Search

In this paper, we study user association and wireless bandwidth allocation for a hierarchical federated learning system that consists of mobile users, edge servers, and a cloud server. To minimize the length of a global round in hierarchical federated learning with equal bandwidth allocation, we formulate a combinatorial optimization problem. We design the twin sorting dynamic programming (TSDP) algorithm that obtains a globally optimal solution in polynomial time when there are two edge servers. In addition, we put forward the TSDP-assisted algorithm for user association when there are three or more edge servers. Furthermore, given a user association matrix, we formulate and solve a convex optimization problem for optimal wireless bandwidth allocation. Simulation results show that the proposed approach outperforms a number of alternative schemes., Comment: 14 pages

This paper studies the structure of graphs with given tree-width and excluding a fixed complete bipartite subgraph, which generalises the bounded degree setting. We give a new structural description of such graphs in terms of so-called quasi-tree-partitions. We demonstrate the utility of this result through applications to (fractional) clustered colouring. Further generalisations of these structural and colouring results are presented.

Motivated by Hadwiger's conjecture and related problems for list-coloring, we study graphs $H$ for which every graph with minimum degree at least $|V(H)|-1$ contains $H$ as a minor. We prove that a large class of apex-outerplanar graphs satisfies this property. Our result gives the first examples of such graphs whose vertex cover numbers are significantly larger than half of the number of its vertices, which breaks a barrier for attacking related coloring problems via extremal functions, and recovers all known such graphs that have arbitrarily large maximum degree. Our proof can be adapted to directed graphs to show that if $\vec H$ is the digraph obtained from a directed cycle or an in-arborescence by adding an apex source, then every digraph with minimum out-degree $|V(\vec H)|-1$ contains $\vec H$ as a subdivision or a butterfly minor respectively. These results provide the optimal upper bound for the chromatic number and dichromatic number of graphs and digraphs that do not contain the aforementioned graphs or digraphs as a minor, butterfly minor and a subdivision, respectively. Special cases of our results solve an open problem of Aboulker, Cohen, Havet, Lochet, Moura and Thomass\'{e} and strengthen results of Gishboliner, Steiner and Szab\'{o}.

We introduce peaceful colourings, a variant of $h$-conflict free colourings. We call a colouring with no monochromatic edges $p$-peaceful if for each vertex $v$, there are at most $p$ neighbours of $v$ coloured with a colour appearing on another neighbour of $v$. An $h$-conflict-free colouring of a graph is a (vertex)-colouring with no monochromatic edges so that for every vertex $v$, the number of neighbours of $v$ which are coloured with a colour appearing on no other neighbour of $v$ is at least the minimum of $h$ and the degree of $v$. If $G$ is $\Delta$-regular then it has an $h$-conflict free colouring precisely if it has a $(\Delta-h)$-peaceful colouring. We focus on the minimum $p_\Delta$ of those $p$ for which every graph of maximum degree $\Delta$ has a $p$-peaceful colouring with $\Delta+1$ colours. We show that $p_\Delta > (1-\frac{1}{e}-o(1))\Delta$ and that for graphs of bounded codegree, $p_\Delta \leq (1-\frac{1}{e}+o(1))\Delta$. We ask if the latter result can be improved by dropping the bound on the codegree. As a partial result, we show that $p_\Delta \leq \frac{8000}{8001}\Delta$ for sufficiently large $\Delta$.

A proper conflict-free colouring of a graph is a colouring of the vertices such that any two adjacent vertices receive different colours, and for every non-isolated vertex $v$, some colour appears exactly once on the neighbourhood of $v$. Caro, Petru\v{s}evski and \v{S}krekovski conjectured that every connected graph with maximum degree $\Delta \geq 3$ has a proper conflict-free colouring with at most $\Delta+1$ colours. This conjecture holds for $\Delta=3$ and remains open for $\Delta \geq 4$. In this paper we prove that this conjecture holds asymptotically; namely, every graph with maximum degree $\Delta$ has a proper conflict-free colouring with $(1+o(1))\Delta$ colours.

Weak diameter coloring of graphs recently attracted attention partially due to its connection to asymptotic dimension of metric spaces. We consider weak diameter list-coloring of graphs in this paper. Dvo\v{r}\'{a}k and Norin proved that graphs with bounded Euler genus are 3-choosable with bounded weak diameter. In this paper, we extend their result by showing that for every graph $H$, $H$-minor free graphs are 3-choosable with bounded weak diameter. The upper bound 3 is optimal and it strengthens an earlier result for non-list-coloring $H$-minor free graphs with bounded weak diameter. As a corollary, $H$-minor free graphs with bounded maximum degree are 3-choosable with bounded clustering, strengthening an earlier result for non-list-coloring. When $H$ is planar, we prove a much stronger result: for every 2-list-assignment $L$ of an $H$-minor free graph, every precoloring with bounded weak diameter can be extended to an $L$-coloring with bounded weak diameter. As a corollary, for any planar graph $H$ and $H$-minor free graph $G$, there are exponentially many list-colorings of $G$ with bounded weak diameter (and with bounded clustering if $G$ also has bounded maximum degree); and every graph with bounded layered tree-width and bounded maximum degree has exponentially many 3-colorings with bounded clustering. We also show that the aforementioned results for list-coloring cannot be extended to odd minor free graphs by showing that some bipartite graphs with maximum degree $\Delta$ are $k$-choosable with bounded weak diameter only when $k=\Omega(\log\Delta/\log\log\Delta)$. On the other hand, we show that odd $H$-minor graphs are 3-colorable with bounded weak diameter, implying an earlier result about clustered coloring of odd $H$-minor free graphs with bounded maximum degree.

Assouad-Nagata dimension addresses both large and small scale behaviors of metric spaces and is a refinement of Gromov's asymptotic dimension. A metric space $M$ is a minor-closed metric if there exists an (edge-)weighted graph $G$ satisfying a fixed minor-closed property such that the underlying space of $M$ is the vertex-set of $G$, and the metric of $M$ is the distance function in $G$. Minor-closed metrics naturally arise when removing redundant edges of the underlying graphs by using edge-deletion and edge-contraction. In this paper, we determine the Assouad-Nagata dimension of every minor-closed metric. Our main theorem simultaneously generalizes known results about the asymptotic dimension of $H$-minor free unweighted graphs and about the Assouad-Nagata dimension of complete Riemannian surfaces with finite Euler genus.

Terahertz (THz) communication with ultra-wide available spectrum is a promising technique that can achieve the stringent requirement of high data rate in the next-generation wireless networks, yet its severe propagation attenuation significantly hinders its implementation in practice. Finding beam directions for a large-scale antenna array to effectively overcome severe propagation attenuation of THz signals is a pressing need. This paper proposes a novel approach of federated deep reinforcement learning (FDRL) to swiftly perform THz-beam search for multiple base stations (BSs) coordinated by an edge server in a cellular network. All the BSs conduct deep deterministic policy gradient (DDPG)-based DRL to obtain THz beamforming policy with limited channel state information (CSI). They update their DDPG models with hidden information in order to mitigate inter-cell interference. We demonstrate that the cell network can achieve higher throughput as more THz CSI and hidden neurons of DDPG are adopted. We also show that FDRL with partial model update is able to nearly achieve the same performance of FDRL with full model update, which indicates an effective means to reduce communication load between the edge server and the BSs by partial model uploading. Moreover, the proposed FDRL outperforms conventional non-learning-based and existing non-FDRL benchmark optimization methods.

A proper coloring of a graph is \emph{conflict-free} if, for every non-isolated vertex, some color is used exactly once on its neighborhood. Caro, Petru\v{s}evski, and \v{S}krekovski proved that every graph $G$ has a proper conflict-free coloring with at most $5\Delta(G)/2$ colors and conjectured that $\Delta(G)+1$ colors suffice for every connected graph $G$ with $\Delta(G)\ge 3$. Our first main result is that even for list-coloring, $\left\lceil 1.6550826\Delta(G)+\sqrt{\Delta(G)}\right\rceil$ colors suffice for every graph $G$ with $\Delta(G)\ge 10^{8}$; we also prove slightly weaker bounds for all graphs with $\Delta(G)\ge 750$. These results follow from our more general framework on proper conflict-free list-coloring of a pair consisting of a graph $G$ and a "conflict" hypergraph ${\mathcal H}$. As another corollary of our results in this general framework, every graph has a proper $(\sqrt{30}+o(1))\Delta(G)^{1.5}$-list-coloring such that every bi-chromatic component is a path on at most three vertices, where the number of colors is optimal up to a constant factor. Our proof uses a fairly new type of recursive counting argument called Rosenfeld counting, which is a variant of the Lov\'{a}sz Local Lemma or entropy compression. We also prove an asymptotically optimal result for a fractional analogue of our general framework for proper conflict-free coloring for pairs of a graph and a conflict hypergraph. A corollary states that every graph $G$ has a fractional $(1+o(1))\Delta(G)$-coloring such that every fractionally bi-chromatic component has at most two vertices. In particular, it implies that the fractional analogue of the conjecture of Caro et al.\ holds asymptotically in a strong sense.

The clustering of a graph coloring is the maximum size of monochromatic components. This paper studies colorings with bounded clustering in graph classes with bounded layered treewidth, which include planar graphs, graphs of bounded Euler genus, graphs embeddable on a fixed surface with a bounded number of crossings per edge, map graphs, amongst other examples. Our main theorem says that every graph with layered treewidth at most $k$ and with maximum degree at most $\Delta$ is $3$-colorable with clustering $O(k^{19}\Delta^{37})$. This is the first known polynomial bound on the clustering. This greatly improves upon a corresponding result of Esperet and Joret for graphs of bounded genus., Comment: This paper is extracted from arXiv:1905.08969. The corresponding part in arXiv:1905.08969 will be deleted

The defective chromatic number of a graph class is the infimum $k$ such that there exists an integer $d$ such that every graph in this class can be partitioned into at most $k$ induced subgraphs with maximum degree at most $d$. Finding the defective chromatic number is a fundamental graph partitioning problem and received attention recently partially due to Hadwiger's conjecture about coloring minor-closed families. In this paper, we prove that the defective chromatic number of any minor-closed family equals the simple lower bound obtained by the standard construction, confirming a conjecture of Ossona de Mendez, Oum, and Wood. This result provides the optimal list of unavoidable finite minors for infinite graphs that cannot be partitioned into a fixed finite number of induced subgraphs with uniformly bounded maximum degree. As corollaries about clustered coloring, we obtain a linear relation between the clustered chromatic number of any minor-closed family and the tree-depth of its forbidden minors, improving an earlier exponential bound proved by Norin, Scott, Seymour, and Wood and confirming the planar case of their conjecture.

Sparse subspace clustering (SSC) using greedy-based neighbor selection, such as orthogonal matching pursuit (OMP), has been known as a popular computationally-efficient alternative to the popular L1-minimization based methods. This paper proposes a new SSC scheme using generalized OMP (GOMP), a soup-up of OMP whereby multiple neighbors are identified per iteration, along with a new stopping rule requiring nothing more than a knowledge of the ambient signal dimension. Compared to conventional OMP, which identifies one neighbor per iteration, the proposed GOMP method involves fewer iterations, thereby enjoying lower algorithmic complexity; advantageously, the proposed stopping rule is free from off-line estimation of subspace dimension and noise power. Under the semi-random model, analytic performance guarantees, in terms of neighbor recovery rates, are established to justify the advantage of the proposed GOMP. The results show that, with a high probability, GOMP (i) is halted by the proposed stopping rule, and (ii) can retrieve more true neighbors than OMP, consequently yielding higher final data clustering accuracy. Computer simulations using both synthetic data and real human face data are provided to validate our analytic study and evidence the effectiveness of the proposed approach.

Proper conflict-free coloring is an intermediate notion between proper coloring of a graph and proper coloring of its square. It is a proper coloring such that for every non-isolated vertex, there exists a color appearing exactly once in its (open) neighborhood. Typical examples of graphs with large proper conflict-free chromatic number include graphs with large chromatic number and bipartite graphs isomorphic to the $1$-subdivision of graphs with large chromatic number. In this paper, we prove two rough converse statements that hold even in the list-coloring setting. The first is for sparse graphs: for every graph $H$, there exists an integer $c_H$ such that every graph with no subdivision of $H$ is (properly) conflict-free $c_H$-choosable. The second applies to dense graphs: every graph with large conflict-free choice number either contains a large complete graph as an odd minor or contains a bipartite induced subgraph that has large conflict-free choice number. These give two incomparable (partial) answers of a question of Caro, Petru\v{s}evski and \v{S}krekovski. We also prove quantitatively better bounds for minor-closed families, implying some known results about proper conflict-free coloring and odd coloring in the literature. Moreover, we prove that every graph with layered treewidth at most $w$ is (properly) conflict-free $(8w-1)$-choosable. This result applies to $(g,k)$-planar graphs, which are graphs whose coloring problems have attracted attention recently., Comment: Hickingbotham recently independently announced a paper (arXiv:2203.10402) proving a result similar to the ones in this paper. Please see the notes at the end of this paper for details. v2: add results for odd minors, which applies to graphs with unbounded degeneracy, and change the title of the paper

This paper presents a novel approach to conduct highly efficient federated learning (FL) over a massive wireless edge network, where an edge server and numerous mobile devices (clients) jointly learn a global model without transporting the huge amount of data collected by the mobile devices to the edge server. The proposed FL approach is referred to as spatio-temporal FL (STFL), which jointly exploits the spatial and temporal correlations between the learning updates from different mobile devices scheduled to join STFL in various training epochs. The STFL model not only represents the realistic intermittent learning behavior from the edge server to the mobile devices due to data delivery outage, but also features a mechanism of compensating loss learning updates in order to mitigate the impacts of intermittent learning. An analytical framework of STFL is proposed and employed to study the learning capability of STFL via its convergence performance. In particular, we have assessed the impact of data delivery outage, intermittent learning mitigation, and statistical heterogeneity of datasets on the convergence performance of STFL. The results provide crucial insights into the design and analysis of STFL-based wireless networks., Comment: 3 figures, conference

Federated learning (FL) is a promising distributed learning technique particularly suitable for wireless learning scenarios since it can accomplish a learning task without raw data transportation so as to preserve data privacy and lower network resource consumption. However, current works on FL over wireless networks do not profoundly study the fundamental performance of FL over wireless networks that suffers from communication outage due to channel impairment and network interference. To accurately exploit the performance of FL over wireless networks, this paper proposes a novel intermittent FL model over a cellular-connected unmanned aerial vehicle (UAV) network, which characterizes communication outage from UAV (clients) to their server and data heterogeneity among the datasets at UAVs. We propose an analytically tractable framework to derive the uplink outage probability and use it to devise a simulation-based approach so as to evaluate the performance of the proposed intermittent FL model. Our findings reveal how the intermittent FL model is impacted by uplink communication outage and UAV deployment. Extensive numerical simulations are provided to show the consistency between the simulated and analytical performances of the proposed intermittent FL model., Comment: 6 pages, 3 figures, conference

This paper studies how to achieve a high and flexible coverage performance of a large-scale cellular network that enables unmanned aerial vehicles (UAVs) for non-orthogonal multiple access (NOMA) transmission to simultaneously serve multiple users. The considered cellular network consists of a tier of base stations and a tier of UAVs. Each UAV is mounted with an intelligent reflecting surface (IRS) in order to serve as an aerial IRS reflecting signals between a base station and a user in the network. All the UAVs in the network are deployed based on a newly proposed three-dimensional (3D) point process that leads to a tractable and accurate analysis of the association statistics, which is traditionally difficult to analyze due to the mobility of UAVs. In light of this, we are able to analyze the downlink coverage of UAV-IRS-assisted NOMA transmission for two users and derive the corresponding coverage probabilities. Our coverage analyses shed light on the optimal allocations of transmit power between NOMA users and UAVs to accomplish the goal of ubiquitous and flexible NOMA transmission. We also conduct numerical simulations to validate our coverage analytical results while demonstrating the improved coverage performance achieved by aerial IRSs., Comment: 6 pages; 3 figures; conference