting and sampling, and they want to minimize their own expected total cost. We study the supermarket game in a mean field model that corresponds to the limit as $N$ converges to infinity in the sense that (i) for a fixed symmetric customer strategy, the joint equilibrium distribution of any fixed number of queues converges as $N \to \infty$ to a product distribution determined by the mean field model and (ii) a Nash equilibrium for the mean field model is an $ε$-Nash equilibrium for the finite $N$ model with $N$ sufficiently large. It is shown that there always exists a Nash equilibrium for $λ<1$ and the Nash equilibrium is unique with homogeneous waiting cost for $λ^2 \le 1/2$. Furthermore, we find that the action of sampling more queues by some customers has a positive externality on the other customers in the mean field model, but can have a negative externality for finite $N$.

cs.IT↗

arXiv2011

Broadcast Channels with Delayed Finite-Rate Feedback: Predict or Observe?

Most multiuser precoding techniques require accurate transmitter channel state information (CSIT) to maintain orthogonality between the users. Such techniques have proven quite fragile in time-varying channels because the CSIT is inherently imperfect due to estimation and feedback delay, as well quantization noise. An alternative approach recently proposed by Maddah-Ali and Tse (MAT) allows for significant multiplexing gain in the multi-input single-output (MISO) broadcast channel (BC) even with transmit CSIT that is completely stale, i.e. uncorrelated with the current channel state. With $K$ users, their scheme claims to lose only a $\log(K)$ factor relative to the full $K$ degrees of freedom (DoF) attainable in the MISO BC with perfect CSIT for large $K$. However, their result does not consider the cost of the feedback, which is potentially very large in high mobility (short channel coherence time). In this paper, we more closely examine the MAT scheme and compare its DoF gain to single user transmission (which always achieves 1 DoF) and partial CSIT linear precoding (which achieves up to $K$). In particular, assuming the channel coherence time is $N$ symbol periods and the feedback delay is $N_{\rm fd}$ we show that when $N < (1+o(1)) K \log K$ (short coherence time), single user transmission performs best, whereas for $N> (1+o(1)) (N_{\rm fd}+ K / \log K)(1-\log^{-1}K)^{-1}$ (long coherence time), zero-forcing precoding outperforms the other two. The MAT scheme is optimal for intermediate coherence times, which for practical parameter choices is indeed quite a large and significant range, even accounting for the feedback cost.

cs.IT↗

arXiv2010

On the Accuracy of the Wyner Model in Cellular Networks

The Wyner model has been widely used to model and analyze cellular networks due to its simplicity and analytical tractability. Its key aspects include fixed user locations and the deterministic and homogeneous interference intensity. While clearly a significant simplification of a real cellular system, which has random user locations and interference levels that vary by several orders of magnitude over a cell, a common presumption by theorists is that the Wyner model nevertheless captures the essential aspects of cellular interactions. But is this true? To answer this question, we consider both uplink and downlink transmissions, and both outage-based and average-based metrics. For the uplink, for both metrics, we conclude that the Wyner model is in fact quite accurate for systems with a sufficient number of simultaneous users, e.g. CDMA. Conversely, it is broadly inaccurate otherwise. With multicell processing, intracell TDMA is shown to be suboptimal in terms of average throughput, in sharp contrast to predictions using the Wyner model. Turning to the downlink, the Wyner model is highly inaccurate for outage since it depends largely on the user locations. However, for average or sum throughput, the Wyner model serves as an acceptable simplification in certain special cases if the interference parameter is set appropriately.

cs.IT↗