Open Hub
Open Hub

 Settings |  Report Duplicate

0
I Use This!
Very Low Activity

Commits : Listings

Analyzed about 8 hours ago. based on code collected 1 day ago.
Nov 05, 2025 — Nov 05, 2026
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
792. (FEATURE) Social Network Analytics - LinkedIn Connections Analysis, Name filter and Python 3.7.5 upgrades - 4 March 2020
Srinivasan Kannan
More... over 6 years ago
--------------------------------------------------------------------------------------------------------------------------------------------- 784. (FEATURE) Social Network Analytics - LinkedIn Connections Analysis, Name filter and Python 3.7.5 upgrades - 4 March 2020 --------------------------------------------------------------------------------------------------------------------------------------------- 1.SocialNetworkAnalysis_PeopleAnalytics.py - Connections in a Professional Network Profile have been analyzed by Recursive Lambda Function Growth (RLFG) after applying name filters to strip off Human names of connections and only extract their professional roles. 2.Name filter works by looking up words in profile connections text in a dictionary and filters Proper nouns (Human Names) - rationale is a language dictionary should not contain human names or contain least human names and is a short-cut stop-gap algorithm to recognize proper nouns while an artificial intelligence culture-neutral name parsing alternative might involve extensive training data and deep learning. 3.Dictionary.txt file already committed for CompressedSensing has been preferred to nondictionaryword() in RecursiveGlossOverlap_Classifier.py which repetitively invokes slower REST lookup - first word in each line of Dictionary.txt (Concise Oxford Dictionary) has been parsed to create an array of words 4.Quite a few dependent python source files have been upgraded to Python 3.7.5 including the very necessary RecursiveLambaFunctionGrowth.py implementation for performance by autopep8 and 2to3 utilities. 5.Dense subgraph extracted from definition graph of professional descriptions of connections is an indication of prominent designations/skills pattern. 6.Machine Translation has been commented because of slow REST calls to Google Translate. 7.Unfeasible Exception has been handled in korner_entropy() - maximal independent set computation. 8.Only a fraction of connections profile contents is analyzed because of intensive computation. 9.logs at testlogs/SocialNetworkAnalysis_PeopleAnalytics.log.ConnectionsRLFGAnalysis.4March2020.gz contain exhaustive graph theoretic analysis of LinkedIn Connections of the author (K.Srinivasan - https://www.linkedin.com/in/srinivasan-kannan-608a671/) and to large extent core numbers reflect the designations/skills profile. 10.Of all the graph complexity measures, closeness centrality is more accurate than others - top ranked vertices which directly relate to designations are (and still there are Human names because dictionary attack is just an approximation): [('system', 0.07260504201680672), ('computer_architecture', 0.05979238754325259), ('made', 0.05585003232062055), ('ecosystem', 0.05464895635673624), ('son', 0.05163025210084034), ('John', 0.050760079312623926), ('male', 0.050760079312623926), ('servicing', 0.049373191899710706), ('logic', 0.04934323243860651), ('considered', 0.048403361344537814), ('forming', 0.048403361344537814), ('components', 0.048403361344537814), ('interdependent', 0.048403361344537814), ('unified', 0.048403361344537814), ('body', 0.048403361344537814), ('whole', 0.048403361344537814), ('living', 0.048403361344537814), ('act', 0.04657368101879927), ('computer', 0.04537815126050421), ('permanent_wave', 0.044582043343653247), ('engagement', 0.04364876385336743), ('science', 0.042352941176470586), ('organization', 0.042352941176470586), ('software', 0.042352941176470586), ('structure', 0.042352941176470586), ('hardware', 0.042352941176470586), ('human', 0.04069952305246423), ('offspring', 0.04069952305246423), ('1215', 0.04015686274509804), ('sign', 0.04015686274509804), ('Henry', 0.04015686274509804), ('youngest', 0.04015686274509804), ('French', 0.04015686274509804), ('Magna', 0.04015686274509804), ('compelled', 0.04015686274509804), ('Richard', 0.04015686274509804), ('barons', 0.04015686274509804), ('1167-1216', 0.04015686274509804), ('1216', 0.04015686274509804), ('throne', 0.04015686274509804), ('brother', 0.04015686274509804), ('England', 0.04015686274509804), ('death', 0.04015686274509804), ('1199', 0.04015686274509804), ('II', 0.04015686274509804), ('lost', 0.04015686274509804), ('King', 0.04015686274509804), ('possessions', 0.04015686274509804), ('Carta', 0.04015686274509804), ('succeeded', 0.04015686274509804), ('group', 0.03997917751171265), ('organisms', 0.039705882352941174), ('physical', 0.039705882352941174) ...
Srinivasan Kannan
More... over 6 years ago
Merge branch 'master' of https://github.com/shrinivaasanka/asfer-github-code
Srinivasan Kannan
More... over 6 years ago
NeuronRain AsFer - git pull - working copy merge
Srinivasan Kannan
More... over 6 years ago
Updated NeuronRain Theory Drafts - Set Partition Analytics
Srinivasan Kannan
More... over 6 years ago
Updated NeuronRain Theory Drafts - Set Partition Analytics
Srinivasan Kannan
More... over 6 years ago
Updated NeuronRain Theory Drafts - Set Partition Analytics
Srinivasan Kannan
More... over 6 years ago
783. (THEORY) Phonetic Syllable Boundaries Compression and Approximate Decompression - error bounds and success amplification - 1 March 2020 - related to all sections on Text Compression, Compressed Sensing and People Analytics - Similar name match
Srinivasan Kannan
More... over 6 years ago
Section number corrected to 782
Srinivasan Kannan
More... over 6 years ago
(THEORY and FEATURE) Text Compression, Compressed Sensing, Syllable boundaries decompression implementation - Python 3.7.5 upgrade and refactoring - 28 February 2020, 29 February 2020
Srinivasan Kannan
More... over 6 years ago
---------------------------------------------------------------------------------------------------------------------------------------------- 750. (THEORY and FEATURE) Text Compression, Compressed Sensing, Syllable boundaries decompression implementation - Python 3.7.5 upgrade and refactoring - 28 February 2020, 29 February 2020 ---------------------------------------------------------------------------------------------------------------------------------------------- 1.TextCompression.py vowelless text compression-decompression implementation has been upgraded to Python 3.7.5 by autopep8 and 2to3 utilities 2.Syllable boundaries text compression function from CompressedSensing.py has been imported to optionally choose between syllable boundary compression of text and earlier vowelless compression (which strips vowels across the board) by a boolean flag syllable_boundary 3.Earlier commented code for PyEnchant spelling suggestions has been uncommented to choose between PyEnchant maximum likelihood (prob() function) and Hidden Markov Model maximum likelihood (HiddenMarkovModel_MLE() function) estimators based on if suggestions is not-None and None respectively. 4.Both PyEnchant and HMM computed maximum likelihood words are printed to logs. 5.Logs at testlogs/TextCompression.log.28February2020 contain PyEnchant and Hidden Markov Model maximum likelihood estimations per decompressed string for both vowel compressed and syllable boundary compressed words 6.decompressedtext.txt contains the newly introduced approximate syllable boundary decompressed text:
Srinivasan Kannan
More... over 6 years ago
749. (THEORY and FEATURE) PAC Learning - Python 3.7.5 upgrade, Riemann Hypothesis, Patterns in Primes, class HNc and Algebraic Circuits - 21 February 2020, 22 February 2020 - related to 24,629 and all sections on Complement diophantines, Complementary Sets and Equations, Prime-Composite complementation and Patterns in Primes
Srinivasan Kannan
More... over 6 years ago
749. (THEORY and FEATURE) PAC Learning - Python 3.7.5 upgrade, Riemann Hypothesis, Patterns in Primes, class HNc and Algebraic Circuits - 21 February 2020 - related to 24,629 and all sections on Complement diophantines, Complementary Sets and Equations, Prime-Composite complementation and Patterns in Primes
Srinivasan Kannan
More... over 6 years ago
---------------------------------------------------------------------------------------------------------------------------------------------- 749. (THEORY and FEATURE) PAC Learning - Python 3.7.5 upgrade, Riemann Hypothesis, Patterns in Primes, class HNc and Algebraic Circuits - 21 February 2020 - related to 24,629 and all sections on Complement diophantines, Complementary Sets and Equations, Prime-Composite complementation and Patterns in Primes ---------------------------------------------------------------------------------------------------------------------------------------------- 1. PACLearning.py PAC learning implementation has been refactored and upgraded to Python 3.7.5. Binary encoded and bits mapped UTM Primes dataset has been PAC Learnt again for per-prime-bit-position boolean disjunctions of first 10000 primes (16 disjunctions for each bit of 16-bits prime integer). New refactored PACLearning() function accepts a generic datasource and a hypothesis dictionary as arguments. 2. Example logs for this PAC learning are at testlogs/PACLearning.PrimeBitsMappingConjunctions.Python3.7.5.21February2020 which are different from Python 2.7 approximate learning. 3. Prime pattern References in Section 24 (24.31, 24.32, 24.33) mention a prima facie pattern in juxtaposed primes especially their avoidance of same digit ending. PAC learnt hypothesis in testlogs/PACLearning.PrimeBitsMappingConjunctions.Python3.7.5.21February2020 show an intriguing avoidance pattern in 14th,15th and 16th bits for first 10000 primes though approximate. 4. Prime counting function (number of primes <= N) can be written in terms of non-trivial zeros of Riemann Zeta Function. Riemann Hypothesis (RH) conjectures all non-trivial zeros of Riemann Zeta Function are along critical line Re(z)=0.5. Riemann Zeta Function is just a rephrased version of Sieve of Eratosthenes and is expanded by Euler product. 5. From Section 629, Riemann Zeta Function has an algebraic computation tree circuit in HNc: "... Riemann Zeta Function can be represented as a disjunctive clauses or non-uniform Hilbert Nullstellensatz algebraic circuit in 0-1-HNc (0 if s is a non-trivial zero else 1) of product polynomials having complex roots (non-trivial zeros): RZF=product_of(1/[1-prime(k)^(-s)]) for all primes 2,3,5,7,..." 6.Approximate binary digit hypothesis disjunctions for primes dataset learnt by PAC and exact learning by Euler-Fourier polynomial imply pattern in spacings between consecutive primes and thus are indirect boolean prime counting functions. 7.Riemann Zeta Function has diophantine equation representation and thus in RE by MRDP theorem. Recent result of MIP*=RE implies a quantum multiparty interactive prover-classical verifier algorithm for diophantine and algebraic circuit versions of Riemann Zeta Function. It is notable and coincidental that distribution of Gaussian Unitary Ensemble eigenvalues in quantum world are identical to spacings of non-trivial zeros of RZF.
Srinivasan Kannan
More... over 6 years ago
748. (THEORY and FEATURE) Leaky Bucket Timeseries Analyzer Implementation - Python 3.7.5 - 12 February 2020
Srinivasan Kannan
More... over 6 years ago
---------------------------------------------------------------------------------------------------------------------------------------------- 748. (THEORY and FEATURE) Leaky Bucket Timeseries Analyzer Implementation - Python 3.7.5 - 12 February 2020 ---------------------------------------------------------------------------------------------------------------------------------------------- 1.This commit implements Leaky Bucket Timeseries Analyzer Algorithm mentioned in GRAFIT course material https://github.com/shrinivaasanka/Grafit/blob/master/course_material/NeuronRain/AdvancedComputerScienceAndMachineLearning/AdvancedComputerScienceAndMachineLearning.txt for generic timeseries analytics. 2.Streaming_LeakyBucket.py instantiates Streaming Abstract Generator buffer incoming stream (example Dictionary.txt file data source is analyzed as timeseries) for generic data source and data storage of specified buffer size. 3.leaky_bucket_timeseries_analyzer() iterates through buffer incoming stream and in a loop populates a deque object till it reaches maximum size . Elements exceeding max size are overflown and printed. Deque simulates a leaky bucket - incoming elements fill the bucket from front end (insert() at index 0) and outgoing elements leak through other end. 4.Separate thread outgoing_stream_thread() independently scans the leaky bucket buffer and pops the elements from the deque. It busy waits if the leaky bucket deque is empty and continues to pop after buffer is refilled. 5.Synchronization is by a binary semaphore guarding the deque in main thread and outgoing stream thread. 6.Theoretical overtone of leaky bucket is its ability to flag outliers by overflows and peaks in timeseries are often the abnormalities in stream causing overflow. For example excessive network traffic corresponds to peaks in timeseries data and frequent overflows in leaky bucket as buffer fills faster than leakage. 7.Leaky Bucket Analyzer thus is a universal tool for mining any stream of text,audio,video,people data and most apt for wireless network stream, stream of celestial bodies configurations, climate change datastream (e.g http://climate.nasa.gov data stream on surface temperature, melting polar ice sheets, carbon emission, ocean currents and acidity), economic datasets having lot of outlier spikes. 8.Streaming_LeakyBucket.py has been implemented on Python 3.7.5 and Streaming_AbstractGenerator.py has been upgraded to Python 3.7.5 9.Logs for dictionary time series example are at testlogs/Streaming_LeakyBucket.log.12February2020
Srinivasan Kannan
More... over 6 years ago
747. (THEORY and FEATURE) Computational Geometric Spark NC-PRAM-BSP Factorization Benchmarks - 2014 bits, 2037 bits, 2050 bits integers - Quadcore Spark 2.4.3 + Python 3.7.5 - 5 February 2020
Srinivasan Kannan
More... almost 7 years ago
----------------------------------------------------------------------------------------------------------------------------------------------- 747. (THEORY and FEATURE) Computational Geometric Spark NC-PRAM-BSP Factorization Benchmarks - 2014 bits, 2037 bits, 2050 bits integers - Quadcore Spark 2.4.3 + Python 3.7.5 - 5 February 2020 ----------------------------------------------------------------------------------------------------------------------------------------------- 1.Computational Geometric Factorization Spark NC-PRAM-BSP implementation has been benchmarked on 3 2000+ bits integers both smooth and hard and factorization logs have been captured in testlogs/DiscreteHyperbolicFactorizationUpperbound_TileSearch_Optimized.Python3.7.5Upgrade.2000plusbits.5February2020 2.Following are the 2000+ bits integers benchmarked: 2.1 2014 bits smooth integer (heuristically integers having lot of 9s are easy to factorize) - all 9s - 999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999 (first few factors are 3,9,13,33,... found from 18 seconds onwards) 2.2 2037 bits less smooth integer - mixed digits,most 9s - 12891720999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999987108279 (first few factors are 3,9,11,27,33,... found from 18 seconds onwards) 2.3 2050 bits hard integer (heuristically integers having lot of 7s are hard to factorize) - mixed digits,most 7s - 163335395099872743147480777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777777614442382677905034630297 (first few factors are 7,9,11,21,27,... found from 204 seconds onwards - unusually high duration compared to previous 2000+ bits integers implies hardness) 3. Previous 2000plus bits integers are the largest factorized by NeuronRain AstroInfer Spark NC-PRAM-BSP implementation. 4. Depth of the non-uniform circuits are 182,184 and 184 respectively for 2014 bits, 2037 bits and 2050 bits integers. PRAM time O(logN)^k) for 2050 bits factorization circuit of depth 184 therefore is O(2050^184) and number of PRAMs required are O(2^2050/2050^184) = O(2^2050/2^2025) = O(2^25) = 33554432 which is high number of multicores but plausible in a geographically distributed BSP cloud. 5. Increasing depth slows down the PRAM time but reduces the parallel RAMs. Notwithstanding this theoretically prohibitive parallelism factors of 2050 bits hard integer were found in quadcore in few minutes.
Srinivasan Kannan
More... almost 7 years ago
--------------------------------------------------------------------------------------------------------------------------------------------- 745. (THEORY and FEATURE) Compressed Sensing - Alphabet-Syllable vectorspace embedding - Python 3.7.5 upgrade - update - 4 February 2020 --------------------------------------------------------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------------------------------------------------------------------------- 746. (THEORY and FEATURE) Social Network Analysis - People Analytics - Alphabet-Syllable vectorspace embedding - Python 3.7.5 upgrade - Name clustering - 4 February 2020 ----------------------------------------------------------------------------------------------------------------------------------------------
Srinivasan Kannan
More... almost 7 years ago
--------------------------------------------------------------------------------------------------------------------------------------------- 745. (THEORY and FEATURE) Compressed Sensing - Alphabet-Syllable vectorspace embedding - Python 3.7.5 upgrade - update - 4 February 2020 --------------------------------------------------------------------------------------------------------------------------------------------- 1. Function syllable_boundary_text_compression(self, text, vectorspace_embedding=True) in CompressedSensing python implementation has been updated to parametrically choose alphabet vectorspace embedding of words in text described in GRAFIT course material on Geometric Search - https://github.com/shrinivaasanka/Grafit/blob/master/course_material/NeuronRain/LinuxKernelAndCloud/BigdataAnalyticsCloud_CourseNotes.txt - words of text are embedded in |A|^m vectorspace (|A| is the size of natural language alphabet and m is the longest word in text). For English, this is a space of 26^m. 2.Words are python tokenized to vectors and concatenated to an array creating a vectorspace embedding words in the text document. 3.Return value is a tuple which includes or excludes the word embedding of text depending on vectorspace_embedding boolean flag 4.CompressedSensing.py has been upgraded to Python 3.7.5 by autopep8 and 2to3 utilities. Python 2.7 version is separately committed. 5.An Example sentence is syllable hyphenated (PyHyphen), vowel stripped-compressed and embedded as: Vowelless Syllable Vector Compression for text - This sentence is alphabet-syllable vector space embedded : (['This', 'sen', 'tence', 'al', 'pha', 'bet', '-', 'syl', 'la', 'ble', 'vec', 'tor', 'space', 'em', 'bed', 'ded'], 'Th_s-s_n-t_nc_-_l-ph_-b_t---syl-l_-bl_-v_c-t_r-sp_c_-_m-b_d-d_d') Alphabet vectorspace embedding of text: [['T', 'h', 'i', 's'], ['s', 'e', 'n', 't', 'e', 'n', 'c', 'e'], ['i', 's'], ['a', 'l', 'p', 'h', 'a', 'b', 'e', 't', '-', 's', 'y', 'l', 'l', 'a', 'b', 'l', 'e'], ['v', 'e', 'c', 't', 'o', 'r'], ['s', 'p', 'a', 'c', 'e'], ['e', 'm', 'b', 'e', 'd', 'd', 'e', 'd'], []] 6.This kind of embedding is unconventional because it lifts an one dimensional text to an m-dimensional vectorspace nullifying the meaning of the word whereas traditional word embeddings reduce dimensionality and map words in higher dimensional space to real vectorspace of lower dimensions. Following the linguistic philosophy of "word is characterized by company it keeps" by [Firth] (Studies in Linguistic Analysis - http://cs.brown.edu/courses/csci2952d/readings/lecture1-firth.pdf), alphabet embedding above could be a framework for assessing colocated words by euclidean distance e.g "vector" and "space" are colocated in previous example and are closer in meaning (e.g WordNet distance) but have high euclidean distance. This kind of alphabet embedding is quite relevant to similar name clustering where same name is spelt differently ("What is in a name" is a non-trivial question to answer).
Srinivasan Kannan
More... almost 7 years ago
744. (THEORY and FEATURE) Chaotic Pseudorandom Generator - Python 3.7.5 upgrade, Computational Geometric Factorization - 30 January 2020 - +related to all sections on Pseudorandom Generators, Chaos and Factorization
Srinivasan Kannan
More... almost 7 years ago
--------------------------------------------------------------------------------------------------------------------------------------------- 744. (THEORY and FEATURE) Chaotic Pseudorandom Generator - Python 3.7.5 upgrade, Computational Geometric Factorization - 30 January 2020 - related to all sections on Pseudorandom Generators, Chaos and Factorization --------------------------------------------------------------------------------------------------------------------------------------------- 1.This commit upgrades Chaos Pseudorandom Generator implementation to Python 3.7.5 by autopep8 followed by 2to3, which is the defacto NeuronRain recommended python version for performance reasons. 2.Both Logistic and Lehmer-Palmore PRGs have been tested on Python 3.7.5 + R-py2 combine and logs are committed to testlogs/ChaosAttractor.Python3.7.5Upgrade.30January2020 3.GRAFIT course material on Geometric search of BigData in https://github.com/shrinivaasanka/Grafit/blob/master/course_material/NeuronRain/LinuxKernelAndCloud/BigdataAnalyticsCloud_CourseNotes.txt describe a Randomized NC factorization algorithm based on Parallel Chaotic Pseudorandom Generator and Central Limit Theorem based Space filling Linear Program mentioned in drafts - https://sites.google.com/site/kuja27/ChaoticPRG.pdf and https://sites.google.com/site/kuja27/Analysis%20of%20a%20Randomized%20Space%20Filling%20Algorithm%20and%20its%20Linear%20Program%20Formulation.pdf . 4.Exact PRAM-NC-BSP computational geometric factorization algorithm implemented on Spark in NeuronRain AstroInfer is the derandomized version of Randomized NC Bernoulli trials factorization algorithm in https://github.com/shrinivaasanka/Grafit/blob/master/course_material/NeuronRain/LinuxKernelAndCloud/BigdataAnalyticsCloud_CourseNotes.txt which has the success probability of finding atleast one prime factor pr(m) = (1 - sqrt(2N)/(N*sqrt(2) - sqrt(2) - sqrt(N)))^(m-1)*(sqrt(2N)/(N*sqrt(2) - sqrt(2) - sqrt(N)) after m trials for square embedded hyperbolic arc. Expected number of trials E(m) before atleast one factor is found = sum_x=1-to-N(x*pr(x)) 5.Thus NeuronRain Theory drafts define 3 different Nick's class Computational Geometric Integer factorization algorithms - Randomized NC, Quantum NC, Exact NC - which transition from a randomized, cohered quantum state to classical exact decohered state.
Srinivasan Kannan
More... almost 7 years ago
728. (THEORY and FEATURE) Set Partition to Tile Cover Reduction by Computational Geometric Factorization - Least Squares Approximate Solution for the underdetermined linear equations, Exact Money Changing Problem(MCP) Solution, Quadratic Programs, Pell Equation, Shell Turing Machines(Category of Topological Spaces) - related to 577,624,651 - 19 November 2019, 20 November 2019,6 December 2019, 24 January 2020, 26 January 2020,27 January 2020
Srinivasan Kannan
More... almost 7 years ago
728. (THEORY and FEATURE) Set Partition to Tile Cover Reduction by Computational Geometric Factorization - Least Squares Approximate Solution for the underdetermined linear equations, Exact Money Changing Problem(MCP) Solution, Quadratic Programs, Pell Equation, Shell Turing Machines(Category of Topological Spaces) - related to 577,624,651 - 19 November 2019, 20 November 2019,6 December 2019, 24 January 2020, 26 January 2020,27 January 2020
Srinivasan Kannan
More... almost 7 years ago
728. (THEORY and FEATURE) Set Partition to Tile Cover Reduction by Computational Geometric Factorization - Least Squares Approximate Solution for the underdetermined linear equations, Exact Money Changing Problem(MCP) Solution, Quadratic Programs, Pell Equation, Shell Turing Machines(Category of Topological Spaces) - related to 577,624,651 - 19 November 2019, 20 November 2019,6 December 2019, 24 January 2020, 26 January 2020
Srinivasan Kannan
More... almost 7 years ago
728. (THEORY and FEATURE) Set Partition to Tile Cover Reduction by Computational Geometric Factorization - Least Squares Approximate Solution for the underdetermined linear equations, Exact Money Changing Problem(MCP) Solution, Quadratic Programs, Pell Equation, Shell Turing Machines(Category of Topological Spaces) - related to 577,624,651 - 19 November 2019, 20 November 2019,6 December 2019, 24 January 2020, 26 January 2020
Srinivasan Kannan
More... almost 7 years ago
728. (THEORY and FEATURE) Set Partition to Tile Cover Reduction by Computational Geometric Factorization - Least Squares Approximate Solution for the underdetermined linear equations, Exact Money Changing Problem(MCP) Solution, Quadratic Programs, Pell Equation, Shell Turing Machines(Category of Topological Spaces) - updated
Srinivasan Kannan
More... almost 7 years ago
743. (THEORY and FEATURE) Chaos Pseudorandom Generators - implementation update - Logistic and Lehmer-Palmore PRGs - 21 January 2020 - related to 318, 740 and all sections on Intrinsic Merit, Majority Voting, Ramsey coloring, Hardness Amplification, One-Way Functions, Chaos, Pseudorandomness and draft Randomized NC Chaos PRG defined in https://sites.google.com/site/kuja27/ChaoticPRG.pdf - References
Srinivasan Kannan
More... almost 7 years ago
743. (THEORY and FEATURE) Chaos Pseudorandom Generators - implementation update - Logistic and Lehmer-Palmore PRGs - 21 January 2020 - related to 318, 740 and all sections on Intrinsic Merit, Majority Voting, Ramsey coloring, Hardness Amplification, One-Way Functions, Chaos, Pseudorandomness and draft Randomized NC Chaos PRG defined in https://sites.google.com/site/kuja27/ChaoticPRG.pdf
Srinivasan Kannan
More... almost 7 years ago
----------------------------------------------------------------------------------------------------------------------------------------------- 743. (THEORY and FEATURE) Chaos Pseudorandom Generators - implementation update - Logistic and Lehmer-Palmore PRGs - 21 January 2020 - related to 318, 740 and all sections on Intrinsic Merit, Majority Voting, Ramsey coloring, Hardness Amplification, One-Way Functions, Chaos, Pseudorandomness and draft Randomized NC Chaos PRG defined in https://sites.google.com/site/kuja27/ChaoticPRG.pdf ----------------------------------------------------------------------------------------------------------------------------------------------- (*) Logistic Equation Chaos PRG implementation in ChaosAttractor.py has been updated to also compute Lehmer-Palmore PRG by a function argument. An if clause has been introduced to choose between parabola attractor rx(1-x) and Lehmer-Palmore PRG (Bx mod M) (*) ChaosAttractor.py Lorenz Attractor code has been refactored and made a function of parameters: ChaosPRG(algorithm="Logistic",seqlen=100,radix=10,initialcondition=0.7,prime=104729) which respectively choose the algorithm ("Logistic" equation or "Lehmer-Palmore" prime modulus PRGs), length of pseudorandom sequence, radix (k for "Logistic" and B for "Lehmer-Palmore"), initial condition for sensitive dependence and a Large prime modulus (*) R cor() correlation function has been invoked by R+rpy2 to print a correlation between pseudorandom sequence and a sample DJIA stock quotes dataset. (*) ChaosPRG() function can be optionally invoked in NeuronRain wherever pseudorandomness is necessary - e.g Pseudorandom shuffle simulation of paper ballot in Drone Electronic Voting Machine, Pseudorandom Shuffle of 2-colored (0 and 1) leaves in a boolean majority voting circuit which gives rise to monochromatic arithmetic progression order from pseudorandom chaotic shuffled votes disorder by Ramsey theory (Van Der Waerden numbers) (*) Section 318 states a Majority version of XOR Hardness Amplification Lemma and points 318.15 to 318.19 define hardness of inverting boolean majority + voter SAT composition which might imply existence of an one way function (a #P-complete counting problem). Emergence of arithmetic progression order from binary colored voter leaves of Boolean Majority Circuit stipulates a prima facie indistinguishability condition between a good PRG and true randomness - Good pseudorandom shuffle must minimize probability of emergence of monochromatic arithmetic progressions in the leaves of boolean majority circuit. (*) Probability of monochromatic arithmetic progressions in any pseudorandom bichromatic sequence of length N (including shuffled 2-colored voter leaves of boolean majority circuit) is inversely proportional to Van Der Waerden number VDW(N,2,r) - N elements in sequence(voters), 2 colors (candidates 0 and 1), arithmetic progression of length >= r (voters who chose same candidate are equally spread out in the shuffled bichromatic votes string). (*) Chaos implies disorder from order (Period Three Theorem implies any sequence of period 3 can have larger periods while Feigenbaum sequences define self-referential recursive universality order in Chaos) and monochromatic arithmetic progressions in Pseudorandom multichromatic sequences imply order from disorder - these are two mutually conflicting trends in diametrically opposite directions. (*) logs testlogs/ChaosAttractor.LehmerPalmorePRG.log.21January2020 and testlogs/ChaosAttractor.LogisticEquationPRG.log.21January2020 print a correlation between DJIA data and two chaotic PRGs - Logistic and Lehmer-Palmore. Opposing correlation signs between Logistic and Lehmer-Palmore PRGs and DJIA sample is an indicator of extent of chaos and pattern in economic data: Logistic: correlation coefficient of pseudorandom and DJIA sequences is: -0.00173395205970038 Lehmer-Palmore: correlation coefficient of pseudorandom and DJIA sequences is: 0.00648048041159691 (*) Equilibrium point where the aforementioned Orderly Disorder (Ramsey) and Disorderly Order (Chaos) coincide might be of independent area of research.
Srinivasan Kannan
More... almost 7 years ago