Search engine run on: http://users.monash.edu.au/
Glookbib search for: zz0418
%A A. M. Wood
%A et al ...
%A J. Danesh
%T Risk thresholds for alcohol consumption: combined analysis of
individual-participant data for 599,912 current drinkers in 83 prospective
studies
%J The Lancet
%V 391
%N 10129
%P 1513-1523
%M APR
%D 2018
%K jrnl, c2018, c201x, c20xx, zz0418, moderate alcohol consumption,
drink, drinking, human health, risk, risks, zzAlcohol, negative
%X "... studied individual-participant data from 599,912 current drinkers
without previous cardiovascular disease. ... Alcohol consumption was roughly
linearly associated with a higher risk of stroke (HR per 100 g per week
higher consumption 1.14, 95% CI, 1.10-1.17), coronary disease excluding
myocardial infarction (1.06, 1.00-1.11), heart failure (1.09, 1.03-1.15),
fatal hypertensive disease (1.24, 1.15-1.33); and fatal aortic aneurysm
(1.15, 1.03-1.28). By contrast, increased alcohol consumption was
log-linearly associated with a lower risk of myocardial infarction (HR 0.94,
0.91-0.97). In comparison to those who reported drinking > 0- <= 100 g per
week, those who reported drinking > 100- <= 200 g per week, > 200- <= 350
g per week, or > 350 g per week had lower life expectancy at age 40 years of
approximately 6 months, 1-2 years, or 4-5 years, respectively. ...
the threshold for lowest risk of all-cause mortality was about 100 g/week.
For cardiovascular disease subtypes other than myocardial infarction, there
were no clear risk thresholds below which lower alcohol consumption stopped
being associated with lower disease risk. ..."
-- [doi:10.1016/S0140-6736(18)30134-X]['18].
[Also search for: zzAlcohol].
%A T. Kavitha
%T The popular roommates problem
%J arXiv
%M MAR
%D 2018
%K TR, c2018, c201x, c20xx, zz0418, popular roommates problem, room mate,
roommate, matching, dominant, matchings, matching, preference, NPH
%X "... in a roommates instance with strict preference lists. While popular
matchings always exist in a bipartite instance, they need not exist in a
roommates instance. The complexity of the popular matching problem in a
roommates instance has been an open problem for several years and here we
show it is NP-hard. A sub-class of max-size popular matchings called dominant
matchings has been well-studied in bipartite graphs. We show that the
dominant matching problem in a roommates instance is also NP-hard and this is
the case even when the instance admits a stable matching."
-- 1804.00141@[arXiv]['18].
[Also search for: stable marriage]. Note, 'NPH'.
(Also see Gupta roommates c2018 !)
%A S. Gupta
%A P. Misra
%A S. Saurabh
%A M. Zehavi
%T Popular matching in roommates setting is NP-hard
%J arXiv
%M MAR
%D 2018
%K TR, c201x, c20xx, zz0418, popular roommates problem, room mate,
roommate, matching, NPC
%X "An input to the Popular Matching problem, in the roommates setting, consists
of a graph G and each vertex ranks its neighbors in strict order, known as
its preference. In the Popular Matching problem the objective is to test
whether there exists a matching M⋆ such that there is no matching M where
more people are happier with M than with M⋆. In this paper we settle the
computational complexity of the Popular Matching problem in the roommates
setting by showing that the problem is NP-complete. Thus, we resolve an open
question that has been repeatedly, explicitly asked over the last decade."
-- [arXiv]['18]. Note 'NPC'.
(Also see Kavitha roommates c2018 !)
%A A. Polak
%T Why is it hard to beat O(n^2) for longest common weakly increasing
subsequence?
%J IPL
%V 132
%P 1-5
%M APR
%D 2018
%K jrnl, IPL, c2018, c201x, c20xx, zz0418, LCWIS, LCIS, LCS, problem,
Longest Common Weakly Increasing Subsequence, algorithm, SETH
%X "... we show that LCWIS cannot be solved in O(n^(2-eps)) time unless the
Strong Exponential Time Hypothesis (SETH) is false. The ideas which we
developed can also be used to obtain a lower bound based on a safer
assumption of SETH, i.e. a version of SETH which talks about circuits instead
of less expressive CNF formulas."
-- [doi:10.1016/j.ipl.2017.11.007]['18].
%A M. Kowaluk
%A A. Lingas
%T Are unique subgraphs not easier to find?
%J IPL
%V 134
%P 57-61
%M JUN
%D 2018
%K jrnl, IPL, c2018, c201x, c20xx, zz0418, maths, graph, subgraph, instance,
motif, unique
%X "Consider a pattern graph H with l edges, and a host graph G which may
contain several occurrences of H. ... We show a counterexample to this
too strong claim and correct it by providing an O~((l(d-1)+2)^l) bound on the
multiplicative factor instead, where d is the max. # of occurrences of H that
can share the same edge in the input host graph. We provide also an analogous
correction in the induced case when occurrences of induced subgs. isomorphic
to H are sought."
-- [doi:10.1016/j.ipl.2018.02.010]['18].
%A D. Sumanaweera
%A L. Allison
%A A. S. Konagurthu
%T The bits between proteins
%J Data Compression Conference (DCC)
%I IEEE
%W Snowbird, Utah, USA
%M MAR
%D 2018
%K conf, DCC, DCC2018, MolBio, c2018, c201x, c20xx, zz0418, bioinformatics,
LAllison, ArunK, Dinithi, DCC, DCC2018, protein, sequence, alignment, DPA,
algorithm, evolutionary, model, minimum message length, MDL, MML, AIC,
homology, information, similarity
%X "Comparison of protein sequences via alignment is an important routine in
modern biological studies. Although the technologies for aligning proteins
are mature, the current state of the art continues to be plagued by many
shortcomings, chiefly due to the reliance on: (i) naive objective functions,
(ii) fixed substitution scores independent of the sequences being considered,
(iii) arbitrary choices for gap costs, and (iv) reporting, often, one
optimal alignment without a way to recognise other competing sequence
alignments. Here, we address these shortcomings by applying the
compression-based Minimum Message Length (MML) inference framework to the
protein sequence alignment problem. This grounds the problem in statistical
learning theory, handles directly the complexity-vs-fit trade-off without
ad hoc gap costs, allows unsupervised inference of all the statistical
parameters, and permits the visualization and exploration of competing
sequence alignment landscape."
-- [more],
[doi:10.1109/DCC.2018.00026]['18].
(Also see [protein].)
%A S. Skansi
%T Introduction to Deep Learning: From Logical Calculus to
Artificial Intelligence
%I SpringerVerlag
%S Undergrad. Topics in Comp. Sci.
%P 191
%M FEB
%D 2018
%K intro, book, text, c2018, c201x, c20xx, zz0418, AI, deep learning, ANN, NN,
convolutional, LSTM, Word2vec, RBM, DBN
%X 1st ed 2018; pb us$44; uk us isbn:3319730037; uk us isbn13:978-3319730035.
"... coverage includes convolutional networks, LSTMs, Word2vec, RBMs,
DBNs, neural Turing machines, memory networks and autoencoders. ..."
%A J. Preskill
%T Stephen Hawking (1942-2018)
%J Science
%V 360
%N 6385
%P 156-156
%M APR
%D 2018
%K Hawking, obit, obituary, c2018, c201x, c20xx, zz0418, physics, cosmology,
black hole, holes
%X "Stephen William Hawking died on 14 March (Albert Einstein's birthday) at
the age of 76 ..."
-- [doi:10.1126/science.aat6775]['18].
[Also search for: Stephen Hawking biog].
%A L. Cheng
%A et al ...
%A J. Corander
%T Hierarchical and spatially explicit clustering of DNA sequences with BAPS
software
%J Mol. Biol. Evol.
%V 30
%N 5
%P 1224-1228
%M MAY
%D 2013
%K jrnl, MBE, MolBio, c2013, c201x, c20xx, zz0418, BAPS, DNA, sequence, nested,
genetic population structure, phylogeographics, evolutionary epidemiology,
Borrelia burgdorferi, spatialepidemiology, epidemiology, rhierbaps
%X "... introduce two upgrades to the Bayesian Analysis of Population Structure
(BAPS) s/w, which enable 1) spatially explicit modeling of variation in DNA
sequences & 2) hierarchical clustering of DNA seq. data to reveal nested
genetic popn structures. We provide a direct interface to map the results
from spatial clustering with Google Maps using the portal
www.spatialepidemiology.net & illustrate this approach using seq. data from
Borrelia burgdorferi. The usefulness of hierarchical clustering is demo.
through an analysis of the metapopulation structure within a bacterial
popn experiencing a high level of local horizontal gene transfer. The tools
that are introduced are freely available at [www]."
-- [doi:10.1093/molbev/mst028]['18].
(Also see rhierbaps@[github]['18].)
%A J. M. Dybas
%A A. Fiser
%T Development of a motif-based topology-independent structure comparison method
to identify evolutionarily related folds
%J Proteins: Structure, Function, and Bioinformatics
%I Wiley
%V 84
%N 12
%P ?-?
%M SEP
%D 2016
%K jrnl, MolBio, c2016, c201x, c20xx, zz0418, SmotifCOMP, protein, structure,
fold, super-secondary-structure, motif, library
%X "... describe a novel, super-secondary-structure motif-based,
topology-independent structure comparison method (SmotifCOMP) that is able to
quantitatively identify structural relationships between disparate
topologies. The basis of SmotifCOMP is a systematically defined SSS motif
library whose representative geometries are shown to be saturated in the
PDB & exhibit a unique distn within the known folds. SmotifCOMP offers a
robust & quantitative technique to compare domains that adopt different
topologies since [it] does not rely on a global superposition. SmotifCOMP is
used to perform an exhaustive comparison of the known folds & the identified
relationships are used to produce a nonhierarchical repn of the fold space
that reflects the notion of a cts & connected fold universe. ..."
-- [doi:10.1002/prot.25169]['18],
& PMC5118133@[ncbi]['18].
[Also search for: MolBio protein structure motif].
%A V. M. Narasimhan
%A et al ...
%A D. Reich
%T The genomic formation of South and Central Asia
%J bioRxiv
%M MAR
%D 2018
%K TR, c2018, c201x, c20xx, zz0418, MolBio, human evolution, South, Asia,
Europe, Eurasia, Steppe, Indo-European, India, Indus, Iran, genetics,
BMAC, Bactria Margiana, ancient DNA, spread, language
%X "... generated genome-wide data from 362 ancient individuals, inc. the first
from eastern Iran, Turan (Uzbekistan, Turkmenistan, & Tajikistan), Bronze Age
Kazakhstan, & South Asia. ... document a southward spread of genetic ancestry
from the Eurasian Steppe, correlating with the archaeologically known
expansion of pastoralist sites from the Steppe to Turan in the Middle Bronze
Age (2300-1500 BCE). These Steppe communities mixed genetically with peoples
of the Bactria Margiana Archaeological Complex (BMAC) whom they encountered
in Turan ... but there is no evidence that the main BMAC population
contributed genetically to later S.Asians. Instead, Steppe communities
integrated farther south throughout the 2nd millennium BCE, & we show that
they mixed with a more S.population ... By co-analyzing ancient DNA & genomic
data from diverse present-day S.Asians, we show that Indus Periphery-related
people are the single most important source of ancestry in S.Asia ...
Our results show how ancestry from the Steppe genetically linked Europe &
S.Asia in the Bronze Age, & identifies the popns that almost certainly were
responsible for spreading Indo-European languages across much of Eurasia."
-- [doi:10.1101/292581]['18].
(Also see ReichLab@[harv'rd]['18].)
%A M. Valenzuela
%A Vu Ha
%A O. Etzioni
%T Identifying meaningful citations
%J Workshop on Scholarly Big Data: AI Perspectives, Challenges, and Ideas
%D 2015
%K wkShop, c2015, c201x, c20xx, zz0418, modified bibliometrics, papers, impact,
scientific research, cite, important, significant, citations, references,
semanticscholar, knowledge, science
%X "We introduce the novel task of identifying important citations in scholarly
literature, i.e., citations that indicate that the cited work is used(!) or
extended in the new effort. We believe this task is a crucial component in
algorithms that detect & follow research topics & in methods that measure the
quality of publications. We model this task as a supervised classification
problem at two levels of detail: a coarse one with classes (important vs.
non-important), & a more detailed one with four importance classes. We
annotate a dataset of approximately 450 citations with this information, &
release it publicly. We propose a supervised classification approach that
addresses this task with a battery of features that range from citation
counts to where the citation appears in the body of the paper, & show that,
our approach achieves a precision of 65% for a recall of 90%."
-- [am'zonaws]['22],
[AAAI]['18],
@ [SScclr]['22].
(Also see semanticscholar@[www]['18].)
[Also search for: modified bibliometrics].
%A M. Harvey
%A R. Mailhammer
%T Reconstructing remote relationships
%J Diachronica
%V 34
%N 4
%P 470-515
%D 2017
%K jrnl, c2017, c201x, c20xx, zz0418, linguistics, libguistic, proto-Australian,
natural, language, evolution, noun, nouns, indigenous, aboriginal, koori,
Australia, Australian
%X "... Hypotheses on remote relationships commonly involve greater geographical
& temporal ranges. Consequently, we propose that there are two factors which
are likely to play a greater role in comparing hypotheses of chance, contact
& inheritance for remote relationships: (i) spatial distribution of
corresponding forms; & (ii) language specific unpredictability in related
paradigms. Concentrated spatial distributions disfavour hypotheses of chance,
& discontinuous distributions disfavour contact hypotheses, whereas
hypotheses of inheritance may accommodate both. Higher levels of
language-specific unpredictability favour remote over recent transmission.
We consider a remote relationship hypothesis, the Proto-Australian
hypothesis. We take noun class prefixation as a test dataset for evaluating
this hypothesis against these two criteria, & we show that inheritance is
favoured over chance & contact."
-- [doi:10.1075/dia.15032.har]['18].
auth1@{schlr]['18],
auth2@[schlr]['18].
(Also see [abc][28/3/2018].)
[Also search for: natural language evolution].
%A A. McCook
%T University is quick to disclose misconduct
%J Science
%V 360
%N 6384
%P 13-14
%M APR
%D 2018
%K news, views, c2018, c201x, c20xx, zz0418, academic, university, scientific,
research, Ohio State University (OSU), cancer, Ching-Shih Chen, misconduct,
ethics, Retraction Watch, NIH, grant, grants
%X "... (OSU) in Columbus last week released a detailed account of the
scientific misbehavior of one of its former faculty members. The 75-page
report was damning: It concluded that cancer researcher Ching-Shih Chen -
once lauded as an 'Innovator of the Year' and the winner of millions of
dollars in federal funding - had committed misconduct in eight papers.
Chen resigned last September. ..."
-- [doi:10.1126/science.360.6384.13]['18].
[Also search for: academic misconduct].
%A J. Brainard
%T NIH looks to punish reviewers who violate confidentiality
%J Science
%V 360
%N 6384
%P 17
%M APR
%D 2018
%K news, views, c2018, c201x, c20xx, zz0418, academic, scientific, research,
grant, grants, NIH, ARC, peer review, reviewing, ethics, misconduct, scandal,
cheating, corruption, scam, fraud, reciprocal favour, favor, USA
%X "The U.S. National Institutes of Health (NIH) announced last week it is
investigating violations of confidentiality rules designed to protect the
integrity of its peer reviews that score funding applications. A violation
of NIH rules involving one peer-review panel led the agency recently to
re-review 60 applications that the committee had evaluated. NIH officials
were mum about the number of cases and most other details but said at least
one violation involved 'reciprocal favors' - a term that typically means an
applicant offered a favor to a reviewer in exchange for a favorable
evaluation of a proposal, or vice versa. NIH rules suggest possible sanctions
could include suspending or barring violators from obtaining federal
research funds."
-- [doi:10.1126/science.360.6384.17]['18].
[Also search for: peer review].
%A A. Nunes
%A B. Reimer
%A J. F. Coughlin
%T People must retain control of autonomous vehicles
%J Nature
%V 556
%N ?
%P 169-171
%M APR
%D 2018
%K news, views, c2018, c201x, c20xx, zz0418, autonomous, self driving, car,
cars, automobile, law, laws, liability, Tesla, accident, collision, casualty,
safety, driverless, policy, politics, risk, risks
%X "Last month, for the first time, a pedestrian was killed in an accident
involving a self-driving car. ..."
-- [doi:10.1038/d41586-018-04158-5]['18].
[Also search for: autonomous cars].
%A K. Gross
%A C. T. Bergstrom
%T Contest models highlight inefficiencies of scientific funding
%J arXiv
%D 2018
%K TR, c2018, c201x, c20xx, zz0418, scientific, academic, university, research,
grant, grants, writing, proposals, ARC, NIH, waste, KPI, KPIs, model
%X "... At a first approximation, the work invested in writing proposals
provides no value to the funder. Here, we use the economic theory of contests
to analyze the scientific efficiency of the proposal system, & compare it to
recently proposed, partially randomized alternatives such as lotteries. We
find that the effort researchers waste in writing proposals may be comparable
to the total scientific value of the additional funding, especially when only
a small percentage of proposals are funded. Moreover, when professional
pressures motivate investigators to seek funding for reasons that extend
beyond the value of the proposed science, the entire program can act as a
drag on scientific progress when paylines are low. We suggest that lost
efficiency may be recouped at low paylines either by partial lotteries for
funding, or by funding researchers based on past scientific success instead
of proposals for future work."
-- 1804.03732@[arXiv]['18].
%A S. Muggleton
%T Learning from positive data
%J Proc. 6th Int. Workshop on Inductive Logic Programming
%I SpringerVerlag
%S LNAI
%V 1314
%D 1997
%K wkShop, c1997, c199x, c19xx, zz0418, Inductive Logic Programming, ILP,
Progol, Inverse Entailment, compression, information, AI, II
%X progol@[ic]['18].
(Also see ILP@[wikip]['18].)
%A N. Gillis
%T Multiplicative updates for polynomial root finding
%J IPL
%V 132
%P 14-18
%M APR
%D 2018
%K jrnl, IPL, c2018, c201x, c20xx, zz0418, polynomial, root, roots, find,
finding, solve, solving, nonnegative, factorization, factorisation,
multiplicative updates, algorithm
%X "Let f(x)=p(x)-q(x) be a polynomial with real coeffs. whose roots have
nonnegative real part, where p & q are polynomials with nonnegative coeffs..
... prove the following: Given an initial point x0, the multiplicative update
x_(t+1) = x_t p(x_t)/q(x_t) (t=0,1,...) monotonically and linearly converges
to the largest (resp. smallest) real roots of f smaller (resp. larger) than
x0 if p(x0) < q(x0) (resp. q(x0) < p(x0)). The motivation to study this alg.
comes from the multiplicative updates proposed ... to solve optimization
problems with nonnegativity constraints; in particular many variants of
nonnegative matrix factorization."
-- [doi:10.1016/j.ipl.2017.11.008]['18].
%A N. Cesana-Arlotti
%A et al ...
%A L. L. Bonatti
%T Precursors of logical reasoning in preverbal human infants
%J Science
%V 359
%N 6381
%P 1263-1266
%M MAR
%D 2018
%K jrnl, c2018, c201x, c20xx, zz0418, human development, baby, babbies, infant,
logic, reasoning, brain, mind
%X "Infants are able to entertain hypotheses about complex events & to modify
them rationally when faced with inconsistent evidence. These capacities
suggest that infants can use elementary logical representations to frame &
prune hypotheses. By presenting scenes containing ambiguities about the
identity of an object, here we show that 12- & 19-month-old infants look
longer at outcomes that are inconsistent with a logical inference necessary
to resolve such ambiguities. At the moment of a potential deduction,
infants' pupils dilated, & their eyes moved toward the ambiguous object when
inferences could be computed, in contrast to transparent scenes not requiring
inferences to identify the object. These oculomotor markers resembled those
of adults inspecting similar scenes, suggesting that intuitive & stable
logical structures involved in the interpretation of dynamic scenes may be
part of the fabric of the human mind."
-- [doi:10.1126/science.aao3539]['18].
[Also search for: baby logic].
%A M. S. Gazzaniga
%T The Consciousness Instinct: Unraveling the Mystery of How the Brain Makes
the Mind
%I FarrarStrausGiroux
%P 288
%M APR
%D 2018
%K book, c2018, c201x, c20xx, zz0418, brain, mind, consciousness,
meat machine
%X 1st ed 2018; hb us$22; uk us isbn:0374715505; uk us isbn13:978-0374715502.
"... Gazzaniga asserts that this model has it backward - brains make machines
but they cannot be reduced to one. New research suggests the brain is
actually a confederation of independent modules working together. ..."
%A D. F. Schmidt
%A E. Makalic
%A J. L. Hopper
%T Approximating message lengths of hierarchical Bayesian models using
posterior sampling
%J Australasian Joint Conf. in AI
%W Hobart
%P 482-294
%M NOV
%D 2016
%K conf, AJCAI, AI, c2016, c201x, c20xx, zz0418, II, MML, MMLh, dowe, MDL,
EnesM, dfschmidt
%X "... we adapt ideas from the information theoretic minimum message length
principle & propose a powerful yet simple model selection criteria for
general hierarchical Bayesian models called MML-h. Computation of this
criterion requires only that a set of samples from the posterior distribution
be available. The flexibility of this new algorithm is demonstrated by a
novel application to state-of-the-art Bayesian hierarchical regression
estimation. Simulations show that the MML-h criterion is able to adaptively
select between classic ridge regression & sparse horseshoe regression
estimators, & the resulting procedure exhibits excellent robustness to the
underlying structure of the regression coefficients."
-- [doi:10.1007/978-3-319-50127-7_41]['18].
(Conf., print pp.729, uk us isbn:3319501267; uk us isbn13:978-3-319-50126-0.)
%A P. Ball
%T Beyond Weird: Why Everything you Thought you knew about Quantum Physics is
Different
%I ChicagoUP
%D 2018
%K book, physics, c2018, c201x, c20xx, zz0418, quantum, theory, weirdness,
intro, information
%X 1st ed 2018; hb us$26; uk us isbn: 022655838X; uk us isbn13:978-0226558387.
(Also see a review [doi:10.1038/d41586-018-03272-8]['18].)
%A A. Becker
%T What Is Real?: The Unfinished Quest for the Meaning of Quantum Physics
%I BasicBooks
%% MAR
%D 2018
%K book, text, c2018, c201x, c20xx, zz0418, physics, quantum mechanics,
reality, Copenhagen
%X 1st ed 2018; us$25; uk us isbn:0465096050; uk us isbn13:978-0465096053.
%A P. Bierhorst
%A et al ...
%A L. K. Shalm
%T Experimentally generated randomness certified by the impossibility of
superluminal signals
%J Nature
%V 556
%P 223-226
%M APR
%D 2018
%K jrnl, physics, c2018, c201x, c20xx, zz0418, random numbers, RNG, quantum,
QRNG, Bell test, number, cryptography
%X "... present certified randomness obtained from a photonic Bell experiment &
extract 1,024 random bits that are uniformly distributed to within 10^(−12).
These random bits could not have been predicted according to any physical
theory that prohibits faster-than-light (superluminal) signalling ..."
-- [doi:10.1038/s41586-018-0019-0]['18].
[Also search for: quantum RNG].
%A C. Rovelli
%T The Order of Time
%I RiverheadBooks
%P 256
%M MAY
%D 2018
%K book, text, c2018, c201x, c20xx, zz0418, physics, arrow of time, past,
future, philosophy
%X 1st ed 2018; hb us$18; uk us isbn:073521610X; uk us isbn13:978-0735216105.
"... Why do we remember the past and not the future? What does it mean for
time to "flow"? Do we exist in time or does time exist in us? In lyric,
accessible prose, [CR] invites us to consider questions about the nature of
time that continue to puzzle physicists and philosophers alike. ..."
Also see CR@[wikip]['22].
[Also search for: arrow time].
%A L. Frey
%A D. Fisher
%A I. Tsamardinos
%A C. F. Aliferis
%A A. Statnikov
%T Identifying Markov blankets with decision tree induction
%I ICDM
%D 2003
%K conf, ICDM, c2003, c200x, c20xx, zz0418, AI, II, local, Markov blanket,
decision tree, DTree, C5.0, C5, Bayesian network
%X "... apply decision tree induction to the task of Markov blanket
identification. Notably, we compare (a) C5.0, a widely used alg. for decision
rule indn, (b) C5C, which post-processes C5.0 's rule set to retain the
most freq. referenced variables & (c) PC, a std method for B.n/wk indn.
C5C performs as well as or better than C5.0 & PC across a # of data sets. Our
modest variation of an inexpensive, accurate, off-the-shelf indn engine
mitigates the need for specialized procedures, & establishes baseline
performance against which specialized algs. can be compared."
-- [doi:10.1109/ICDM.2003.1250903]['18],
or [semanticscholar]['18].
[Also search for: Bayesian network DTree].
%A T. Silander
%A P. Myllymaki
%T A simple approach for finding the globally optimal Bayesian network structure
%J UAI
%P 445-452
%M JUL
%D 2006
%K conf, UAI06, UAI2006, c2006, c200x, c20xx, zz0418, Bayesian network, AI, II,
dynamic programming
%X "We study the problem of learning the best B.n/wk structure wrt a
decomposable score such as BDe, BIC or AIC. This problem is known to be
NP-hard, which means that solving it becomes quickly infeasible as the #
of variables increases. Nevertheless ... show that it is possible to learn
the best B.n/wk struct. with over 30 variables, which covers many practically
interesting cases. Our alg. is less complicated & more efficient than the
techniques presented earlier ... easily parallelized, & offers a possibility
for efficient exploration of the best n/wks consistent with different
variable orderings. ... compare the performance of the alg. to the previous
state-of-the-art alg.. Free source-code & an online-demo can be found at
b-course.hiit.fi/bene [not found 4/'18]."
"... claim feasibility for n < 33 and demonstrate with n=29 ..."
-- [acm]['18].
Also see paper at
1206.6875@[arXiv]['18].
(?Not sure what makes it "DP"?)
(auth1@[schlr]['18].)
[Also search for: SLL algorithm].
%A T. Niinimaki
%A P. Parviainen
%T Local structure discovery in Bayesian networks
%I UAI
%D 2012
%K conf, UAI, UAI2012, c2012, c201x, c20xx, zz0418, AI, II, Bayesian network,
SLL algorithm, HITON, local structure, global, UHelsinki, UniHelsinki
%X "Learning a Bayesian network structure from data is an NP-hard ...
present a score-based local learning alg. called SLL. We conjecture that [it]
is theoretically sound in the sense that it is optimal in the limit of large
sample size. Empirical results suggest that SLL is competitive when compared
to the constraint-based HITON alg.. We also study the prospects of
constructing the n/wk structure for the whole node set based on local results
by presenting two algs. & comparing them to several heuristics."
-- 1210.4888@[arXiv]['18].
(SLL -- Score based Local Learning.)
%A N. Polson
%A V. Rockova
%T Posterior concentration for sparse deep learning
%J arXiv
%P 23
%M MAR
%D 2018
%K TR, c2018, c201x, c20xx, zz0418, AI, II, ANN, NN, deep learning, sparse,
neural network, spike and slab, SSDL, dropout, rectified linear, ReLU,
regularization, posterior, width, depth, sparsity, sparseness, Bayesian
%X "Spike-and-Slab Deep Learning (SS-DL) is a fully Bayesian alternative to
Dropout for improving generalizability of deep ReLU networks. This new type
of regularization enables provable recovery of smooth input-output maps with
unknown levels of smoothness. Indeed, we show that the posterior distribution
concentrates at the near minimax rate for alpha-Holder smooth maps,
performing as well as if we knew the smoothness level alpha ahead of time.
Our result sheds light on architecture design for deep NNs, namely the choice
of depth, width & sparsity level. These n/wk attributes typically depend on
unknown smoothness in order to be optimal. We obviate this constraint with
the fully Bayes construction. As an aside, we show that SS-DL does not
overfit in the sense that the posterior concentrates on smaller networks with
fewer (up to the optimal # of) nodes & links. ... provide new theoretical
justifications for deep ReLU networks from a Bayesian point of view.
-- 1803.09138@[arXiv]['18].
%A S. R. McCurdy
%T Ridge regression and provable deterministic ridge leverage score sampling
%J arXiv
%M MAR
%D 2018
%K TR, c2018, c201x, c20xx, zz0418, stats, linear, ridge regression,
matrix sketch, statistical risk
%X "Ridge leverage scores provide a balance between low-rank approximation and
regularization, and are ubiquitous in randomized linear algebra and machine
learning. ... We provide provable guarantees for deterministic column
sampling using ridge leverage scores. The matrix sketch returned by our
algorithm is a column subset of the original matrix, yielding additional
interpretability. ... Performing ridge regression with the matrix sketch
returned by our algorithm and a particular regularization parameter forces
coefficients to zero and has a provable (1+{epsilon}) bound on the
statistical risk. As such, it is an interesting alternative to elastic net
regularization."
-- 1803.06010@[arXiv]['18].
[Also search for: ridge regression].
%A W. Lutz
%A E. Kebede
%T Education and health: Redrawing the Preston curve
%J Population and Development Review
%A APR
%D 2018
%K jrnl, c2018, c201x, c20xx, zz0418, health, education, schooling, wealth, GDP,
income, life expectancy, longevity, stats
%X "... In all of the models the effect of educational attainment on life
expectancy is highly significant in the expected direction, and the
standardized coefficients are clearly larger than those of income. ..."
-- [doi:10.1111/padr.12141]['18].
%A U. Baier
%T On undetected redundancy in the Burrows-Wheeler transform
%J arXiv
%M APR
%D 2018
%K TR, c2018, c201x, c20xx, zz0418, string, strings, BWT,
Burrows Wheeler transform, compression
%X "... present a new technique to reduce the size of a BWT using its
combinatorial properties, while keeping it invertible. The technique can be
applied to any BWT-based compressor, and, as experiments show, is able to
reduce the encoding size by 8-16% on average and up to 33-57% in the best
cases (depending on the BWT-compressor used), making BWT-based compressors
competitive or even superior to today's best lossless compressors."
-- 1804.01937@[arXiv]['18].
[Also search for: BWT string].
(Also see [BWT].)
%A M. Kosinski
%A D. Stillwell
%A T. Graepel
%T Private traits and attributes are predictable from digital records of human
behavior
%J PNAS
%V ?
%P ?-?
%D 2013
%K jrnl, PNAS, c2013, c201x, c20xx, zz0418, facebook, like, likes, behaviour,
trait, personality, social media, privacy
%X "Likes, can be used to automatically and accurately predict a range of
highly sensitive personal attributes including: sexual orientation,
ethnicity, religious and political views, personality traits, intelligence,
happiness, use of addictive substances, parental separation, age, and gender.
The analysis presented is based on a dataset of over 58,000 volunteers who
provided their Facebook Likes, detailed demographic profiles, and the results
of several psychometric tests. The proposed model uses dimensionality
reduction for preprocessing the Likes data, which are then entered into
logistic/linear regression to predict individual psychodemographic profiles
from Likes. The model correctly discriminates between homosexual and
heterosexual men in 88% of cases, African Americans and Caucasian Americans
in 95% of cases, and between Democrat and Republican in 85% of cases. For
the personality trait 'Openness,' prediction accuracy is close to the
test - retest accuracy of a std personality test. We give examples of
associations between attributes and Likes and discuss implications for
online personalization and privacy." (online Mar 2013)
-- [doi:10.1073/pnas.1218772110]['18].
[Also search for: facebook Kogan].
%A M. H. Yearwood
%A et al ...
%A A. Spectre [Kogan]
%T On wealth and the diversity of friendships: High social class people around
the world have fewer international friends
%J Personality and Individual Differences
%V 87
%P 224-229
%M DEC
%D 2015
%K jrnl, c2015, c201x, c20xx, zz0418, facebook, app, ads, Kogan, internet,
social media, personality, traits, friends, likes, privacy, controversy,
controversial, GSR, SCL, Cambridge Analytica, This Is Your Digital Life app
%X "... , we found that among individuals in the United States, social class
was negatively related to percentage of friends on Facebook that are outside
the United States. In Study 2, we extended these findings to the global level
by analyzing country-level data on Facebook friends formed in 2011 (nearly
50 billion friendships) across 187 countries. We found that people from
higher social class countries (as indexed by GDP per capita) had lower levels
of internationalism - that is, they made more friendships domestically than
abroad." (online Aug 2015.)
-- [doi:10.1016/j.paid.2015.07.040]['18].
(*Note, A.Spectre = A.Kogan, also see
[CamU][11/4/'18] on Facebook & Kogan,
and [bbc][13/4/2018].*)
%A L. Farach-Coulton
%A et al ...
%A M. Mosteiro
%T Closure operators and spam resistance for PageRank
%J arXiv
%M MAR
%D 2018
%K TR, c2018, c201x, c20xx, zz0418, www, HREF, web, internet, spam resistance,
spamming, resistant, pagerank, ranking, rank, pages, graph, network, links,
hub, hubs, Min-k-PRR, MinkPRR
%X "We study the spammablility of ranking fns on the web. Although
graph-theoretic ranking fns, such as Hubs & Authorities & PageRank exist,
there is no graph theoretic notion of how spammable such fns are. We
introduce a very general cost model that only depends on the observation that
changing the links of a page that you own is free, whereas changing the links
on pages owned by others requires effort or money. We define spammability to
be the ratio between the amount of benefit one receives for one's spamming
efforts & the amount of effort/money one must spend to spam. The more
effort/money it takes to get highly ranked, the less spammable the fn. Our
model helps explain why both hubs & authorities & standard PageRank are very
easy to spam. Although standard PageRank is easy to spam, we show that there
exist spam-resistant PageRanks. Specifically, we propose a ranking method,
Min-k-PPR, that is the component-wise min of a set of personalized PageRanks
centered on k trusted sites. Our main results are that Min-k-PPR is, itself,
a type of PageRank & that it is expensive to spam. We elucidate a
surprisingly elegant algebra for PageRank. We define the space of all
possible PageRanks & show that this space is closed under some operations.
Most notably, we show that PageRanks are closed under (normalized)
component-wise min, which establishes that (normalized) Min-k-PPRis a
PageRank. This algebraic structure is also key to demonstrating the spam
resistance of Min-k-PPR."
-- 1803.05001@[arXiv]['18].
%A E. Ormsby
%T Darkest Web: Drugs, death and destroyed lives ... the inside story of the
internet's evil twin
%I AllenAndUnwin
%M MAR
%D 2018
%K book, c2018, c201x, c20xx, zz0418, internet, web, dark web, darkweb,
drugs, Silk Road, SilkRoad, crime, hitmen, hitman, vice, porn,
Daisy's Destruction, hurtcore, red room, rooms, 419, scam, scams, fraud,
bitcoin, Dread Pirate Roberts, Variety Jones, deep web, deepweb
%X 1st ed 2018; pb us$26; uk us isbn:1760297852; uk us isbn13:978-1760297855.
"Hitmen for hire, drugs for sale. Inside the dangerous world that lurks
beneath the bright, friendly light of your internet screen Dark..."
(The author was interviewed on ABC radio
[www][17/4/2018].
Drugs? -- yes. Hitmen? -- mostly (all?) cons.
Also see [wikip]['18].)