Print

Print


*****  To join INSNA, visit http://www.insna.org  *****

Just to add two things. The mathematical background of
non-deterministic models of diffusion on graphs is often based on the
theory of Interacting Particle Systems (cf., Liggett, Durrett etc.)
and one of the first simple models of this type was what was called
(by Kempe et al.) Independent Cascade Model, developed in the context
of marketing by Goldenberg, Libai & Muller. In what concerns
deterministic models, besides the sociological motivations of
Granovetter and Schelling, formal models of that sort are also known
as "bootstrap percolation," originally introduced in the context of
disordered magnetic systems and subsequently studied in combinatorics
(e.g., the work of Daniel Reichman, which is very similar to the paper
in question as this is also acknowledged by the authors of the paper
too) and in random graphs (e.g., the work of Svante Janson).

--Moses

On Sat, Sep 21, 2013 at 10:06 AM, Alex Stivala <[log in to unmask]> wrote:
> ***** To join INSNA, visit http://www.insna.org *****
> I've read the paper in question (and the news article about it). As some
> sort of computer scientist myself their algorithm and their analysis of it
> seems sound to me, and their experimental results look pretty good.
>
> I don't know what the "fundamental problem of viral marketing" is (if there
> is one). And of course there are other models than the (notably,
> deterministic - all the thresholds are specified) tipping model in this
> paper. There was a talk (Wu et al "Maximizing the Performance of
> Advertisements Diffusion: A Simulation Study of the Dynamics of Viral
> Advertising in Social Networks") at INSNA in Xi'an on this topic for
> instance, using a non-deterministic model (following from the work of Kempe
> et al., cited in the "notes on non-deterministic tipping" section of the
> paper).
>
> If the question is about how realistic such models are, or how applicable,
> it seems to me the non-deterministic models and the somewhat different
> problem the INSNA talk and the other work in the area following Kempe et al.
> address seems more realistic. In such models the thresholds are random
> variables (reflecting our lack of knowledge of them), and rather than
> finding a set of nodes that will infect the entire network, the problem is
> to find a a set of nodes of a specified size that maximizes the expected
> number of infected nodes. Seems to me a marketer would care more about
> maximizing this number of infected nodes given a fixed budget (the number of
> initial nodes to give a free sample to or whatever), than really wanting to
> get their message to absolutely everybody in the network.
>
> We should bear in mind that the news article and its headline was not
> written by the authors of the paper  - their paper completely lives up to
> their more modest title "A Scalable Heuristic for Viral Marketing Under the
> Tipping Model". The news article slightly misinterprets their algorithm, but
> does correctly describe its essence as Bryan describes it below.
>
> Obviously a news headline "Military Scientists Solve the Fundamental Problem
> of Viral Marketing" is somewhat more eye-catching than the more accurate
> "Military Scientists Find Efficient Heuristic to Solve Problem of Finding
> Non-Trivial But Not Necessarily Minimal Seed Set Size To Infect Entire
> Network Under the Deterministic Tipping Model in Viral Marketing".
>
> Regards,
> Alex.
>
> (Postdoc, Melbourne School of Psychological Sciences, The University of
> Melbourne, Australia.)
>
>
>
> ________________________________
> From: "Lee, Bryan Leslie" <[log in to unmask]>
> To: [log in to unmask]
> Sent: Friday, 20 September 2013 3:30 AM
> Subject: Re: [SOCNET] Fwd: viral marketing
>
> *****  To join INSNA, visit http://www.insna.org *****
>
> I'm curious as well.  I'm not a computer scientist and can't comment on the
> algorithm, but the end result doesn't strike me as living up to the
> headline.  As I read it, it presents a method of finding a seed group which
> is large enough to infect the entire network (assuming a certain threshold
> infection level) which works on the massive scale of actual online social
> networks.  This is exciting, but I don't think it solves the "fundamental
> problem," which I still see as proving the threshold infection model
> accurately depicts actual human behavior.
>
> I'm eager to hear what the experts on this list have to say.  I'm a neophyte
> and came here to learn.
>
> -Bryan
>
> Bryan L. Lee
> Director, Eurasia Nonproliferation Program
> James Martin Center for Nonproliferation Studies
> Monterey Institute of International Studies
> 460 Pierce Street
> Monterey, CA 93940
> T: (831) 647-3509 | F: (831) 647-3519
> nonproliferation.org
>
>
> -----Original Message-----
> From: Social Networks Discussion Forum [mailto:[log in to unmask]] On
> Behalf Of Simone Gabbriellini
> Sent: Thursday, September 19, 2013 1:46 AM
> To: [log in to unmask]
> Subject: [SOCNET] Fwd: viral marketing
>
> *****  To join INSNA, visit http://www.insna.org *****
>
> Hello everybody,
>
> I have just stumbled upon this paper: "A Scalable Heuristic for Viral
> Marketing Under the Tipping Model" by Paulo Shakarian, Sean Eyre and Damon
> Paulo.
>
> http://arxiv.org/abs/1309.2963
>
> I am just curious to see if someone has any comment on how to interpret the
> (journalistic) claim "Military Scientists Solve the Fundamental Problem of
> Viral Marketing"?
>
> http://www.technologyreview.com/view/519361/us-military-scientists-solve-the-fundamental-problem-of-viral-marketing/
>
> All the best,
> Simone
>
> --
> Simone Gabbriellini, PhD
>
> PostDoc@DISI, University of Bologna
> mobile: +39 340 39 75 626
> email: [log in to unmask]
> academia.edu: http://goo.gl/7pq62
>
> DigitalBrains srl
> Amministratore
> mobile: +39 340 39 75 626
> email: [log in to unmask]
> home: www.digitalbrains.it.
>
> _____________________________________________________________________
> SOCNET is a service of INSNA, the professional association for social
> network researchers (http://www.insna.org). To unsubscribe, send an email
> message to [log in to unmask] containing the line UNSUBSCRIBE SOCNET in
> the body of the message.
>
> _____________________________________________________________________
> SOCNET is a service of INSNA, the professional association for social
> network researchers (http://www.insna.org). To unsubscribe, send
> an email message to [log in to unmask] containing the line
> UNSUBSCRIBE SOCNET in the body of the message.
>
>
> _____________________________________________________________________ SOCNET
> is a service of INSNA, the professional association for social network
> researchers (http://www.insna.org). To unsubscribe, send an email message to
> [log in to unmask] containing the line UNSUBSCRIBE SOCNET in the body of
> the message.

_____________________________________________________________________
SOCNET is a service of INSNA, the professional association for social
network researchers (http://www.insna.org). To unsubscribe, send
an email message to [log in to unmask] containing the line
UNSUBSCRIBE SOCNET in the body of the message.