arXiv subjects

Lili Su

Publications and source records attributed to Lili Su.

58 records · Page 4Linked to original sources

arXiv2015

Byzantine Multi-Agent Optimization: Part II

In Part I of this report, we introduced a Byzantine fault-tolerant distributed optimization problem whose goal is to optimize a sum of convex (cost) functions with real-valued scalar input/ouput. In this second part, we introduce a condition-based variant of the original problem over arbitrary directed graphs. Specifically, for a given collection of $k$ input functions $h_1(x), \ldots, h_k(x)$, we consider the scenario when the local cost function stored at agent $j$, denoted by $g_j(x)$, is formed as a convex combination of the $k$ input functions $h_1(x), \ldots, h_k(x)$. The goal of this condition-based problem is to generate an output that is an optimum of $\frac{1}{k}\sum_{i=1}^k h_i(x)$. Depending on the availability of side information at each agent, two slightly different variants are considered. We show that for a given graph, the problem can indeed be solved despite the presence of faulty agents. In particular, even in the absence of side information at each agent, when adequate redundancy is available in the optima of input functions, a distributed algorithm is proposed in which each agent carries minimal state across iterations.

cs.DC↗

arXiv2015

Reaching Approximate Byzantine Consensus with Multi-hop Communication

We address the problem of reaching consensus in the presence of Byzantine faults. In particular, we are interested in investigating the impact of messages relay on the network connectivity for a correct iterative approximate Byzantine consensus algorithm to exist. The network is modeled by a simple directed graph. We assume a node can send messages to another node that is up to $l$ hops away via forwarding by the intermediate nodes on the routes, where $l\in \mathbb{N}$ is a natural number. We characterize the necessary and sufficient topological conditions on the network structure. The tight conditions we found are consistent with the tight conditions identified for $l=1$, where only local communication is allowed, and are strictly weaker for $l>1$. Let $l^*$ denote the length of a longest path in the given network. For $l\ge l^*$ and undirected graphs, our conditions hold if and only if $n\ge 3f+1$ and the node-connectivity of the given graph is at least $2f+1$ , where $n$ is the total number of nodes and $f$ is the maximal number of Byzantine nodes; and for $l\ge l^*$ and directed graphs, our conditions is equivalent to the tight condition found for exact Byzantine consensus. Our sufficiency is shown by constructing a correct algorithm, wherein the trim function is constructed based on investigating a newly introduced minimal messages cover property. The trim function proposed also works over multi-graphs.

cs.DC↗

arXiv2014

Computing Similarity Distances Between Rankings

We address the problem of computing distances between rankings that take into account similarities between candidates. The need for evaluating such distances is governed by applications as diverse as rank aggregation, bioinformatics, social sciences and data storage. The problem may be summarized as follows: Given two rankings and a positive cost function on transpositions that depends on the similarity of the candidates involved, find a smallest cost sequence of transpositions that converts one ranking into another. Our focus is on costs that may be described via special metric-tree structures and on complete rankings modeled as permutations. The presented results include a quadratic-time algorithm for finding a minimum cost decomposition for simple cycles, and a quadratic-time, $4/3$-approximation algorithm for permutations that contain multiple cycles. The proposed methods rely on investigating a newly introduced balancing property of cycles embedded in trees, cycle-merging methods, and shortest path optimization techniques.

cs.DS↗

arXiv2014

Synchronizing Rankings via Interactive Communication

We consider the problem of exact synchronization of two rankings at remote locations connected by a two-way channel. Such synchronization problems arise when items in the data are distinguishable, as is the case for playlists, tasklists, crowdvotes and recommender systems rankings. Our model accounts for different constraints on the communication throughput of the forward and feedback links, resulting in different anchoring, syndrome and checksum computation strategies. Information editing is assumed of the form of deletions, insertions, block deletions/insertions, translocations and transpositions. The protocols developed under the given model are order-optimal with respect to genie aided lower bounds.

cs.IT↗