Open Hub
Open Hub

 Settings |  Report Duplicate

0
I Use This!
Very Low Activity

Commits : Listings

Analyzed about 3 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
Section 1522 updated
K.Srinivasan
More... over 2 years ago
+1522. (THEORY and FEATURE) Connections between Condorcet Jury Theorem (CJT) Asymptotic Derandomization and Bell's inequality-CHSH test, CHSH test generalized to arbitrary observers - a conjecture and its implications - 11 June 2024 - related to 1152,1399,1521 and all sections on Boolean composition of Majority function and VoterSAT, KRW Conjecture, Communication Complexity, Hardness Amplification, BQP factorization
Srinivasan Kannan
More... over 2 years ago
Section 1521 updated
K.Srinivasan
More... over 2 years ago
Section 1515 updated
K.Srinivasan
More... over 2 years ago
Section 1521 updated
K.Srinivasan
More... over 2 years ago
Section 1206 updated
Srinivasan Kannan
More... over 2 years ago
Section 1486 updated
Srinivasan Kannan
More... over 2 years ago
1521. (THEORY and FEATURE) Random Matrix Least Square SAT Solver - Python 3.7 Upgrade and implementation of Majority+VoterSAT boolean composition - related to 161,318,368,429,453,517,518,592,620,678,682,683,687 and all sections on Boolean Majority + VoterSAT composition, KRW Conjecture, Election analytics, Opinion Mining - 6 June 2024
Srinivasan Kannan
More... over 2 years ago
----------------------------------------------------------------------------------------------------------------------------------------- 1521. (THEORY and FEATURE) Random Matrix Least Square SAT Solver - Python 3.7 Upgrade and implementation of Majority+VoterSAT boolean composition - related to 161,318,368,429,453,517,518,592,620,678,682,683,687 and all sections on Boolean Majority + VoterSAT composition, KRW Conjecture, Election analytics, Opinion Mining - 6 June 2024 ----------------------------------------------------------------------------------------------------------------------------------------- 1521.1 CNFSATSolver.py has been upgraded to Python 3.x by 2to3 and autopep8 utilities. 1521.2 New multipurpose function majorityvotingCSP(self,VoterSATs,Assignments) has been defined that simulates Majority+VoterSAT boolean composition voting on 2 candidates 0 and 1 which has been discussed in multiple places in NeuronRain theory drafts in the context of hardness amplification, KRW conjecture among others. This function accepts an array of VoterSATs (one per voter) and an array of assignments (one per voter) and verifies if the assignment to variables satisfies the CNFSAT formula of the voter by function satisfy(). 1521.3 Binary assignment and Fractional assignment are set to be equal while invoking satisfy(). 1521.4 Total number of binary votes (number of formulas satisfied by assignments - which is equal to number of votes to candidate 1) and total number of MAXSAT votes (which is the percentage of clauses satisfied - real values) are computed. 1521.5 New boolean flag MajorityVotingCSP chooses between two clauses - (*) to compute Majority+VoterSAT boolean composition by majorityvotingCSP() or (*) Invoke SAT Solver by any of the standard least square algorithms by default set to LSMR. 1521.6 Following logs respectively capture: 1521.6.1 CNFSATSolver.Python3.7Upgrade.MajorityVotingCSP.log2.6June2024 - An example set of 1000 (Randomly generated CNFs) Random Matrix Least Square SAT Solver iterations for 5 variables and 5 clauses (alpha=1). All 1000 CNFSATs have been satisified 100%. 1521.6.2 CNFSATSolver.Python3.7Upgrade.MajorityVotingCSP.log.6June2024 - Simulation of a majority voting with 5 voters each having their own VoterCNFSAT and associated assignments to those VoterSATs by function majorityvotingCSP() - input CNFSATs for this function have to be generated based on Voter preferences. For example, following 3 constraints by voter can be converted to 3 binary variables in a CNFSAT clause which requires a preprocessing routine: (1) Candidate should be above 60 years age (x1 set to 1 if constraint is satisfied) (2) Candidate should have atleast 30 years experience (x2 set to 1 if constraint is satisfied) (3) Candidate should have atleast 2 academic degrees (x3 set to 1 if constraint is satisfied). 1521.7 In real world majority voting elections, identifying VoterSATs for individual voters could be a breakthrough in psephology as satisfying assignments to them found by an efficient SAT or MAXSAT Solver could maximize the votes.
Srinivasan Kannan
More... over 2 years ago
Section 385 updated
K.Srinivasan
More... over 2 years ago
Section 216 updated
K.Srinivasan
More... over 2 years ago
Recursive Lambda Function Growth compositions visualized by DALL-E 3 - OpenArtAI - https://openart.ai
K.Srinivasan
More... over 2 years ago
Section 1515 updated
Srinivasan Kannan
More... over 2 years ago
Section 1206 updated
Srinivasan Kannan
More... over 2 years ago
+1518. (THEORY and FEATURE) Governing Equation Discovery from Non-linear timeseries by SINDy (Sparse Identification of Non-Linear Dynamics) - IMD Rainfall timeseries (1901-2017) - 24 May 2024 - related to 690,691,692,1257,1315,1316,1327,1348,1367,1375,1376,1436,1437,1438,1515,1517 and all sections on Computational Economics-Quantitative Finance and Stockquote TimeSeries Analysis, Climate and Weather analytics, Graphical Event Models (GEM), Causal Event Models (CEM), Chain Event Graphs(CEG), Chaos and Fractals, EventNet causality, Multifractal Detrended Fluctuation Analysis (MFDFA), Linear and Logistic regression, Global Warming
Srinivasan Kannan
More... over 2 years ago
--------------------------------------------------------------------------------------------------------------------------------------------- 1518. (THEORY and FEATURE) Governing Equation Discovery from Non-linear timeseries by SINDy (Sparse Identification of Non-Linear Dynamics) - IMD Rainfall timeseries (1901-2017) - 24 May 2024 - related to 690,691,692,1257,1315,1316,1327,1348,1367,1375,1376,1436,1437,1438,1515,1517 and all sections on Computational Economics-Quantitative Finance and Stockquote TimeSeries Analysis, Climate and Weather analytics, Graphical Event Models (GEM), Causal Event Models (CEM), Chain Event Graphs(CEG), Chaos and Fractals, EventNet causality, Multifractal Detrended Fluctuation Analysis (MFDFA), Linear and Logistic regression, Global Warming --------------------------------------------------------------------------------------------------------------------------------------------- 1518.1 Subdivisional monthly rainfall timeseries dataset (for 116 years - 1901 to 2017) from https://data.gov.in/resource/sub-divisional-monthly-rainfall-1901-2017 has been fit to SINDy model to discover governing equations for 2 subdivisions - Tamil Nadu and Coastal Andhra Pradesh. 1518.2 python-src/SINDy.py has been updated to print SINDy model coefficients and model equations within SINDy_fit(). New function read_rainfall_dataset(rainfalltimeseriesjson,datasource='IMD',subdivision='Tamil Nadu') has been defined to process the rainfall timeseries dataset subdivision-wise with clause for datasource (presently IMD). JSON and CSV datasets from https://data.gov.in/resource/sub-divisional-monthly-rainfall-1901-2017 have been downloaded and timeseries data for subdivision is extracted (1287 elements for 12 months per each year from 1901 to 2017 excluding Not Available - NA - columns). 1518.3 Following JSON and CSV datasets from Open Government Data (OGD) Platform India (https://data.gov.in/catalog/rainfall-india) have been added to NeuronRain AstroInfer repositories: RainfallTimeseries_Sub_Division_IMD_2017.csv RainfallTimeseries_Sub_Division_IMD_2017.json 1518.4 Following governing equations have been discovered for 2 subdivisions over the range of 116 years (1901-2017) - excerpts from testlogs/SINDy.log.24May2024 : ----------------------------------- Tamil Nadu ----------------------------------- model coefficients: [[1. 0. 0. 0. 0. 0. ] [0.1019425 0. 0. 0. 0. 0. ]] model equations: ['1.000 1', '0.102 1'] (x)' = 1.000 1 (y)' = 0.102 1 ----------------------------------- Coastal Andhra Pradesh ----------------------------------- model coefficients: [[1. 0. 0. 0. 0. 0.] [0. 0. 0. 0. 0. 0.]] model equations: ['1.000 1', '0.000'] (x)' = 1.000 1 (y)' = 0.000
Srinivasan Kannan
More... over 2 years ago
Section 1517 updated
Srinivasan Kannan
More... over 2 years ago
Updated Requirements.txt
Srinivasan Kannan
More... over 2 years ago
+1517. (THEORY and FEATURE) Governing Equation Discovery from Non-linear timeseries by SINDy (Sparse Identification of Non-Linear Dynamics) - Climate and Stock quote timeseries - 18 May 2024 - related to 690,691,692,1257,1315,1316,1327,1348,1367,1375,1376,1436,1437,1438,1515 and all sections on Computational Economics-Quantitative Finance and Stockquote TimeSeries Analysis, Climate and Weather analytics, Graphical Event Models (GEM), Causal Event Models (CEM), Chain Event Graphs(CEG), Chaos and Fractals, EventNet causality, Multifractal Detrended Fluctuation Analysis (MFDFA), Linear and Logistic regression
Srinivasan Kannan
More... over 2 years ago
--------------------------------------------------------------------------------------------------------------------------------------------- 1517. (THEORY and FEATURE) Governing Equation Discovery from Non-linear timeseries by SINDy (Sparse Identification of Non-Linear Dynamics) - Climate and Stock quote timeseries - 18 May 2024 - related to 690,691,692,1257,1315,1316,1327,1348,1367,1375,1376,1436,1437,1438,1515 and all sections on Computational Economics-Quantitative Finance and Stockquote TimeSeries Analysis, Climate and Weather analytics, Graphical Event Models (GEM), Causal Event Models (CEM), Chain Event Graphs(CEG), Chaos and Fractals, EventNet causality, Multifractal Detrended Fluctuation Analysis (MFDFA), Linear and Logistic regression --------------------------------------------------------------------------------------------------------------------------------------------- 1517.1 New source file SINDy.py has been added to NeuronRain repositories which implements SINDy governing equation discovery mentioned in 1515.5 by invoking PySINDy. Sparse Identification of Non-Linear Dynamics or SINDy is based on sparse regression that depends on minimum number of variables or predictors (number of terms in linear regression is minimized which is crucial for massive datasets). 1517.2 SINDy.py defines 3 functions SINDy_fit(),stockquote_SINDy_model(),precipitation_SINDy_model() which respectively fit (*) an arbitrary timeseries (*) stockquote ticker timeseries and (*) rainfall precipitation timeseries to a governing equation by SINDy.fit(). 1517.3 Logs python-src/testlogs/SINDy.log.18May2024 demonstrate equation discovery for arbitrary timeseries, historic stockquotes timeseries of 3 stocks MSFT,GOOG,AAPL and rainfall timeseries.
Srinivasan Kannan
More... over 2 years ago
Section 1469 updated
Srinivasan Kannan
More... over 2 years ago
Updated Requirements.txt
Srinivasan Kannan
More... over 2 years ago
1516. (THEORY and FEATURE) Kolmogorov-Arnold Networks - a fundamental replacement for perceptrons - implementation based on PyKAN - 17 May 2024 - related to 116 and all sections on Single and Multilayered Perceptron implementations in NeuronRain, BackPropagation, Transformers, Polynomial interpolation and Complement Diophantines, Circuit Range Avoidance (AVOID), Special case of Majority voting
Srinivasan Kannan
More... over 2 years ago
--------------------------------------------------------------------------------------------------------------------------------------------- 1516. (THEORY and FEATURE) Kolmogorov-Arnold Networks - a fundamental replacement for perceptrons - implementation based on PyKAN - 17 May 2024 - related to 116 and all sections on Single and Multilayered Perceptron implementations in NeuronRain, BackPropagation, Transformers, Polynomial interpolation and Complement Diophantines, Circuit Range Avoidance (AVOID), Special case of Majority voting --------------------------------------------------------------------------------------------------------------------------------------------- 1516.1 Initial version of utility implementation of Kolmogorov-Arnold Networks (KAN) based on recent paper https://arxiv.org/pdf/2404.19756 and its python implementation - PyKAN documentation - https://kindxiaoming.github.io/pykan/ has been (which has been found to be 100 times faster than conventional multilayer perceptrons in some benchmarks) added to NeuronRain AstroInfer repositories. 1516.2 KAN is based on Kolmogorov-Arnold Representation theorem (which solved Hilbert 13th problem) that essentially expresses a function of several variables f:[0,1]^n -> R as compostion sum of functions of single variables. 1516.3 KAN differs from perceptrons by replacing the linear weights by variable function weights called "edges". This initial implementation is based on PyTorch datasets and defines a function kolmogorov_arnold_fit(function,n_var=2,x=[],y=[],trainnum=1000,testnum=1000) which takes torch lambda function, number of variables and x-y values and computes: (1) the training_loss and test_loss of the KAN model learnt (2) and finds parameters a,b,c,d minimizing MSE |y - (cf(ax+b) + d)|^2 for function f. 1516.4 KAN could be useful in fitting a complementary or mutually range-avoiding set to a function (or a circuit) over several variables by bottom-up function (or circuit) composition of functions over single variable. For example Majority function over n voters Maj:[0,1]^n -> {0,1} could be broken down into composition sum of functions of individual voters or Majority circuit is decomposed to individual voter circuits and composed. 1416.5 Spline and KAN plots-logs are written to: figures/ testlogs/KolmogorovArnoldNetworks.fit.jpeg testlogs/KolmogorovArnoldNetworks.log.17May2024
Srinivasan Kannan
More... over 2 years ago
Sections 1375 and 1426 updated
K.Srinivasan
More... over 2 years ago
Section 1426 updated
K.Srinivasan
More... over 2 years ago
Section 1398 updated
K.Srinivasan
More... over 2 years ago
Section 1428 updated
K.Srinivasan
More... over 2 years ago
Section 1296 updated
K.Srinivasan
More... over 2 years ago
Secton 1489 updated
Srinivasan Kannan
More... over 2 years ago