Search engine run on: http://users.monash.edu.au/
Glookbib search for: zz0419
%T Turing award
%M MAR
%D 2019
%K news, views, c2018, c201x, c0xx, zz0419, AI, deep learning, NN, ANN,
neural net, nets, Hinton, Bengio, LeCun
%X Geoffrey Hinton, Yoshua Bengio and Yann LeCun won the 2018(!) Turing award
for work on "deep learning".
-- bbc[27/3/2019],
AMT@[acm]['19].
Also see DL@[wikip]['19].
%A I. McEwan
%T Machines Like Me: A Novel
%I Penguin
%P 352
%M APR
%D 2019
%K book, c2019, c201x, c20xx, zz0419, SciFi, novel, robot, robots, android
%X 1st ed 2019; hb stg.13, uk isbn:1787331660, uk isbn13:978-1787331662;
hb us$18; us isbn:0385545118; us isbn13:978-0385545112.
"Machines Like Me takes place in an alternative 1980s London. Charlie,
drifting through life and dodging full-time employment, is in love with
Miranda, a bright student who lives with a terrible secret. When Charlie
comes into money, he buys Adam, one of the first synthetic humans and - with
Miranda's help - he designs Adam's personality. The near-perfect human that
emerges is beautiful, strong, and clever. It isn't long before a love
triangle soon forms, and these three beings confront a profound moral
dilemma. ..."
%A S. Arora
%A C. Lund
%A R. Motwani
%A M. Sudan
%A M. Szegedy
%T Proof verification and the hardness of approximation problems
%J JACM
%V 45
%N 3
%P 501–555
%D 1998
%K jrnl, JACM, c1998, c199x, c19xx, zz0419, PCP theorem, complexity, hardness,
NP, NPC, NPH, randomized, hardness, approximation algorithm, optimization,
optimisation, probablistic verifier, MAX SNP, Papadimitriou, Yannakakis,
probabilistically checkable proof, NEXP
%X "We show that every language in NP has a probablistic verifier that checks
membership proofs for it using logarithmic # of random bits & by examining a
const. # of bits in the proof. If a string is in the language, then there
exists a proof s.t. the verifier accepts with prob. 1 (i.e., for every choice
of its random string). For strings not in the language, the verifier rejects
every provided 'proof' with prob. >= 1/2. ... builds upon & improves a recent
result of Arora & Safra [1998] whose verifiers examine a nonconst. # of bits
in the proof (though this # is a v.slowly growing fn of the input length).
... [so] we prove that no MAX SNP-hard problem has a polynomial time
approximation scheme, unless NP=P. The class MAX SNP was defined by
Papadimitriou & Yannakakis [1991] & hard problems for this class include
vertex cover, maximum sat., max. cut, metric TSP, Steiner trees & shortest
superstring. We also improve upon the clique hardness results of Feige et al
[1996] & Arora & Safra [1998] & show that there exists a +ve &egr; such s.t.
the maximum clique size in an N-vertex graph to within a factor of N&egr; is
NP-hard."
-- [doi:10.1145/278298.278306]['19].
Also see,
"... An alternative [formulation is] the max. fraction of satisfiable
constraints of a constraint satisfaction problem is NP-hard to approx. within
some constant factor. ..."
-- PCP@[wikip]['19].
[Also search for: probablistic algorithm].
%A Li Sheng
%T 2-Role assignments on triangulated graphs
%J TCS
%V 304
%N 1-3
%P 201-214
%M JUL
%D 2003
%K jrnl, TCS, c2003, c200x, c20xx, zz0419, role assignment problem, 2role,
roleAssignment, social network, graph, maths, algorithm, indifference,
triangulated
%X "... G graph ... a k-role assignment is a fn mapping each vertex into a role,
a +ve integer 1,2,...,k, so that if x & y have the same role, then the sets
of roles assigned to their neighbors are the same. A graph is called a
triangulated graph [chordal] if ... provide a 'greedy' alg. for finding a
2-role assignment on a connected, non-bipartite triangulated graph with at
most one pendant vertex [leaf]. We characterize indifference graphs that have
a 2-role assignment."
-- [doi:10.1016/S0304-3975(03)00084-7]['19].
[Also search for: role assignment problem].
%A P. Crescenzi
%A M. Di Ianni
%A F. Greco
%A G. Rossi
%A P. Vocca
%T Making role assignment feasible: A polynomial-time algorithm for computing
ecological colorings
%J Graph-Theoretic Concepts in Computer Science
%P 90-100
%D 2008
%O LNCS vol. 5344
%K conf, GTCCS, c2008, c200x, c20xx, zz0419, role assignment problem,
graph, maths, algorithm
%X "... study the problem of ecologically coloring a graph. ... an ecological
coloring of a graph is a role assignment to the nodes of the graph, s.t. two
nodes surrounded by the same set of roles must be assigned the same role ...
prove that, for any simple undirected graph G with n_G distinct
neighborhoods and for any integer k with 1 <= k <= n_G , G admits an
ecological coloring which uses exactly k roles, and that this coloring can be
computed in polynomial time. ..."
-- [doi:10.1007/978-3-540-92248-3_9]['19].
[Also search for: role assignment problem].
%A D. Harvey
%A J. Van Der Hoeven
%T Integer multiplication in time O(n log n)
%J HALarchive
%M MAR
%D 2019
%K TR, c2019, c201x, c20xx, zz0419, fast, integer, multiplication, arithmetic,
algorithm, BigInt, nLogN, hal02070778, 02070778
%X "... Schonhage and Strassen conjectured in 1971 that the true complexity of
int.multn lies in Theta(n log n), & ... established their famous upper bound
M(n)=O(n log n log log n). In 2007 their result was sharpened by Furer
to M(n)=O(n log(n) K^{log^∗ n}) for some unspec. const. K>1, where
log^∗(n) denotes the iterated log, i.e., log^*(n)=min{k>=0: log^{k}(x)<=1}.
Prior to the present work, the record stood at M(n)=O(n log n 4^{log^*n}).
The main result of this paper is a verification of the upper bound in
S&S's conjecture, thus completely closing the remaining 4^{log^*(n)} gap..."
-- [www]['19].
Also see the abc[10/4/2019],
arith@[wikip]['21]
(and iteratedLog@[wikip]['19]).
[Also search for: fast integer multiplication].
%A A. Rubinstein
%A Z. Song
%T Reducing approximate longest common subsequence to approximate edit distance
%J arXiv
%M APR
%D 2019
%K TR, c2019, c201x, c20xx, zz0419, approximate, approximation, algorithm,
reduction, LCS, LCSS, Koucky, Saks, strings, longest common subsequence,
edit distance
%T "Given a pair of strings, the problems of computing their LCS & ED have been
extensively studied for decades. For exact algs., LCS & ED (with character
insertions & dels.) are equivalent; the state of the art running time is
(almost) quadratic & this is tight under plausible fine-grained complexity
assumptions. But for approximation algs. the picture is different: there is a
long line of works with improved approximation factors for ED, but for LCS
(with binary strings) only a trivial 1/2-approximation was known. ... we give
a reduction from approximate LCS to approximate ED, yielding the first
efficient (1/2+e)-approximation alg. for LCS for some constant e > 0."
-- 1904.05451@[arXiv][10/4/2019].
(Also see Koucky and Saks 1904.05459@[arXiv][10/4/2019].)
[Also search for: edit distance algorithm] and
[also search for: LCS algorithm] and
[also search for: strings approximation algorithm].
%A Y.-K. Yu
%A J. C. Wootton
%A S. F. Altschul
%T The compositional adjustment of amino acid substitution matrices
%V 100
%N 26
%P 15688-15693
%M DEC
%D 2003
%K jrnl, MolBio, PNAS, c2003, c200x, c20xx, zz0419, protein, amino acid,
substitution matrix, PAM, BLOSUM, biased, composition
%X "... shows that if one demands consistency between target & back-ground
freqs., then a log-odds substitution matrix implies a unique set of target &
b/g freqs. as well as a unique scale. Std substitution matrices therefore are
truly appropriate only for the comparison of proteins with standard AA
composition. [So] present & evaluate a rationale for transforming the target
freqs. implicit in a std matrix to freqs. appropriate for a non std context.
... yields asymmetric matrices for the comparison of proteins with divergent
compositions. Earlier approaches are unable to deal with this case in a fully
consistent manner. Composition-specific subs. matrix adjustment is shown
to be of utility for comparing compositionally biased proteins, inc. those of
organisms with nucleotide-biased, & therefore codon-biased, genomes or
isochores." [23 refs.]
-- [doi:10.1073/pnas.2533904100]['19].
[Also search for: MolBio substitution matrix].
%A Y.-K. Yu
%A S. F. Altschul
%T The construction of amino acid substitution matrices for the comparison of
proteins with non-standard compositions
%J Bioinformatics
%V 21
%N 7
%P 902-911
%M APR
%D 2005
%K jrnl, MolBio, c2005, c200x, c20xx, zz0419, protein, substitution matrix,
bias, PAM, BLOSUM
%X "... It has been argued elsewhere that standard matrices are not ideal for
such comparisons &, furthermore, a rationale has been presented for
transforming a std matrix for use in a non-std compositional context. ...
presents the mathematical details underlying the compositional adjustment of
amino acid or DNA substitution matrices. ..."
-- [doi:10.1093/bioinformatics/bti070]['19].
[Also search for: MolBio substitution matrix].
%A F. Chao
%A P. Gerland
%A A. R. Cook
%A L. Alkema
%T Systematic assessment of the sex ratio at birth for all countries and
estimation of national imbalances and regional reference levels
%J PNAS
%M APR
%D 2019
%K jrnl, PNAS, c2019, c201x, c20xx, zz0419, population, selective abortion,
birth, births, son, boys, daughter, girls, babies, sex ratio, stats, balance,
imbalance, SRB, China, India, human, biology, biol
%X "The sex ratio at birth (SRB; ratio of male to female live births) imbalance
in parts of the world over the past few decades is a direct consequence of
sex-selective abortion, driven by the coexistence of son preference ...
develop Bayesian methods for SRB estimation for all countries from 1950 to
2017. We model the SRB regional & national reference levels, the fluctuation
around national reference levels, & the inflation. The estimated regional
reference levels range from 1.031 (95% uncertainty interval [1.027; 1.036])
in sub-Saharan Africa to 1.063 [1.055; 1.072] in southeastern Asia, 1.063
[1.054; 1.072] in eastern Asia, & 1.067 [1.058; 1.077] in Oceania. We
identify 12 countries with strong statistical evidence of SRB imbalance
during 1970-2017, resulting in 23.1 [19.0; 28.3] million missing f.births
globally. The majority of those missing f.births are in China, with
11.9 [8.5; 15.8] million, & in India, with 10.6 [8.0; 13.6] million."
("... Under normal circumstances, SRB varies in a narrow range around 1.05
... maximum SRB since the start of [SRB-] inflation are in China (1.179
[1.141; 1.221] in 2005), Armenia (1.176 [1.150; 1.203] in 2000), ...")
-- [doi:10.1073/pnas.1812593116]['19].
[Also search for: sex ratio biology].
%A M. Parker
%T University, Ltd: Changing a business school
%J Organization
%M MAR
%D 2014
%K jrnl, c2014, c201x, c20xx, zz0419, university, management, faculty, teaching,
research, administration, deans, change, managerialism, KIP, KPIs, impact,
league tables, Leicester, UK
%X "... recounts two years in the change management of a European business
school. Using evidence gained whilst I was an employee, I describe the
condns & institutional mechanisms which allowed an 'earth shattering' change
programme to take place. This involved discounting the past & claiming that
anyone who is against change is either self-interested or doesn't
understand the 'real world'. The lack of collective & successful resistance
is of particular concern here, since it suggests that (in certain contexts)
academics are incapable of preventing their own institutions changing around
them. Exit appears to be the only choice, when loyalty is questioned &
voice impossible. I conclude with some observations on the relationship
between the managerial business school & the idea of the university."
-- [doi:10.1177/1350508413502646]['19]
or [le.ac.uk]['19].
[Also search for: university management].
%A M. Hoschele
%A A. Zeller
%T Mining input grammars from dynamic taints
%J Automated SWE (ASE) Conf.
%D 2016
%K conf, c2016 c201x, c20xx, zz0419, infer, learn, test, grammar, CFG, grammars,
program, behaviour, AUTOGRAM, SWE, UniSaarland
%X "Knowing which part of a program processes which parts of an input can
reveal the structure of the input [&] the structure of the program. In
a URL http://www.example.com/path/, ... the protocol http, the host
www.example.com, & the path 'path' would be handled by different fns & stored
in different variables. Given a set of sample inputs, we use dynamic
tainting to trace the data flow of each input char, & aggregate those input
fragments that would be handled by the same fn into lexical & syntactical
entities. The result is a CFG that reflects valid input structure. In its
evaluation, our AUTOGRAM prototype automatically produced readable &
structurally accurate grammars for inputs like URLs, spread-sheets or
config. files. ..."
-- pdf@[sa'rland]['19].
Also see AUTOGRAM@[sa'rland]['19].
[Also search for: infer grammar].
%A H. Childress
%T The Adjunct Underclass: How America's Colleges Betrayed Their Faculty,
Their Students, and Their Mission
%I UChicagoPress
%P 208
%M APR
%D 2019
%K book, text, c2019, c20x, c20xx, zz0419, university, academia, job, jobs,
management, exploitation, prof, professor, dean, deans, deanitis, pay,
tenure, part-time, teaching, higher education, college, gig economy, USA,
casual, casualisation
%X 1st ed 2019; hb us$16; uk us isbn:022649666X; uk us isbn13:978-0226496665.
"... Over the past few decades, the job of college professor has been utterly
transformed - for the worse. America's colleges & unis. were designed to
serve students & create knowledge through the teaching, research, & stability
that come with the longevity of tenured faculty, but higher ed. today is
dominated by adjuncts. In 1975, only 30% of faculty held temporary or
part-time posns. By 2011, as unis. faced both a decrease in public support
& ballooning admin. costs(!), that # topped 50%. Now, some surveys suggest
that as many as 70% of American professors are working course-to-course, with
few benefits, little to no security, & extremely low pay. ..."
[Also search for: University management].
%A H. Else
%T Impact factors are still widely used in academic evaluations
%J Nature
%V ?
%N ?
%M APR
%D 2019
%K news, vews, c2019, c201x, c20xx, zz0419, impact, KPI, KPIs, JIF, DORA,
journal, university, culture, promotion, hire, hiring, job, jobs, employment,
management, research, USA
%X "Almost half [~40%] of research-intensive universities consider jrnl impact
factors when deciding whom to promote, a survey of North American
institutions has found. ... has been widely criticized as a crude and
misleading proxy for the quality of scientists' work. ... unhealthy research
culture ..."
-- [doi:10.1038/d41586-019-01151-4]['19].
(Also see [doi:10.1038/d41586-018-01642-w][2/2019].)
Probably .uk and .au are much the same as .us?
[Also search for: university management].
%A A. Finkel
%T To move research from quantity to quality, go beyond good intentions
%J Nature
%V 566
%N 566
%P 297
%M APR
%D 2019
%K news, views, c2019, c201x, c20xx, zz0419, university, management, faculty,
KPI, KPIs, science, scientific research, Hindex, gIndex, KPI, KPIs, publish,
perish, publications, papers, journals, JIF, SNIP, citations, bibliometrics,
audit culture, misuse, impact, quality, ERA, ARC, Australia, chief scientist,
quote, quotable, Monash, Australia, topN, top5
%X "... Next, institutions must heed growing calls to abandon paper counting and
similar metrics for evaluating researchers. One alternative approach, the
Rule of Five, demonstrates a clear commitment to quality: candidates present
their best five papers over the past five years, accompanied by a description
of the research, its impact and their individual contribution. The exact
numbers are immaterial: what matters is the focus on quality. A handful of
institutions have required reviewers to consider individual contributions
rather than lists of publications, and the shift has not been easy. Reviewers
should be admonished for Googling individuals' h-indices and citation lists,
for example. Perseverance and self-reflection are essential. ..."
-- [doi:10.1038/d41586-019-00613-z]['19].
(Alan Finkel is Australian chief scientist and was
Chancellor of Monash University 2008-2016.)
[Also search for: publish perish] & [also search for: university management].
%A G. Tononi
%A C. Koch
%T Consciousness: Here, there and everywhere?
%J Phil. Trans. B
%I Royal Soc.
%M MAY
%D 2015
%K jrnl, c2015, c201x, c20xx, zz0419, human, brain, mind, consciousness,
conscious, thought, information, quale, philosophy, AI, NN, ANN,
%X "... we need not only more data but also a theory of consciousness one that
says what experience is & what type of physical systems can have it.
Integrated information theory (IIT) does so by starting from experience
itself via five phenomenological axioms: intrinsic existence, composition,
information, integration & exclusion. From these it derives five postulates
about the properties required of physical mechanisms to support
consciousness. The theory provides a principled account of both the quantity
& the quality of an individual experience (a quale), & a calculus to evaluate
whether or not a particular physical system is c. & of what. ... The theory
holds that c. is a fundamental property possessed by physical systems having
specific causal properties. It predicts that c. is graded, is common among
biological organisms & can occur in some very simple systems. Conversely, it
predicts that feed-forward networks, even complex ones, are not c., nor are
aggregates such as groups of individuals or heaps of sand. Also, in sharp
contrast to widespread functionalist beliefs, IIT implies that digital
computers, even if their behaviour were to be functionally equivalent to
ours, & even if they were to run faithful simulations of the human brain,
would experience next to nothing."
-- [doi:10.1098/rstb.2014.0167]['19].
Also see the bbc[27/3/2019].
[Also search for: consciousness].
%A D. Toker
%A F. T. Sommer
%T Information integration in large brain networks
%J PLoS Comp. Biol.
%V ?
%M FEB
%D 2019
%K jrnl, c2019, c201x, c20xx, zz0419, human, monkey, brain, mind, consciousness,
information, anterior, posterior, cortex
%X "... info. flow in the massively recurrent n/wks of the rest of the brain
have been lacking. To address this, recent work in info.theory has produced
a sound measure of network-wide 'integrated info.', which can be estimated
from time-series data. But, a computational hurdle has stymied attempts to
measure large-scale info. integration in real brains. Specifically, the
measurement of integrated info. involves a combinatorial search for the
informational 'weakest link' of a n/wk, a process whose computation time
explodes super-exponentially with n/wk size. Here, we show that spectral
clustering, applied on the correlation matrix of time-series data, provides
an approximate but robust soln to the search for the informational weakest
link of large n/wks. ... We evaluate this soln in brain-like systems of
coupled oscillators as well as in high-density electrocortigraphy data from
two macaque monkeys, & show that the informational 'weakest link' of the
monkey cortex splits posterior sensory areas from anterior association areas.
Finally, we use our soln to provide evidence in support of the long-standing
hypothesis that info. integration is maximized by n/wks with a high global
efficiency, & that modular n/wk structures promote the segregation of info.."
-- [doi:10.1371/journal.pcbi.1006807]['19].
Also see bbc[27/3/2019].
[Also search for: consciousness].
%A F. Mollica
%A S. T. Piantadosi
%J Humans store about 1.5 megabytes of information during language acquisition
%J Royal Soc. Open Sci.
%V 6
%N 3
%P MAR
%D 2019
%K jrnl, c2019, c201x, c20xx, zz0419, human, brain, mind, natural language,
linguistics, learn, knowledge, information theory, content
%X "... provide estimates at several levels of linguistic analysis: phonemes,
wordforms, lexical semantics, word frequency and syntax. Our best guess is
that the average English-speaking adult has learned 12.5 million bits of
information, the majority of which is lexical semantics. Interestingly, very
little of this information is syntactic, even in our upper bound analyses.
Generally, our results suggest that learners possess remarkable inferential
mechanisms capable of extracting, on average, nearly 2000 bits of information
about how language works each day for 18 years."
-- [doi:10.1098/rsos.181393]['19],
[mollicaf]['19].
[Also search for: natural language information].
%A K. Lee
%A D. Hoon Lee
%A J. Hwan Park
%A M. Yung
%T CCA security for self-updatable encryption: Protecting cloud data when
clients read/write ciphertexts
%J COMPJ
%V 62
%N 4
%P 545-562
%M APR
%D 2019
%K jrnl, COMPJ, c2019, c201x, c20xx, zz0419, encryption, chosen ciphertext, CCA,
cyphertext, SUE, public-key, cloud server, revocable, RS-ABE, RSABE,
networks, computer security
%X "Self-updatable encryption (SUE) is a new kind of public-key encryption,
motivated by cloud computing, which enables anyone (i.e., cloud server with
no access to private keys) to update a past ciphertext to a future ciphertext
by using a public key. The main applications of SUE are revocable-storage
attribute-based encryption (RS-ABE) that provides an efficient & secure
access control to encrypted data stored in cloud storage. ... a new threat
s.t. a revoked user still can access past ciphertexts ... RS-ABE solves this
problem by combining user revocation & ciphertext updating functionalities.
We propose the first SUE & RS-ABE schemes secure against a relevant form of
chosen-ciphertext security (CCA). Due to the fact that some ciphertexts are
easily derived from others, we employ a different notion of CCA that avoids
easy challenge related messages. Specifically, we define 'time extended
challenge' CCA security for SUE which excludes ciphertexts that are easily
derived from the challenge (over time periods) from being queried on. We then
propose an efficient SUE scheme with such CCA security, & we also present an
RS-ABE scheme with this CCA security."
-- [doi:10.1093/comjnl/bxy122]['19].
%A T. Mark Ellison
%T Discovering planar segregations
%R TR 90/5
%D 1990
%I Dept. Computer Science, University of Western Australia
%O Proc. of AAAI Spring Symp. on Machine Learning of
Natural Language and Ontology, pp.42-47, 1991
%K UWA, natural language, vowel, consonant, phoneme, inductive inference II,
information theory, MML, MDL, Occam, computational linguistics, languages,
algorithmic complexity, c1990, c199x, c19xx, zz0419
%X (simulated annealing, classifies letters / phonemes into
vowel / consonant / ... and infers patterns for words.)
"this report I present an algorithm for finding planar segregations of
phonemes for particular languages. This algorithm requires no domain-specific
knowledge of phonology or phonetics. Despite this lack of knowledge, the
implemented algorithm has identified the structurally significant
segregations for thirty languages"
-- [cites'er]['19],
[res.g.]['19].
[Also search for: natural language information].