representation   4122

« earlier    

[1710.03395] Efficient Dynamic Dictionary Matching with DAWGs and AC-automata
The dictionary matching is a task to find all occurrences of pattern strings in a set D (called a dictionary) on a text string T. The Aho-Corasick-automaton (AC-automaton) is a data structure which enables us to solve the dictionary matching problem in O(dlogσ) preprocessing time and O(nlogσ+occ) matching time, where d is the total length of the patterns in D, n is the length of the text, σ is the alphabet size, and occ is the total number of occurrences of all the patterns in the text. The dynamic dictionary matching is a variant where patterns may dynamically be inserted into and deleted from D. This problem is called semi-dynamic dictionary matching if only insertions are allowed. In this paper, we propose two efficient algorithms. For a pattern of length m, our first algorithm supports insertions in O(mlogσ+logd/loglogd) time and pattern matching in O(nlogσ+occ) time for the semi-dynamic setting and supports both insertions and deletions in O(σm+logd/loglogd) time and pattern matching in O(n(logd/loglogd+logσ)+occ(logd/loglogd)) time for the dynamic setting by some modifications. This algorithm is based on the directed acyclic word graph. Our second algorithm, which is based on the AC-automaton, supports insertions in O(mlogσ+uf+uo) time for the semi-dynamic setting and supports both insertions and deletions in O(σm+uf+uo) time for the dynamic setting, where uf and uo respectively denote the numbers of states of which the failure function and the output function need to be updated. This algorithm performs pattern matching in O(nlogσ+occ) time for both settings. Our algorithm achieves optimal update time for AC-automaton based methods, since any algorithm which explicitly maintains the AC-automaton requires Ω(uf+uo) update time.
algorithms  strings  rather-interesting  formal-languages  representation  rewriting-systems  to-write-about  consider:rediscovery  nudge-targets 
7 days ago by Vaguery
[1711.01012] Genetic Policy Optimization
Genetic algorithms have been widely used in many practical optimization problems. Inspired by natural selection, operators, including mutation, crossover and selection, provide effective heuristics for search and black-box optimization. However, they have not been shown useful for deep reinforcement learning, possibly due to the catastrophic consequence of parameter crossovers of neural networks. Here, we present Genetic Policy Optimization (GPO), a new genetic algorithm for sample-efficient deep policy optimization. GPO uses imitation learning for policy crossover in the state space and applies policy gradient methods for mutation. Our experiments on Mujoco tasks show that GPO as a genetic algorithm is able to provide superior performance over the state-of-the-art policy gradient methods and achieves comparable or higher sample efficiency.
machine-learning  genetic-algorithm  representation  reinvented-wheel  to-understand  to-write-about  evolutionary-algorithms 
7 days ago by Vaguery
(4510) Jordan Peterson LIVE: 12 Rules for Life - An Antidote to Chaos - YouTube - H3 Podcast #37 - Jordan Peterson - 59:00 Hiterl was a mouthpiece (not the man w the ideas) for resentment in population (exploitation w cheap nationalistic populism), Same again w Trump and Brexit and AfD. 'It was a collaboration between Hitler and the people.' again Brexit Trump - bridge between poor people and elites. Allister Heath "sacred and historic compact between elites and people” // Carl Jung - people dont have ideas, ideas have people. // Ben Shapiro / Milo Yiannopoulos / Douglas Murray / et al // Jordan Peterson Tweet about sexual assault/rape/abuse/violence post-Weinstein 'its time to get sex back into marriage.' is a anti post-modernism answer - retrotopia to the problem. Finds Michel Foucault a reprehensible, resentment-ridden individual // - With all the accusations of sex assault emerging (eg Louis CK) we are going to soon remember why sex was traditionally enshrined in marriage... // anti-ideology cult, pro free speech, anti professional demonstarator(s) indoctrinating feable //
Jordan  Peterson  Psychology  Ideology  book  post-modernism  Identity  LGBT  far-right  right-wing  alt-right  IQ  Sociology  free  speech  White  Supremacy  Conservative  radicalism  Feminism  feminist  sexism  Sexismus  gender-based  democracy  No  Representation  Culture  Politics  DonaldTrump  Donald  Trump  PC  political  correctness  Brexit  AfD  Islam  Racism  Rassismus  Xenophobia  Homophobia  hate  crime  Islamophobia  Retrotopia  Utopia  Foucault  Heidegger  Psychopath  Hitler 
12 days ago by asterisk2a
[1710.05183] Inferring Mesoscale Models of Neural Computation
Recent years have seen dramatic progress in the development of techniques for measuring the activity and connectivity of large populations of neurons in the brain. However, as these techniques grow ever more powerful---allowing us to even contemplate measuring every neuron in entire brain---a new problem arises: how do we make sense of the mountains of data that these techniques produce? Here, we argue that the time is ripe for building an intermediate or "mesoscale" computational theory that can bridge between single-cell (microscale) accounts of neural function and behavioral (macroscale) accounts of animal cognition and environmental complexity. Just as digital accounts of computation in conventional computers abstract away the non-essential dynamics of the analog circuits that implementing gates and registers, so too a computational account of animal cognition can afford to abstract from the non-essential dynamics of neurons. We argue that the geometry of neural circuits is essential in explaining the computational limitations and technological innovations inherent in biological information processing. We propose a blueprint for how to employ tools from modern machine learning to automatically infer a satisfying mesoscale account of neural computation that combines functional and structural data, with an emphasis on learning and exploiting regularity and repeating motifs in neuronal circuits. Rather than suggest a specific theory, we present a new class of scientific instruments that can enable neuroscientists to design, propose, implement and test mesoscale theories of neural computation.
dynamical-systems  machine-learning  deep-learning  representation  temporal-models  rather-interesting  inference  to-write-about 
12 days ago by Vaguery
[1711.02818] Enumeration of lozenge tilings of a hexagon with a shamrock missing on the symmetry axis
In their paper about a dual of MacMahon's classical theorem on plane partitions, Ciucu and Krattenthaler proved a closed form product formula for the tiling number of a hexagon with a "shamrock", a union of four adjacent triangles, removed in the center (Proc. Natl. Acad. Sci. USA 2013). Lai later presented a q-enumeration for lozenge tilings of a hexagon with a shamrock removed from the boundary (European J. Combin. 2017). It appears that the above are the only two positions of the shamrock hole that yield nice tiling enumerations. In this paper, we show that in the case of symmetric hexagons, we always have a simple product formula for the number of tilings when removing a shamrock at any position along the symmetry axis. Our result also generalizes Eisenk\"olbl's related work about lozenge tilings of a hexagon with two unit triangles missing on the symmetry axis (Electron. J. Combin. 1999).
tiling  domino-tiling  rather-interesting  combinatorics  enumeration  the-mangle-in-practice  representation  to-write-about  consider:the-convoluted-approach 
12 days ago by Vaguery
[1707.05589] On the State of the Art of Evaluation in Neural Language Models
Ongoing innovations in recurrent neural network architectures have provided a steady influx of apparently state-of-the-art results on language modelling benchmarks. However, these have been evaluated using differing code bases and limited computational resources, which represent uncontrolled sources of experimental variation. We reevaluate several popular architectures and regularisation methods with large-scale automatic black-box hyperparameter tuning and arrive at the somewhat surprising conclusion that standard LSTM architectures, when properly regularised, outperform more recent models. We establish a new state of the art on the Penn Treebank and Wikitext-2 corpora, as well as strong baselines on the Hutter Prize dataset.
natural-language-processing  representation  machine-learning  deep-learning  to-write-about 
12 days ago by Vaguery
11 Women of Color Politicians Running in 2018
1. Hiral Tipirneni (Arizona’s 8th Congressional district)
2. Dr. Mai Khanh Tran (California’s 39th Congressional district)
3. Aruna Miller (Maryland’s 6th Congressional district)
4. Fayrouz Saad (Michigan’s 11th Congressional district)
5. Angela Angel (Maryland’s District 25 State Delegate)
6. Mazie Hirono (U.S. Senate, Hawaii)
7. Virginia Madueño (U.S. House, California)
8. Stacey Abrams (Governor, Georgia)
9. Debra Haaland (U.S. House, New Mexico)
10. Lucy McBath (U.S. House, Georgia)
11. Sema Hernandez (U.S. Senate, Texas)
women  politics  socialJustice  representation 
14 days ago by campylobacter
Minus Infinity |
Today we’ll talk about some paradoxical things, like the logarithm of zero, and the maximum element of a set of real numbers that doesn’t contain any real numbers at all. More importantly, we’ll see how mathematicians try to wrap their heads around such enigmas.

All today’s logarithms will be base ten logarithms; so the logarithm of 100 is 2 (because 100 is 102) and the logarithm of 1/1000 is −3 (because 1/1000 is 10−3)). The logarithm of 0 would have to be an x that satisfies the equation 10x = 0. Since there’s no such number, we could just say “log 0 is undefined” and walk away, with our consciences clear and our complacency unruffled.
mathematical-recreations  math  representation  to-read  to-write-about 
15 days ago by Vaguery
[1711.02450] Shape Representation by Zippable Ribbons
Shape fabrication from developable parts is the basis for arts such as papercraft and needlework, as well as modern architecture and CAD in general, and it has inspired much research. We observe that the assembly of complex 3D shapes created by existing methods often requires first fabricating many small flat parts and then carefully following instructions to assemble them together. Despite its significance, this error prone and tedious process is generally neglected in the discussion. We propose an approach for shape representation through a single developable part that attaches to itself and requires no assembly instructions. Our inspiration comes from the so-called zipit bags, which are made of a single, long ribbon with a zipper around its boundary. In order to "assemble" the bag, one simply needs to zip up the ribbon. Our method operates in the same fashion, but it can be used to approximate any shape. Given a 3D model, our algorithm produces plans for a single 2D shape that can be laser cut in few parts from flat fabric or paper. We can then attach a zipper along the boundary for quick assembly and disassembly, or apply more traditional approaches, such as gluing and stitching. We show physical and virtual results that demonstrate the capabilities of our method and the ease with which shapes can be assembled.
via:gigasquid  computational-geometry  construction  manufacturability  representation  algorithms  to-write-about  via:twitter  topology 
15 days ago by Vaguery
[1710.07179] Increasing Labelings, Generalized Promotion, and Rowmotion
We generalize Bender-Knuth promotion on linear extensions to an analogous action on increasing labelings of any finite poset, in which the restrictions on the values of the labels satisfy a natural consistency condition. We give an equivariant bijection between such increasing labelings under this generalized promotion and order ideals in an associated poset under rowmotion. Additionally, we give a criterion for when certain kinds of toggle group actions on order ideals of a finite poset will be conjugate to rowmotion. These results build upon work of O.\ Pechenik with the first two authors in the case of rectangular increasing tableaux and work of N.\ Williams with the second author relating promotion and rowmotion on ranked posets. We apply these results to posets embedded in the Cartesian product of ranked posets and increasing labelings with labels between 1 and q, in which case we obtain new instances of the resonance phenomenon.
partial-ordering  combinatorics  representation  rather-interesting  to-understand  to-write-about  to-simulate  algorithms  consider:fitness  consider:inverse-problem 
15 days ago by Vaguery
[1708.08319] Fast Access to Columnar, Hierarchically Nested Data via Code Transformation
Big Data query systems represent data in a columnar format for fast, selective access, and in some cases (e.g. Apache Drill), perform calculations directly on the columnar data without row materialization, avoiding runtime costs.
However, many analysis procedures cannot be easily or efficiently expressed as SQL. In High Energy Physics, the majority of data processing requires nested loops with complex dependencies. When faced with tasks like these, the conventional approach is to convert the columnar data back into an object form, usually with a performance price.
This paper describes a new technique to transform procedural code so that it operates on hierarchically nested, columnar data natively, without row materialization. It can be viewed as a compiler pass on the typed abstract syntax tree, rewriting references to objects as columnar array lookups.
We will also present performance comparisons between transformed code and conventional object-oriented code in a High Energy Physics context.
databases  data-modeling  rather-interesting  algorithms  representation  to-understand  big-data 
15 days ago by Vaguery
[1710.01410] Learning Registered Point Processes from Idiosyncratic Observations
A parametric point process model is developed, with modeling based on the assumption that sequential observations often share latent phenomena, while also possessing idiosyncratic effects. An alternating optimization method is proposed to learn a "registered" point process that accounts for shared structure, as well as "warping" functions that characterize idiosyncratic aspects of each observed sequence. Under reasonable constraints, in each iteration we update the sample-specific warping functions by solving a set of constrained nonlinear programming problems in parallel, and update the model by maximum likelihood estimation. The justifiability, complexity and robustness of the proposed method are investigated in detail. Experiments on both synthetic and real-world data demonstrate that the method yields explainable point process models, achieving encouraging results compared to state-of-the-art methods.
modeling-is-not-mathematics  rather-interesting  representation  machine-learning  algorithms  nudge  nudge-targets  consider:representation  consider:performance-measures 
15 days ago by Vaguery
[1605.06848] Nonnegative Matrix Factorization Requires Irrationality
Nonnegative matrix factorization (NMF) is the problem of decomposing a given nonnegative n×m matrix M into a product of a nonnegative n×d matrix W and a nonnegative d×m matrix H. A longstanding open question, posed by Cohen and Rothblum in 1993, is whether a rational matrix M always has an NMF of minimal inner dimension d whose factors W and H are also rational. We answer this question negatively, by exhibiting a matrix for which W and H require irrational entries.
matrices  computational-geometry  rather-interesting  to-write-about  to-understand  representation  proof  consider:classification 
19 days ago by Vaguery

« earlier    

related tags

1%  10%  1980s  2002  2007  2011  2017  500  aapi  abstraction  accent  afd  algebra  algorithms  alt-right  alternative-computational-models  alternative  analog-circuits  analogy  analytical-expressions  anarchism  anxiety  aperiodic-tiling  approach  approximation  apps  archive  ariadnesthread  arrangements  art  artist  asynchronous  audience  austerity  automata  automated-conjecture  avatar  avoidance  award  bailout  bame  bank  basque  behavior  berlin  big-data  bigotry  biography  bioinformatics  blackness  blackwomen  book  boolean-networks  boxed  bretvictor  brexit  btw17  ca  capitalism  career  catalonia  cellular-automata  chaos  cities  clairebishop  class  classification  coding  cohesion  collective-behavior  collective-intelligence  combinatorics  comics  commentary  commodity  communication  community-detection  complexology  compressed-sensing  compression  computational-complexity  computational-geometry  computer-science  computer-vision  computing  concurrency  conference  congress  conservative  consider:algorithms  consider:cellular-automata  consider:classification  consider:feature-discovery  consider:fitness  consider:for-behavior-classification  consider:fractel-type-in-klapaucius  consider:game-applications  consider:including-operators  consider:inverse-problem  consider:looking-to-see  consider:mathematical-recreation  consider:open-questions  consider:out-of-box  consider:performance-measures  consider:rediscovery  consider:representation  consider:simplifying  consider:the-convoluted-approach  constraint-satisfaction  constructibility  construction  context  corporate  correctness  course.use  crime  criticism  culture  culturewars  data-modeling  data  database  databases  deep-learning  define-your-terms  democracy  design  diegoluna  digital-processing  digital  discourse  discrimination  display  dissonance  distributed-processing  diversity  domain-specific-languages  domino-tiling  donald  donaldtrump  door  dougashford  dynamical-systems  east  education  elite  embodiment  engineering-design  engineering  enumeration  equality  establishment  ethics  evasion  evolutionary-algorithms  evolvable  experimental-design  exploration  far-right  fashion  feature-construction  feature-extraction  feedback  feminism  feminist  finance  floating-point  formal-languages  foucault  fractals  free  gap  gender-based  gender  generative-models  genetic-algorithm  genetic-programming  geog107  geog327  germany  gesellschaft  gfc  glamourmagazine  goodiepal  government  graffiti  graph-theory  groupmaterial  grpc  hate  have-written-about  heidegger  heuristics  history  hitler  homophobia  horse-races  identity  ideology  ifs  image-processing  image-segmentation  image  income  independence  inequality  inference  information-theory  injustice  institution  institutional  intelligenceamplification  interaction  interest  internet  intersectionality  inverse-problems  iq  islam  islamophobia  jas  jeffreyscudder  jordan  journalism  julieault  justice  klassenkampf  knot-theory  kurdish  kurds  ladder  landscape  language  laws  learning-algorithms  legalaid  lens  lgbt  library  linear-algebra  linguists  lobby  looking-to-see  loops  lucylippard  machine-learning  magazine  manufacturability  marcelduchamp  market  martharosler  marx  materials-science  math  mathematical-programming  mathematical-recreations  mathematics  matrices  maybe-not-the-representation-i'd-start-from  media  meritocracy  meritocratic  meta-optimization  metamarket  metaphor  miwonkwon  mobility  modeling-is-not-mathematics  modeling  multivalent  nasty  natural-language-processing  neoliberal  neoliberalism  network-theory  neural-networks  no  non-binary-representation  nonlinear-dynamics  nudge-targets  nudge  null  number-theory  numbers  old-project  oligarchy  online  open-letter  open-problems  operations-research  optimization  ossi  out-of-the-box  painting  panamapapers  paradisepapers  partial-ordering  participatory  party  paulchan  pay  pc  pegida  performance-measure  performance  peterson  philosophy-of-engineering  philosophy-of-science  philosophy  photography  physics  plane-geometry  planning  play  plutocracy  poc  policy  political  political_knowledge  political_parties  politicians  politics  polyominoes  poor  portland  post-modernism  post-racial  poverty  precariat  prejudice  privacy  privilege  privileged  probability-theory  programming  proof  proprioception  protests  protobuf  psychology  psychopath  public  quantums  race  racism  radicalism  ramsey-theory  rassismus  rather-interesting  recovery  reinvented-wheel  reinventing-the-wheel-again  relational  religion  research  responsiveness  retrotopia  reunification  review  revolving  rewriting-systems  rich  right-wing  robotics  roleplay  run  satisfiability  saxony  science  scotland  screen  secondlife  secular  sexism  sexismus  sexuality  signal-processing  simulation  social-networks  social  social_choice  socialjustice  socialmedia  socialpractice  socialturn  society  sociology  software-development-is-not-programming  software-development  software  spain  speech  sports  stagnation  starwars  starwarsrogueone  statistics  steuerbetrug  steuerhinterziehung  steuertricks  stigma  strategy  strings  super  superresolution  supremacy  surface  system  systems-biology  tanlin  tasrillsieyes  tax  taxation  tech  television  temporal-models  text  the-mangle-in-practice  theoretical-biology  tiling  time-series  time-warping  to-cartoon-about  to-do  to-learn  to-read  to-simulate  to-understand  to-write-about-more  to-write-about  to-write  topology  tories  totalitarianism  towatch  transients  trap  traversal  tricky-rabbit  trump  typography  tyranny  uk  universality  unstable  unthinkable  usa  utopia  very-nice  vested  videogame  videos  votes  voting  w.a.g.e.  walls  what-gets-measured-gets-fudged  white  whiteness  women  work  working  worklife  writing  xenophobia  youtube  zusammenhalt 

Copy this bookmark: