The Semantic DB Project write up.

Knowledge representation and a language to reason with that knowledge.

The central idea is to represent knowledge, and networks, in the form of: "OP KET => SUPERPOSITION", combined with a compact language to reason with that knowledge. This is the so called BKO scheme (Bra-Ket-Operator), and for the project as a whole I call it the Feynman Knowledge Engine (a minimalist, but extensible, knowledge engine).

Here are some example learn rules in this notation:
the-friends-of |Fred> => |Sam> + |Rob> + |Emma> + |Bella> + |George>
the-age-of |Fred> => |age: 33>
the-list-of |week days> => |Monday> + |Tuesday> + |Wednesday> + |Thursday> + |Friday>
the-parents-of |*> #=> the-mother-of |_self> + the-father-of |_self>
is-a-teenager |person: *> #=> is-in-range[13,19] the-age-of |_self>
links-to |WP: Autism> => 3|WP: Asperger_syndrome> + 3|WP: intellectual_disability> + 2|WP: autism_spectrum> + ...
words-1 |WP: Albedo> => 193|the> + 167|of> + 115|albedo> + 101|and> + 98|ref> + 65|a> + 59|in> + 58|is> + 57|to> + ...
the-plural-of |word: *> #=> merge-labels(|_self> + |s>)
the-plural-of |word: foot> => |word: feet>
the-plural-of |word: mouse> => |word: mice>
the-building-at |grid: 4 40> => |building: cafe>
Some code, now on github, but definitely pre-alpha!: The below is extracted from a long series of blog posts trying to explain my ideas.
I also have a big collection of sw (semantic web) files here.


  1. announcing the semantic db project
  2. defining some terms
  3. context learn and recall
  4. tim berners lee flowchart in bko
  5. some ket label conventions
  6. personality profile for a bot called bella
  7. introducing the semantic agent console
  8. linearity of operators
  9. exponentiating operators
  10. learning indirectly
  11. supported ops and apply
  12. inverse
  13. train of thought
  14. add_learn and stored_rules
  15. current weather in adelaide in sw format
  16. label descent
  17. the self ket
  18. a fictional george in sw format
  19. some general people rules
  20. the methanol molecule in sw format
  21. the merge labels function
  22. learning plurals in bko
  23. introducing pick elt
  24. random greetings in bko
  25. arithmetic
  26. factorial and fibonacci in bko
  27. diversion the tidy language underneath
  28. simple network in sw format
  29. matrices in sw format
  30. different objects identical network structure
  31. set union and intersection in bko
  32. simple finite sets and soft intersection
  33. set builder in bko
  34. algebra
  35. to base
  36. learning a grid
  37. walking our grid
  38. temperature conversion
  39. context recall vs x apply op
  40. introducing file recall
  41. kevin bacon game numbers
  42. finding common movies and actors
  43. some bigger sw examples
  44. the maths rules for the bko scheme
  45. the first bko claim
  46. announcing phase 2 function operators
  47. some built in functions
  48. introducing sigmoids
  49. a big collection of function operators
  50. foaf vs sw
  51. is teenager and is adult in bko
  52. new function pretty print a table
  53. pretty print some data about australian cities
  54. another pretty print table example
  55. sorting in the bko scheme
  56. new function sort by
  57. is early and is late in bko
  58. tweaked pretty print table code
  59. non abelian algebra
  60. new function int coeffs to word
  61. african capital cities and population in table format
  62. sparql vs bko
  63. another simple one aiming towards natural language
  64. xml vs bko
  65. new functions is greater than is greater equal than etc
  66. brief note on pretty print tables
  67. this concludes phase 2
  68. announcing phase 3 similar and find topic operators
  69. new feature memoizing rules
  70. a similarity metric
  71. a list implementation of the simm
  72. some examples of list simm in action
  73. a superposition implementation of simm
  74. the landscape function
  75. introducing similar op x
  76. some simple similar op examples
  77. simple image recognition
  78. new function such that
  79. similar movies and similar actors
  80. how many movies
  81. top level domains in sw format
  82. a table of integers and their factors
  83. mapping webpages to well defined superpositions
  84. how big are our webpage superpositions
  85. making the ket count differences clearer
  86. creating average webpage superpositions
  87. website similarity matrices
  88. the main event pattern recognition of websites
  89. the second bko claim
  90. supervised pattern recognition
  91. the general supervised pattern recognition algo
  92. supervised learning of iris classes
  93. fixed supervised learning of iris classes
  94. harder supervised pattern recognition example
  95. some wage prediction results
  96. document types to superpositions
  97. spike fourier transform using simm
  98. introducing active buffer
  99. using active buffer
  100. introducing bridging sets
  101. an implementation of categorize code
  102. a bko version of categorize
  103. some categorize examples
  104. the normed frequency class equation
  105. the map to topic and find topic functions
  106. find topic names
  107. new function find unique op
  108. find unique op applied to webpage superpositions
  109. mapping sw files to frequency lists
  110. mapping wikipedia pages to frequency lists
  111. new function intn find topic op
  112. similarity matrices for wikipedia word frequency lists
  113. mapping mindpixels to bko rules
  114. this concludes phase 3
  115. new function subset
  116. new function list kets
  117. new function full exp op n
  118. new function apply weights n1 n2
  119. the full wage prediction results
  120. announcing phase 4 tying it all together
  121. announcing command line file similarity tool
  122. an interesting little observation
  123. new function average categorize
  124. average categorize h i simple pattern recognition example
  125. average categorize websites
  126. comment on ket independence
  127. announcing command line guess file name tool
  128. new function find inverse op
  129. towards processing all of wikipedia
  130. how many wikipage links
  131. what do we know about bananas
  132. more inverse simm results
  133. even more inverse simm results
  134. wikipedia fragment to frequency list
  135. non linear resonance
  136. some more similar inverse links to results
  137. she is out of my league
  138. introducing function matrices
  139. introducing the matsumsig model
  140. simple inhibitory signals in the matsumsig model
  141. simple logic in the matsumsig model
  142. set union and intersection in the matsumsig model
  143. simm in the matsumsig model
  144. difference and smooth in the matsumsig model
  145. simple prolog vs bko example
  146. ebook letter frequencies
  147. on emerging patterns
  148. brief object orientated vs bko example
  149. finding the transpose of a table
  150. working towards natural language
  151. introducing the ngram stitch
  152. some rambler examples
  153. some letter rambler examples
  154. letter 3 grams that precede a full stop
  155. start and end chars for 3grams that precede a full stop
  156. new console feature web load
  157. representing song lyrics in sw format
  158. visualizing superpositions
  159. new function hash
  160. new function process reaction
  161. representing nuclear decay
  162. new functions union op and top k
  163. shopping with process reacion
  164. simple particle entanglement example
  165. new function operator inhibition
  166. an interpretation of my bko scheme
  167. revisiting the letter rambler
  168. revisiting wikipedia inverse links to semantic similarities
  169. visualizing sw files
  170. update on visualizing sw files
  171. a brief look at sum of prime factors
  172. the easy way to make a big binary tree
  173. maybe we dont need projections
  174. one way to handle uris in bko
  175. semantic networks in bko
  176. spreading activation in bko
  177. softmax and log
  178. edit distance in bko
  179. visualizing edit distance
  180. how well do you know x
  181. is bko related to category theory
  182. ket arithmetic
  183. new operators times by divide by plus minus
  184. new feature list 2 sp
  185. new operators mod k and is mod k
  186. introducing the if then machine
  187. learning a sequence using if then machines
  188. learning days of the week using if then machines
  189. learning simple images using if then machines
  190. new operators guess ket and guess operator
  191. towards a definition of intelligence
  192. image histogram similarity
  193. new tool edge enhance
  194. introducing network k similarity
  195. average categorize some mnist digits
  196. new tool phi transform
  197. phi transform of lenna
  198. averaging out noise
  199. new operator simm add
  200. introducing a new tool wikivec and wikivec similarity
  201. new operators append column and random column
  202. learning sequences htm style using if then machines
  203. more htm sequence learning
  204. random encode similarity matrices
  205. simple if then machine classifier
  206. naming htm sequences
  207. visualizing htm sequences
  208. p pattern similarity metric in latex
  209. p pattern similarity metric in python
  210. normalizing mnist digits
  211. full mnist results
  212. predicting cat vs dog based on their noises
  213. identifying a face
  214. new function operator bar chart
  215. new operator words to list
  216. spike wave similarity
  217. smoothed spike wave similarity
  218. learning how to spell
  219. learning and recalling chunked sequences
  220. learning and recalling a simple sentence
  221. generating random grammatically correct sentences
  222. predicting sequences
  223. learning and recalling a sequence of frames
created: 30/6/2015
updated: 19/12/2016
by Garry Morrison
email: garry -at-