EP0157080A2
Expired / Lapsed
EP
⏱️ 0.0 years remaining
Probabilistic learning element
Assignee: International Standard Electric Corp
•
Filing: 1985-01-15
•
Granted: —
•
Expires: 2005-01-15
Abstract
Abstract A probabilistic learning element for performing task independent sequential pattern recognition. Said element receives sequences of objects and outputs sequences of recognized states composed of objects. A plurality of memory means (80, 152, 154, 156) are utilized to store the received objects in sequence and for storing in context learned information including previously learned states, objects contained in previously learned states, positional information for each object in a learned state and other predetermined types of knowledge relating to the previously learned states and the objects contained therein. The element correlates the sequences of received objects with the learned information relating to previously learned states for providing conditional probabilities to possible sequences of recognized states. The most likely state sequence is determined and outputted as a recognized sequence when the element detects that a state has ended. The memory for storing learned information is a context organized memory including a plurality of tree structures having various types of information stored in the nodes thereof with certain of said tree structures including at each node an attribute list referring to other tree structures whereby searching is facilitated and unnecessary searching eliminated. The element derives support coefficients relating to how much information was available when calculating the conditional probabilities and said support coefficients and conditional probabilities are combined to provide a rating of confidence. When the rating of confidence exceeds a predetermined level, the element is caused to store the outputted recognized state sequence as a learned state sequence with said memories storing the various types of knowledge relating to said learned sequence of states.
Inventors
- 👤 Jeffrey Neil Denenberg
- 👤 Allen Richard Smith
- 👤 Thomas Burgoyne Slack
- 👤 Chuan-Chieh Tan
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-10-26 19:15
CPC Classifications
G; G06; G06N; G06N20/00; G06F; G06F18/00; G06F18/20; G06F18/21; G06V; G06V10/00; G06V10/70
Claims Analysis
Total Claims: 31
Claim 1
Independent Claim
255 words
Probabilistic learning element, of the type that receives sequences of objects and outputs sequences of recognized states composed of objects, characterized in that it comprises: means (28) for receiving objects; means (80) for storing said received objects; means (152, 154, 156) for storing learned information including sequences of previously learned states, previously learned states, objects contained in previously learned states, sequences of objects contained in previously learned states, positional information for each object in a learned state and predetermined types of knowledge relating to sequences of previously learned states, previously learned states, objects contained in previously learned states, sequences of objects contained in previously learned states and positional information for each object in a learned state whereby information relating to received objects is stored as well as information relating to previously learned states and the objects contained therein; means (32) for correlating the received object information with the learned state information for assigning conditional probabilities to possible sequences of recognized states or to possible next states in the sequence of recognized states; means (184) responsive to said conditional probabilities for determining the most likely next state or state sequence; means (102) responsive to the stored information for determining the probability of the end of a state or for detecting that a state has ended and for providing an end of state signal; and means (34) responsive to the end-of-state signal or to the end of state probability to output the most likely next state or state sequence as a recognized next state or state sequence.
Claim 2
Dependent Claim
255 words
Probabilistic learning element, of the type that receives sequences of objects and outputs sequences of recognized states composed of objects, characterized in that it comprises: means (28) for receiving objects; means (80) for storing said received objects; means (152, 154, 156) for storing learned information including sequences of previously learned states, previously learned states, objects contained in previously learned states, sequences of objects contained in previously learned states, positional information for each object in a learned state and predetermined types of knowledge relating to sequences of previously learned states, previously learned states, objects contained in previously learned states, sequences of objects contained in previously learned states and positional information for each object in a learned state whereby information relating to received objects is stored as well as information relating to previously learned states and the objects contained therein; means (32) for correlating the received object information with the learned state information for assigning conditional probabilities to possible sequences of recognized states or to possible next states in the sequence of recognized states; means (184) responsive to said conditional probabilities for determining the most likely next state or state sequence; means (102) responsive to the stored information for determining the probability of the end of a state or for detecting that a state has ended and for providing an end of state signal; and means (34) responsive to the end-of-state signal or to the end of state probability to output the most likely next state or state sequence as a recognized next state or state sequence.
Claim 3
Dependent Claim
30 words
A probabilistic learning element as claimed in claim 1, wherein the positional information stored for each object includes the object's distance to begin and distance to end of a state.
Claim 4
Dependent Claim
61 words
A probabilistic learning element as claimed in claim 2, wherein the predetermined types of knowledge stored include the numbers of occurrences of: each object, sequences of objects, states and sequences of states, each of the sequences of learned states, learned states, objects contained in learned states, sequences of objects contained in learned states and the positions of objects in learned states.
Claim 5
Dependent Claim
69 words
A probabilistic learning element as claimed in claim 3, wherein the predetermined types of knowledge stored include the length of each learned state, the number of occurences of each length, sequences of state lengths, the number of occurrences of each sequence of state lengths, state-length pairs, the number of occurrences of each state-length pair, sequences of state-length pairs and the number of occurrences of each sequence of state-length pairs.
Claim 6
Dependent Claim
70 words
A probabilistic learning element as claimed in any one of the preceding claims, wherein the means (80, 152, 154, 156) for storing is adapted to store the information in accordance with the context in which the stored information statistically occurred, whereby from any stored information the stored information which statistically occurs next in context is directly accessible and the conditional probabilities may be easily derived from the contextually stored information.
Claim 7
Dependent Claim
157 words
A probabilistic learning element as claimed in claim 5 wherein the means for storing said received objects is a short term memory means (80) for storing in sequential context said received objects; wherein the means for storing said learned information is a long term memory means (152, 154, 156) for storing in sequential context previously learned states and the objects contained therein, and predetermined types of knowledge relating to the stored objects and states, whereby from any stored information the stored information which occurs next in context is directly accessible and wherein in said means (32) for correlating the received object information with the learned state information, said correlation is facilitated by using the context of the information stored in the short term memory means as a pointer to the context of the information stored in the long term memory means, said correlating means (32) assigning probabilities to possible next states in the sequence of recognized states.
Claim 8
Dependent Claim
117 words
A probabilistic learning element as claimed in any one of the preceding claims, additionally comprising: means (34) for providing a rating of confidence in said recognized next state. means for accumulating the ratings of confidence of recognized states of the recognized state sequence; and means responsive to said accumulated ratings of confidence to cause said means for storing to store the recognized states, state sequence, the objects, sequences of objects and states forming the recognized state sequence and the predetermined types of knowledge relating to the objects, sequences of objects, states and sequences of states forming said recognized state sequence as statistical information relating to previously learned states when the accumulated ratings exceed a predetermined threshold level.
Claim 9
Dependent Claim
56 words
A probabilistic learning element as claimed in claim 6, additionally comprising learning supervision means responsive to external reinforcement signals to cause said long term memory means to store the recognized states and the objects contained therein and the predetermined types of knowledge relating to the stored objects and states as information relating to previously learned states.
Claim 10
Dependent Claim
134 words
A probabilistic learning element as claimed in claim 6, additionally comprising: means (34) for providing a rating of confidence in said recognized next state; learning supervision means adapted to receive said rating of confidence and an external reinforcement signal, said means being responsive to accumulated ratings of confidence of the recognized states of the recognized state sequence and the external reinforcement signal for providing an output signal when either the accumulated ratings of confidence exceed a predetermined threshold level or an external reinforcement signal is received; and means responsive to the output signal from the learning supervision means to cause said mens for storing to store the recognized states and the objects contained therein and the predetermined types of knowledge relating to the stored objects and states as information relating to previously learned states.
Claim 11
Dependent Claim
60 words
A probabilistic learning element as claimed in any one of the preceding claims, wherein the means for storing (152, 154, 156) stores information on a plurality of previously learned states and the objects contained therein, whereby several levels of prior context are stored and the means for correlating implements an nthe order Markov process correlating several levels of stored context.
Claim 12
Dependent Claim
48 words
A probabilistic learning element as claimed in claim 6, wherein the long term memory means comprises a context organized memory including a plurality of tree structures for storing previously learned states and the objects contained therein and predetermined types of knowledge relating to the stored objects and states.
Claim 13
Dependent Claim
44 words
A probabilistic learning element as claimed in claim 11, wherein a tree structure used to store objects of learned states includes at each node thereof an object that is a part of a state whereby the tree includes sequences of objects shown in context.
Claim 14
Dependent Claim
32 words
A probabilistic learning element as claimed in claim 12, wherein each node of the tree has stored therein the number of occurrences of the object in the context represented by the node.
Claim 15
Dependent Claim
32 words
A probabilistic learning element as claimed in claim 11, wherein a separate tree structure is provided for storing in context learned states, objects contained in learned states, state lengths and state-length pairs.
Claim 16
Dependent Claim
122 words
A probabilistic learning element as claimed in claim 6, wherein the long term memory means comprises a context organized memory including a plurality of tree structures for storing the various types of stored information, each tree includes a plurality of nodes with each node having a record of the additional stored item represented by the node, the occurrences of the stored item in the context represented by the node, a pointer to a node which represents the same sequence of items without the last item, a pointer to a node which represents the same sequence of items and with one additional item at each end and a pointer to the node which represents the same sequence of items without the first item.
Claim 17
Dependent Claim
195 words
A probabilistic learning element as claimed in claim 5, wherein the means for storing said received objects is a short term memory means (80) for storing in sequential context said received objects; wherein the means for storing said learned information is a context organized memory means including a plurality of tree structures, including an alltree structure and a plurality of singletree structures, the alltree structure stores the contextual occurrences of all objects received by the probabilistic learning element and at each node of the alltree there is provided an attribute list which refers to singletrees that include the same object context as the node of the alltree, a singletree is provided for each previously learned state, whereby searching is facilitated by using the alltree as a pointer to the less complex singletrees, and wherein in said means (32) for correlating the received object information with the learned state, said correlation is facilitated by use of the context of information stored in the short term memory means as a pointer to the context of the statistically stored information in the context organized memory means, said correlating means assigning conditional probabilities to possible sequences of recognized states.
Claim 18
Dependent Claim
55 words
A probabilistic learning element as claimed in claim 16, wherein each singletree contains object information for a previously learned state with each node representing an object and having an attribute list including the distance to begin, distance to end and the number of times that the object appeared at that particular position within a state.
Claim 19
Dependent Claim
95 words
A probabilistic learning element as claimed in any one of claims 1 to 4 and 16 to 17, wherein the means for correlating comprises: a first means for determining the probabilities that possible states will span an object sequence having a particular begin time and end time and second means for determining the probabilities of state-length pairs given the previous state-length pair context, means responsive to the previously mentioned probabilities to implement an algorithm to provide probabilities of possible states, with a particular length that span an object sequence given the previous state-length pair context.
Claim 20
Dependent Claim
67 words
A probabilistic learning element as claimed in claim 18, wherein the first means is responsive to two types of conditional probability signals, a first type signal corresponding to the conditional probabilities of object sequences occurring within a state given the state and a second type signal corresponding to the conditional probabilities of states with a particular begin time given an end time, object sequence and a state.
Claim 21
Dependent Claim
47 words
A probabilistic learning element as claimed in claim 19, wherein the first type probability signal is derived from the conditional probabilities of an object occurring given the previous object context and state which probabilities are calculated from the stored learned information relating to objects and object occurrences.
Claim 22
Dependent Claim
61 words
A probabilistic learning element as claimed in claim 19 or 20, wherein the second type probability signal is derived from conditional probabilities of an object occurring in a particular position in a state given the previous object context its position and state which probabilities are derived from the stored learned information relating to objects, object occurrences and the object positional information.
Claim 23
Dependent Claim
78 words
A probabilistic learning element as claimed in claim 18, wherein the second means for determining the conditional probabilities of state-length pairs is responsive to the stored learned information relating to previously learned states and their occurrences, lengths of previously learned states and their occurrences, state-length pairs form previously learned states and their occurrences, sequences of previously learned states and their occurrences, sequences of lengths of previously learned states and their occurrences, sequences of state-length pairs and their occurrences.
Claim 24
Dependent Claim
33 words
A probabilistic learning element as claimed in any one of claims 1 to 5 and 16 to 22, additionally comprising means for providing a rating of confidence in said sequence of recognized states.
Claim 25
Dependent Claim
67 words
A probabilistic learning element as claimed in claim 23, wherein said means for providing a rating of confidence in said recognized state sequence includes means for deriving support coefficients relating to how much information was available when calculating the conditional probabilities, said confidence rating being a function of the conditional probabilities and the support coefficients for the conditional probabilities used to determine the most likely state sequence.
Claim 26
Dependent Claim
82 words
A probabilistic learning element as claimed in claim 23 or 24, additionally comprising means responsive to said rating of confidence to cause said means for storing to store the recognized state sequence, the objects, sequences of objects and states forming the recognized state sequence and the predetermined types of knowledge relating to the objects, sequences of objects, states and sequences of states forming said recognized state sequence as information relating to previously learned states when the rating exceeds a predetermined threshold level.
Claim 27
Dependent Claim
81 words
A probabilistic learning element as claimed in claim 23 or 24, additionally comprising means responsive to said rating of confidence to cause said context organized memory to store the objects and states forming the recognized state sequence, the lengths and state-length pairs of said states and the predetermined types of knowledge relating to the objects, states, state lengths and state-length pairs from said recognized state sequence as information relating to previously learned states when the rating exceeds a predetermined threshold level.
Claim 28
Dependent Claim
86 words
A probabilistic learning element as claimed in any one of claims 16 to 22 without reference being had to claims 1 to 5, additionally comprising learning supervision means responsive to external reinforcement signals to cause said context organized memory to store the objects and states forming the recognized state sequence the lengths and state-lengths pairs of said states and the predetermined types of knowledge relating to the objects, states, state lengths and state-length pairs from the recognized state sequence as information relating to previously learned states.
Claim 29
Dependent Claim
82 words
A probabilistic learning element as claimed in any one of claims 1 to 5 and 16 to 22, additionally comprising learning supervision means responsive to external reinforcement signals to cause said means for storing to store the recognized state sequence, the objects, sequences of objects and states forming the recognized state sequence and the predetermined types of knowledge relating to the objects, sequences of objects, states and sequences of states forming the recognized state sequence as information relating to previously learned states.
Claim 30
Dependent Claim
157 words
A probabilistic learning element as claimed in any one of claims 1 to 5 and 16 to 22, additionally comprising; means for providing a rating of confidence in said sequence of recognized states; learning supervision means adapted to receive said rating of confidence and an external reinforcement signal, said means being responsive to the rating of confidence of the recognized state sequence and the external reinforcement signal for providing an output signal when either the rating of confidence exceed a predetermined threshold level or an external reinforcement signal is received; and means responsive to the output signal from the learning supervision means to cause said means for storing to store the recognized state sequence, the objects, sequences of objects, and states forming the recognized state sequence and the predetermined types of knowledge relating to the objects, sequences of objects, states and sequences of states forming said sequence of recognized states as information relating to previously learned states.
Claim 31
Dependent Claim
32 words
A probabilistic learning element as claimed in any one of claims 8, 9, 27, 28, 29, additionally comprising means for correcting a recognized state sequence prior to initiating an external reinforcement signal.
References Cited (Backward Citations)
No citation records indexed.
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.