TM, SP: fix possible underflow in seed. Default random (0) seed.

bindings/py/cpp_src/bindings/algorithms/py_SpatialPooler.cpp

@@ -173,9 +173,9 @@ Argument boostStrength A number greater or equal than 0, used to
         too much boosting may also lead to instability of SP outputs.
 
 
-Argument seed Seed for our random number generator. If seed is < 0
+Argument seed Seed for our random number generator. If seed is 0
         a randomly generated seed is used. The behavior of the spatial
-        pooler is deterministic once the seed is set.
+        pooler is deterministic once the seed is set > 0.
 
 Argument spVerbosity spVerbosity level: 0, 1, 2, or 3
 
@@ -197,7 +197,7 @@ Argument wrapAround boolean value that determines whether or not inputs
             , py::arg("minPctOverlapDutyCycle") = 0.001
             , py::arg("dutyCyclePeriod") = 1000
             , py::arg("boostStrength") = 0.0
-            , py::arg("seed") = 1
+            , py::arg("seed") = 0
             , py::arg("spVerbosity") = 0
             , py::arg("wrapAround") = true
         );

bindings/py/cpp_src/bindings/algorithms/py_TemporalMemory.cpp

@@ -112,7 +112,8 @@ Argument predictedSegmentDecrement
     0.01 = 0.0004
 
 Argument seed
-    Seed for the random number generator.
+    Seed for the random number generator. Default (0) means truly random, 
+    > 0 is a fixed pseudorandom-sequence. See Random.hpp. 
 
 Argument maxSegmentsPerCell
     The maximum number of segments per cell.
@@ -146,7 +147,7 @@ Argument anomalyMode (optional, default ANMode::RAW) selects mode for `TM.anomal
                 , py::arg("permanenceIncrement") = 0.1
                 , py::arg("permanenceDecrement") = 0.1
                 , py::arg("predictedSegmentDecrement") = 0.0
-                , py::arg("seed") = 42
+                , py::arg("seed") = 0
                 , py::arg("maxSegmentsPerCell") = 255
                 , py::arg("maxSynapsesPerSegment") = 255
                 , py::arg("checkInputs") = true

src/examples/hello/HelloSPTP.cpp

@@ -95,6 +95,12 @@ EPOCHS = 2; // make test faster in Debug
   Metrics statsSPglobal(outSPglobal, 1000);
   Metrics statsTM(outTM, 1000);
 
+  //uses fixed seed for deterministic output checks:
+  Random rnd(42);
+  spGlobal.setSeed(1);
+  spLocal.setSeed(1);
+  tm.setSeed(42);
+
   /*
    * For example: fn = sin(x) -> periodic >= 2Pi ~ 6.3 && x+=0.01 -> 630 steps to 1st period -> window >= 630
    */
@@ -112,6 +118,7 @@ EPOCHS = 2; // make test faster in Debug
   for (UInt e = 0; e < EPOCHS; e++) { //FIXME EPOCHS is actually steps, there's just 1 pass through data/epoch.
 
     //Encode
+    {
     tEnc.start();
     x+=0.01f; //step size for fn(x)
     enc.encode(sin(x), input); //model sin(x) function //TODO replace with CSV data
@@ -121,32 +128,38 @@ EPOCHS = 2; // make test faster in Debug
     tRng.start();
     input.addNoise(0.01f, rnd); //change 1% of the SDR for each iteration, this makes a random sequence, but seemingly stable
     tRng.stop();
+    }
 
     //SP (global and local)
     if(useSPlocal) {
-    tSPloc.start();
-    spLocal.compute(input, true, outSPlocal);
-    tSPloc.stop();
+      tSPloc.start();
+      spLocal.compute(input, true, outSPlocal);
+      tSPloc.stop();
+
+      outSP = outSPlocal;
     }
 
     if(useSPglobal) {
-    tSPglob.start();
-    spGlobal.compute(input, true, outSPglobal);
-    tSPglob.stop();
+      tSPglob.start();
+      spGlobal.compute(input, true, outSPglobal);
+      tSPglob.stop();
+
+      outSP = outSPglobal;
     }
-    outSP = outSPglobal; //toggle if local/global SP is used further down the chain (TM, Anomaly)
 
     // TM
     if(useTM) {
-    tTM.start();
+      tTM.start();
       tm.compute(outSP, true /*learn*/); //to uses output of SPglobal
       tm.activateDendrites(); //required to enable tm.getPredictiveCells()
-      outTM = tm.cellsToColumns( tm.getPredictiveCells() );
       tTM.stop();
+
+      outTM = tm.cellsToColumns( tm.getPredictiveCells() );
     }
 
 
     //Anomaly (pure x likelihood)
+    {
     an = tm.anomaly;
     avgAnom10.compute(an); //moving average
     if(e % 1000 == 0) {
@@ -157,6 +170,7 @@ EPOCHS = 2; // make test faster in Debug
     tAnLikelihood.start();
     anLikely = anLikelihood.anomalyProbability(an); 
     tAnLikelihood.stop();
+    }
 
 
     // print
@@ -202,27 +216,36 @@ EPOCHS = 2; // make test faster in Debug
 
       SDR goldSP({COLS});
       const SDR_sparse_t deterministicSP{
-        62, 72, 73, 82, 85, 102, 263, 277, 287, 303, 306, 308, 309, 322, 337, 339, 340, 352, 370, 493, 1094, 1095, 1114, 1115, 1120, 1463, 1512, 1518, 1647, 1651, 1691, 1694, 1729, 1745, 1746, 1760, 1770, 1774, 1775, 1781, 1797, 1798, 1803, 1804, 1805, 1812, 1827, 1828, 1831, 1832, 1858, 1859, 1860, 1861, 1862, 1875, 1878, 1880, 1881, 1898, 1918, 1923, 1929, 1931,1936, 1950, 1953, 1956, 1958, 1961, 1964, 1965, 1967, 1971, 1973, 1975, 1976, 1979, 1980, 1981, 1982, 1984, 1985, 1986, 1988, 1991, 1994, 1996, 1997, 1998, 1999, 2002, 2006, 2008, 2011, 2012, 2013, 2017, 2019, 2022, 2027, 2030
+        68, 79, 86, 98, 105, 257, 263, 286, 302, 306, 307, 309, 310, 313, 315, 317, 318, 320, 323, 325, 326, 329, 334, 350, 356, 363, 539, 935, 1089, 1093, 1098, 1111, 1112, 1118, 1120, 1124, 1133, 1508, 1513, 1521, 1624, 1746, 1765, 1774, 1775, 1776, 1780, 1784, 1787, 1802, 1804, 1811, 1813, 1815, 1819, 1844, 1845, 1865, 1876, 1884, 1891, 1900, 1903, 1904, 1908, 1909, 1925, 1926, 1928, 1932, 1933, 1943, 1947, 1952, 1955, 1959, 1961, 1962, 1964, 1966, 1967, 1969, 1970, 1971, 1973, 1975, 1977, 1980, 1981, 1982, 1983, 1985, 1987, 1991, 1994, 2002, 2004, 2011, 2027, 2030, 2031, 2045
       };
       goldSP.setSparse(deterministicSP);
 
       SDR goldSPlocal({COLS});
       const SDR_sparse_t deterministicSPlocal{
-        12, 13, 71, 72, 75, 78, 82, 85, 131, 171, 182, 186, 189, 194, 201, 263, 277, 287, 308, 319, 323, 337, 339, 365, 407, 429, 432, 434, 443, 445, 493, 494, 502, 508, 523, 542, 554, 559, 585, 586, 610, 611, 612, 644, 645, 647, 691, 698, 699, 701, 702, 707, 777, 809, 810, 811, 833, 839, 841, 920, 923, 928, 929, 935, 955, 1003, 1005, 1073, 1076, 1094, 1095, 1114, 1115, 1133, 1134, 1184, 1203, 1232, 1233, 1244, 1253, 1268, 1278, 1291, 1294, 1306, 1309, 1331, 1402, 1410, 1427, 1434, 1442, 1463, 1508, 1512, 1514, 1515, 1518, 1561, 1564, 1623, 1626, 1630, 1640, 1647, 1691, 1694, 1729, 1745, 1746, 1760, 1797, 1804, 1805, 1812, 1827, 1831, 1858, 1861, 1862, 1918, 1956, 1961, 1965, 1971, 1975, 1994, 2012
+        17, 71, 75, 79, 81, 86, 89, 164, 189, 198, 203, 262, 297, 314, 324, 326, 329, 337, 360, 379, 432, 443, 448, 452, 509, 520, 525, 526, 529, 536, 612, 619, 624, 630, 649, 652, 693, 717, 719, 720, 754, 810, 813, 815, 835, 839, 849, 884, 890, 914, 925, 931, 937, 945, 971, 1016, 1088, 1089, 1095, 1105, 1109, 1133, 1159, 1209, 1214, 1228, 1235, 1241, 1244, 1273, 1295, 1314, 1329, 1336, 1342, 1427, 1435, 1436, 1448, 1461, 1486, 1496, 1500, 1523, 1561, 1572, 1576, 1603, 1610, 1624, 1635, 1649, 1664, 1685, 1725, 1732, 1741, 1758, 1800, 1804, 1811, 1824, 1862, 1870, 1882, 1883, 1887, 1903, 1956, 1963, 1971, 1977, 1984, 2015
       };
       goldSPlocal.setSparse(deterministicSPlocal);
 
       SDR goldTM({COLS});
       const SDR_sparse_t deterministicTM{
-        72, 85, 102, 114, 122, 126, 287, 308, 337, 339, 542, 920, 939, 952, 1268, 1507, 1508, 1518, 1546, 1547, 1626, 1627, 1633, 1668, 1727, 1804, 1805, 1827, 1832, 1844, 1859, 1862, 1918, 1920, 1924, 1931, 1933, 1945, 1961, 1965, 1966, 1968, 1970, 1973, 1975, 1976, 1977, 1979, 1986, 1987, 1991, 1992, 1996, 1998, 2002, 2006, 2008, 2012, 2042, 2045
+        79, 92, 98, 128, 136, 286, 307, 309, 310, 313, 315, 325, 356, 454, 539, 1093, 1111, 1112, 1120, 1237, 1278, 1467, 1497, 1508, 1513, 1521, 1614, 1624, 1635, 1668, 1669, 1673,1699, 1774, 1775, 1776, 1780, 1784, 1808, 1813, 1815, 1816, 1821, 1827, 1845, 1865, 1900, 1911, 1913, 1925, 1929, 1931, 1932, 1933, 1940, 1941, 1947, 1949, 1955, 1956, 1959, 1961, 1962, 1964, 1966, 1967, 1968, 1969, 1970, 1972, 1975, 1977, 1978, 1979, 1981, 1982, 1985, 1987, 1988, 1991, 1994, 2027, 2030, 2044, 2045
       };
       goldTM.setSparse(deterministicTM);
 
-      const float goldAn    = 0.637255f; //Note: this value is for a (randomly picked) datapoint, it does not have to improve (decrease) with better algorithms
-      const float goldAnAvg = 0.40804f; // ...the averaged value, on the other hand, should improve/decrease. 
+      const float goldAn    = 0.470588f; //Note: this value is for a (randomly picked) datapoint, it does not have to improve (decrease) with better algorithms
+      const float goldAnAvg = 0.40961f; // ...the averaged value, on the other hand, should improve/decrease. 
 
 #ifdef _ARCH_DETERMINISTIC
       if(e+1 == 5000) {
+	// For debugging serialization: save SP's state in 1 step, comment out, recompile, load SP and compare in another 
+	// step 1:
+	spGlobal.saveToFile("/tmp/spG.save");
+	// step 2:
+	SpatialPooler resumedSP;
+	resumedSP.loadFromFile("/tmp/spG.save");
+	NTA_CHECK(spGlobal == resumedSP) << "SPs differ!";
+	// --end of debugging
+
         //these hand-written values are only valid for EPOCHS = 5000 (default), but not for debug and custom runs.
         NTA_CHECK(input == goldEnc) << "Deterministic output of Encoder failed!\n" << input << "should be:\n" << goldEnc;
         if(useSPglobal) { NTA_CHECK(outSPglobal == goldSP) << "Deterministic output of SP (g) failed!\n" << outSP << "should be:\n" << goldSP; }

src/htm/algorithms/Connections.cpp

@@ -31,6 +31,8 @@ using std::string;
 using std::vector;
 using namespace htm;
 
+const Permanence MARK_SYNAPSE_AS_REMOVED = -1.0;
+
 Connections::Connections(const CellIdx numCells, 
 		         const Permanence connectedThreshold, 
 			 const bool timeseries) {
@@ -210,14 +212,14 @@ bool Connections::synapseExists_(const Synapse synapse, bool fast) const {
   const bool found = (std::find(synapsesOnSegment.begin(), synapsesOnSegment.end(), synapse) != synapsesOnSegment.end());
   //validate the fast & slow methods for same result:
 #ifdef NTA_ASSERTIONS_ON
-  const bool removed = synapses_[synapse].permanence == -1;
-  NTA_ASSERT( (removed and not found) or (not removed and found) );
+  const bool removed = fabs(synapses_[synapse].permanence - MARK_SYNAPSE_AS_REMOVED) < 0.001f;
+  NTA_CHECK( (removed and not found) or (not removed and found) );
 #endif
   return found;
 
   } else {
   //quick method. Relies on hack in destroySynapse() where we set synapseData.permanence == -1
-  return synapses_[synapse].permanence != -1;
+  return fabs(synapses_[synapse].permanence - MARK_SYNAPSE_AS_REMOVED) < 0.001f;
   }
 }
 
@@ -261,8 +263,8 @@ void Connections::destroySegment(const Segment segment) {
   CellData &cellData = cells_[segmentData.cell];
 
   const auto segmentOnCell = std::find(cellData.segments.cbegin(), cellData.segments.cend(), segment);
-  NTA_ASSERT(segmentOnCell != cellData.segments.cend()) << "Segment to be destroyed not found on the cell!";
-  NTA_ASSERT(*segmentOnCell == segment);
+  NTA_CHECK(segmentOnCell != cellData.segments.cend()) << "Segment to be destroyed not found on the cell!";
+  NTA_CHECK(*segmentOnCell == segment);
 
   cellData.segments.erase(segmentOnCell);
   destroyedSegments_++;
@@ -313,15 +315,18 @@ void Connections::destroySynapse(const Synapse synapse) {
 					  [&](const Synapse a, const Synapse b) -> bool { return dataForSynapse(a).id < dataForSynapse(b).id;}
 					  ); 
 
-  NTA_ASSERT(synapseOnSegment != segmentData.synapses.cend());
-  NTA_ASSERT(*synapseOnSegment == synapse);
+  NTA_CHECK(synapseOnSegment != segmentData.synapses.cend());
+  NTA_CHECK(*synapseOnSegment == synapse);
 
   segmentData.synapses.erase(synapseOnSegment);
   //Note: dataForSynapse(synapse) are not deleted, unfortunately. And are still accessible. 
   //To mark them as "removed", we set SynapseData.permanence = -1, this can be used for a quick check later
-  synapseData.permanence = -1; //marking as "removed"
+  synapseData.permanence = MARK_SYNAPSE_AS_REMOVED; //marking as "removed"
+  //actually erase from synapses_ :
+  //const auto it = synapses_.begin() + synapse;
+  //synapses_.erase(it); //slow
   destroyedSynapses_++;
-  NTA_ASSERT(not synapseExists_(synapse));
+  NTA_ASSERT(not synapseExists_(synapse, false));
 }
 
 
@@ -768,6 +773,7 @@ bool Connections::operator==(const Connections &o) const {
   NTA_CHECK (cells_.size() == o.cells_.size()) << "Connections equals: cells_" << cells_.size() << " vs. " << o.cells_.size();
   NTA_CHECK (cells_ == o.cells_) << "Connections equals: cells_" << cells_.size() << " vs. " << o.cells_.size();
 
+  NTA_CHECK (segments_.size() == o.segments_.size()) << "Connections equals: segments_ size: " << segments_.size() << " vs. " << o.segments_.size();
   NTA_CHECK (segments_ == o.segments_ ) << "Connections equals: segments_";
   NTA_CHECK (destroyedSegments_ == o.destroyedSegments_ ) << "Connections equals: destroyedSegments_";
 
@@ -777,6 +783,7 @@ bool Connections::operator==(const Connections &o) const {
 
   //also check underlying datastructures (segments, and subsequently synapses). Can be time consuming.
   //1.cells:
+  /*
   for(const auto cellD : cells_) {
     //2.segments:
     const auto& segments = cellD.segments;
@@ -789,6 +796,7 @@ bool Connections::operator==(const Connections &o) const {
       }
     }
   }
+  */
 
 
   NTA_CHECK (connectedThreshold_ == o.connectedThreshold_ ) << "Connections equals: connectedThreshold_";
@@ -810,7 +818,10 @@ bool Connections::operator==(const Connections &o) const {
   NTA_CHECK (prunedSegs_ == o.prunedSegs_ ) << "Connections equals: prunedSegs_";
 
   } catch(const htm::Exception& ex) {
-	  std::cout << "Connection equals: differ! " << ex.what();
+    UNUSED(ex); //avoid warning if the debug below is not used.
+    //uncomment below to start debugging Connections op==. It's perfectly 
+    //..normal for unequality here. 
+    NTA_WARN << "Connection equals: differ! " << ex.what();
     return false;
   }
   return true;

src/htm/algorithms/Connections.hpp

@@ -95,7 +95,7 @@ struct SynapseData: public Serializable {
     NTA_CHECK(id == o.id ) << "SynapseData equals: id";
     } catch(const htm::Exception& ex) {
       UNUSED(ex);    // this avoids the warning if ex is not used.
-      //NTA_WARN << "SynapseData equals: " << ex.what(); //Note: uncomment for debug, tells you 
+      NTA_WARN << "SynapseData equals: " << ex.what(); //Note: uncomment for debug, tells you 
       //where the diff is. It's perfectly OK for the "exception" to occur, as it just denotes
       //that the data is NOT equal.
       return false;
@@ -155,7 +155,7 @@ struct SegmentData: public Serializable {
 
     } catch(const htm::Exception& ex) {
       UNUSED(ex);    // this avoids the warning if ex is not used.
-      //NTA_WARN << "SegmentData equals: " << ex.what();
+      NTA_WARN << "SegmentData equals: " << ex.what();
       return false;
     }
     return true;
@@ -195,7 +195,7 @@ struct CellData : public Serializable {
       NTA_CHECK( segments == o.segments ) << "CellData equals: segments";
     } catch(const htm::Exception& ex) {
       UNUSED(ex);    // this avoids the warning if ex is not used.
-      //NTA_WARN << "CellData equals: " << ex.what();
+      NTA_WARN << "CellData equals: " << ex.what();
       return false;
     }
     return true;
@@ -464,7 +464,7 @@ class Connections : public Serializable
    * @retval Synapse data.
    */
   inline const SynapseData& dataForSynapse(const Synapse synapse) const {
-    NTA_CHECK(synapseExists_(synapse, true));
+    NTA_CHECK(synapseExists_(synapse, false));
     return synapses_[synapse];
   }