arXiv subjects

Scott Kirkpatrick

Publications and source records attributed to Scott Kirkpatrick.

16 recordsLinked to original sources

arXiv2025

50 years of spin glass theory

In 1975, two papers were published that together sparked major new directions, conceptual, mathematical and practically applicable, in several previously disparate fields of science. In this short review, we expose key aspects of their thinking, implementations and implications, along with a selection of further crucial and consequential developments. These papers were `Theory of spin glasses' by S.F.Edwards and P.W.Anderson (EA)[1] and `Solvable Model of a Spin-Glass', by D.Sherrington and S.Kirkpatrick (SK)[2], both concerned with trying to understand recent experiments that suggested a new phase of matter.

cond-mat.dis-nn↗

arXiv2023

Simulated annealing, optimization, searching for ground states

The chapter starts with a historical summary of first attempts to optimize the spin glass Hamiltonian, comparing it to recent results on searching largest cliques in random graphs. Exact algorithms to find ground states in generic spin glass models are then explored in Section 1.2, while Section 1.3 is dedicated to the bidimensional case where polynomial algorithms exist and allow for the study of much larger systems. Finally Section 1.4 presents a summary of results for the assignment problem where the finite size corrections for the ground state can be studied in great detail.

cond-mat.dis-nn↗

arXiv2022

Impulse Measurement Methods for Pulsed Laser Ablation Propulsion

Pulsed laser ablation propulsion has the potential to revolutionize space exploration by eliminating the requirement of a spacecraft to carry its propellant and power source as the high-power laser is situated off-board. More experimentation needs to be done to optimize this propulsion system and understand the mechanisms of thrust generation. There are many methods used to calculate the impulse imparted in pulsed laser ablation experiments. In this paper, key performance parameters are derived for some of the impulse measurement methods used in ablation propulsion experiments. Regimes discussed include the torsional pendulum system, simple pendulum system, and solid and liquid microspheres.

physics.app-ph↗

arXiv2020

Large independent sets on random $d$-regular graphs with fixed degree $d$

This paper presents a linear prioritized local algorithm that computes large independent sets on a random $d$-regular graph with small and fixed degree $d$. We studied experimentally the independence ratio obtained by the algorithm when $ d \in [3,100]$. For all $d \in [5,100]$, our results are larger than lower bounds calculated by exact methods, thus providing improved estimates of lower bounds.

cs.DS↗

arXiv2018

From Megabits to CPU~Ticks: Enriching a Demand Trace in the Age of MEC

All the content consumed by mobile users, be it a web page or a live stream, undergoes some processing along the way; as an example, web pages and videos are transcoded to fit each device's screen. The recent multi-access edge computing (MEC) paradigm envisions performing such processing within the cellular network, as opposed to resorting to a cloud server on the Internet. Designing a MEC network, i.e., placing and dimensioning the computational facilities therein, requires information on how much computational power is required to produce the contents needed by the users. However, real-world demand traces only contain information on how much data is downloaded. In this paper, we demonstrate how to {\em enrich} demand traces with information about the computational power needed to process the different types of content, and we show the substantial benefit that can be obtained from using such enriched traces for the design of MEC-based networks.

cs.NI↗

arXiv2018

Revisiting the Challenges of MaxClique

The MaxClique problem, finding the largest complete subgraph in an Erd{\"o}s-R{\'e}nyi $G(N,p)$ random graph in the large $N$ limit, is a well-known example of a simple problem for which finding any approximate solution within a factor of $2$ of the known, probabilistically determined limit, appears to require P$=$NP. This type of search has practical importance in very large graphs. Algorithmic approaches run into phase boundaries long before they reach the size of the largest likely solutions. And, most intriguing, there is an extensive literature of \textit{challenges} posed for concrete methods of finding maximum naturally occurring as well as artificially hidden cliques, with computational costs that are at most polynomial in the size of the problem. We use the probabilistic approach in a novel way to provide a more insightful test of constructive algorithms for this problem. We show that extensions of existing methods of greedy local search will be able to meet the \textit{challenges} for practical problems of size $N$ as large as $10^{10}$ and perhaps more. Experiments with spectral methods that treat a single large clique of size $\alpha N^{1/2}$ \textit{planted} in the graph as an impurity level in a tight binding energy band show that such a clique can be detected when $\alpha \geq \approx1.0$. Belief propagation using a recent \textit{approximate message passing} (\textbf{AMP}) scheme of inference pushes this limit down to $\alpha \sim \sqrt{1/e}$. Exhaustive local search (with early stopping when the planted clique is found) does even better on problems of practical size, and proves to be the fastest solution method for this problem.

cs.DS↗

arXiv2018

Mining the Air -- for Research in Social Science and Networking Measurement

Smartphone apps provide a vitally important opportunity for monitoring human mobility, human experience of ubiquitous information aids, and human activity in our increasingly well-instrumented spaces. As wireless data capabilities move steadily up in performance, from 2&3G to 4G (today's LTE) and 5G, it has become more important to measure human activity in this connected world from the phones themselves. The newer protocols serve larger areas than ever before and a wider range of data, not just voice calls, so only the phone can accurately measure its location. Access to the application activity permits not only monitoring the performance and spatial coverage with which the users are served, but as a crowd-sourced, unbiased background source of input on all these subjects, becomes a uniquely valuable resource for input to social science and government as well as telecom providers

cs.CY↗

arXiv2018

Cellular Network Traces Towards 5G: Usage, Analysis and Generation

Deployment and demand traces are a crucial tool to study today's LTE systems, as well as their evolution toward 5G. In this paper, we use a set of real-world, crowdsourced traces, coming from the WeFi and OpenSignal apps, to investigate how present-day networks are deployed, and the load they serve. Given this information, we present a way to generate synthetic deployment and demand profiles, retaining the same features of their real-world counterparts. We further discuss a methodology using traces (both real-world and synthetic) to assess (i) to which extent the current deployment is adequate to the current and future demand, and (ii) the effectiveness of the existing strategies to improve network capacity. Applying our methodology to real-world traces, we find that present-day LTE deployments consist of multiple, entangled, medium- to large-sized cells. Furthermore, although today's LTE networks are overprovisioned when compared to the present traffic demand, they will need substantial capacity improvements in order to face the load increase forecasted between now and 2020.

cs.NI↗

arXiv2016

How Close to the Edge? Delay/utilization tradeoffs in MEC

Virtually all of the rapidly increasing data traffic consumed by mobile users requires some kind of processing, normally performed at cloud servers. A recent thrust, {\em mobile edge computing}, moves such processing to servers {\em within} the cellular mobile network. The large temporal and spatial variations to which mobile data usage is subject could make the reduced latency that edge clouds offer come at an unacceptable cost in redundant and underutilized infrastructure. We present some first empirical results on this question, based on large scale sampled crowd-sourced traces from several major cities spanning multiple operators and identifying the applications in use. We find opportunities to obtain both high server utilization and low application latency, but the best approaches will depend on the individual network operator's deployment strategy and geographic specifics of the cities we study.

cs.NI↗

arXiv2016

The Impact of Vehicular Traffic Demand on 5G Caching Architectures: a Data-Driven Study

The emergence of in-vehicle entertainment systems and self-driving vehicles, and the latter's need for high-resolution, up-to-date maps, will bring a further increase in the amount of data vehicles consume. Considering how difficult WiFi offloading in vehicular environments is, the bulk of this additional load will be served by cellular networks. Cellular networks, in turn, will resort to caching at the network edge in order to reduce the strain on their core network, an approach also known as mobile edge computing, or fog computing. In this work, we exploit a real-world, large-scale trace coming from the users of the We-Fi app in order to (i) understand how significant the contribution of vehicular users is to the global traffic demand; (ii) compare the performance of different caching architectures; and (iii) studying how such a performance is influenced by recommendation systems and content locality. We express the price of fog computing through a metric called price-of-fog, accounting for the extra caches to deploy compared to a traditional, centralized approach. We find that fog computing allows a very significant reduction of the load on the core network, and the price thereof is low in all cases and becomes negligible if content demand is location specific. We can therefore conclude that vehicular networks make an excellent case for the transition to mobile-edge caching: thanks to the peculiar features of vehicular demand, we can obtain all the benefits of fog computing, including a reduction of the load on the core network, reducing the disadvantages to a minimum.

cs.NI↗

arXiv2016

What is LTE actually used for? An answer through multi-operator, crowd-sourced measurement

LTE networks are commonplace nowadays; however, comparatively little is known about where (and why) they are deployed, and the demand they serve. We shed some light on these issues through large-scale, crowd-sourced measurement. Our data, collected by users of the WeFi app, spans multiple operators and multiple cities, allowing us to observe a wide variety of deployment patterns. Surprisingly, we find that LTE is frequently used to improve the {\em coverage} of network rather than the capacity thereof, and that no evidence shows that video traffic be a primary driver for its deployment. Our insights suggest that such factors as pre-existing networks and commercial policies have a deeper impact on deployment decisions than purely technical considerations.

cs.NI↗

arXiv2016

The Price of Fog: a Data-Driven Study on Caching Architectures in Vehicular Networks

Vehicular users are expected to consume large amounts of data, for both entertainment and navigation purposes. This will put a strain on cellular networks, which will be able to cope with such a load only if proper caching is in place, this in turn begs the question of which caching architecture is the best-suited to deal with vehicular content consumption. In this paper, we leverage a large-scale, crowd-collected trace to (i) characterize the vehicular traffic demand, in terms of overall magnitude and content breakup, (ii) assess how different caching approaches perform against such a real-world load, (iii) study the effect of recommendation systems and local contents. We define a price-of-fog metric, expressing the additional caching capacity to deploy when moving from traditional, centralized caching architectures to a "fog computing" approach, where caches are closer to the network edge. We find that for location-specific contents, such as the ones that vehicular users are most likely to request, such a price almost disappears. Vehicular networks thus make a strong case for the adoption of mobile-edge caching, as we are able to reap the benefit thereof -- including a reduction in the distance traveled by data, within the core network -- with little or no of the associated disadvantages.

cs.NI↗

arXiv2010

Social Networks and Spin Glasses

The networks formed from the links between telephones observed in a month's call detail records (CDRs) in the UK are analyzed, looking for the characteristics thought to identify a communications network or a social network. Some novel methods are employed. We find similarities to both types of network. We conclude that, just as analogies to spin glasses have proved fruitful for optimization of large scale practical problems, there will be opportunities to exploit a statistical mechanics of the formation and dynamics of social networks in today's electronically connected world.

cs.SI↗

arXiv2006

MEDUSA - New Model of Internet Topology Using k-shell Decomposition

The k-shell decomposition of a random graph provides a different and more insightful separation of the roles of the different nodes in such a graph than does the usual analysis in terms of node degrees. We develop this approach in order to analyze the Internet's structure at a coarse level, that of the "Autonomous Systems" or ASes, the subnetworks out of which the Internet is assembled. We employ new data from DIMES (see http://www.netdimes.org), a distributed agent-based mapping effort which at present has attracted over 3800 volunteers running more than 7300 DIMES clients in over 85 countries. We combine this data with the AS graph information available from the RouteViews project at Univ. Oregon, and have obtained an Internet map with far more detail than any previous effort. The data suggests a new picture of the AS-graph structure, which distinguishes a relatively large, redundantly connected core of nearly 100 ASes and two components that flow data in and out from this core. One component is fractally interconnected through peer links; the second makes direct connections to the core only. The model which results has superficial similarities with and important differences from the "Jellyfish" structure proposed by Tauro et al., so we call it a "Medusa." We plan to use this picture as a framework for measuring and extrapolating changes in the Internet's physical structure. Our k-shell analysis may also be relevant for estimating the function of nodes in the "scale-free" graphs extracted from other naturally-occurring processes.

cond-mat.dis-nn↗

arXiv2005

Selfish vs. Unselfish Optimization of Network Creation

We investigate several variants of a network creation model: a group of agents builds up a network between them while trying to keep the costs of this network small. The cost function consists of two addends, namely (i) a constant amount for each edge an agent buys and (ii) the minimum number of hops it takes sending messages to other agents. Despite the simplicity of this model, various complex network structures emerge depending on the weight between the two addends of the cost function and on the selfish or unselfish behaviour of the agents.

cs.NI↗

arXiv2005

Comparing Beliefs, Surveys and Random Walks

Survey propagation is a powerful technique from statistical physics that has been applied to solve the 3-SAT problem both in principle and in practice. We give, using only probability arguments, a common derivation of survey propagation, belief propagation and several interesting hybrid methods. We then present numerical experiments which use WSAT (a widely used random-walk based SAT solver) to quantify the complexity of the 3-SAT formulae as a function of their parameters, both as randomly generated and after simplification, guided by survey propagation. Some properties of WSAT which have not previously been reported make it an ideal tool for this purpose -- its mean cost is proportional to the number of variables in the formula (at a fixed ratio of clauses to variables) in the easy-SAT regime and slightly beyond, and its behavior in the hard-SAT regime appears to reflect the underlying structure of the solution space that has been predicted by replica symmetry-breaking arguments. An analysis of the tradeoffs between the various methods of search for satisfying assignments shows WSAT to be far more powerful that has been appreciated, and suggests some interesting new directions for practical algorithm development.

cond-mat.stat-mech↗