# US 11,113,600 B2 · Translating sensor input into expertise > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 11,113,600 B2 - Document ID: US11113600B2 - Status: Granted - Kind: B2 - Application number: 15/982,665 - Filed: 2018-05-17 - Published: 2018-11-22 - Granted: 2021-09-07 - Earliest claimed: 2017-05-17 - Assignees: BSquare Corporation - Applicants: Bsquare Corp. - Inventors: David Wagstaff - Web page: https://patentgrep.com/patent/US11113600B2 ## CPC - G06N 3/08: Learning methods - G05B 13/0265: {the criterion being a learning criterion} - G06N 3/084: Backpropagation, e.g. using gradient descent - G06N 3/09: Supervised learning - G06N 5/04: Inference or reasoning models - G06N 20/00: Machine learning - G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS - G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G) ## Abstract A method including receiving input data; selecting a classification scheme; transforming the input data into transformation data utilizing the classification scheme; transforming the input data into machine learner outputs; comparing the transformation data to the machine learner outputs; and altering machine state of one or more machines in response to comparing the transformation data to the machine learner outputs. Further, a method including receiving one or more sensor inputs; receiving one or more machine insights, the one or more machine insights comprising one or more states; selecting one of the one or more states; determining conditions of the one of the one or more states; comparing the conditions to the one or more sensor inputs; and altering a machine state of one or more machines in response to comparing the conditions to the one or more sensor inputs. ## Claims ### Claim 1 A method comprising: receiving sensor input data from a plurality of sensors and machine insights, the machine insights comprising one or more states, with a trainer, the trainer generating a selection signal; operating a selector with the selection signal to select a classification scheme, the classification scheme specifying a classifier; transforming the machine insights and the sensor input data into training data utilizing the classifier; transforming the sensor input data into a class prediction with the classifier; applying the machine insights and the sensor input data to a machine learner to generate a machine learner prediction; and comparing the class prediction and the machine learner prediction with a tester and altering a machine state of one or more machines if the difference between the class prediction and the machine learner prediction is less than a threshold value and updating the machine learner if the difference between the class prediction and the machine learner prediction is not less than the threshold value. ### Claim 2 (depends on 1) The method of claim 1, wherein the one or more machines is the trainer. ### Claim 3 (depends on 2) The method of claim 2, wherein the trainer's altered machine state prompts the generation of an update selection signal to operate the selector to select a different classification scheme. ### Claim 4 (depends on 2) The method of claim 2, wherein the trainer's altered machine state prompts the generation of an update signal to update the machine learner. ### Claim 5 (depends on 1) The method of claim 1, wherein the machine learner further comprises a neural network. ### Claim 6 (depends on 1) The method of claim 1, wherein the one or more machines is a message transmitter, the message transmitter sending a message comprising one of the one or more states as a current state. ### Claim 7 (depends on 1) The method of claim 1, wherein the trainer transmits a correlation to the one or more machines if the trainer determines that the machine learner outputs correlate to the machine insights. ### Claim 8 (depends on 1) The method of claim 1, wherein: the trainer selects one of the one or more states; determines conditions of the one or more states; compares the conditions to the sensor input data; and alters the machine state of the one or more machines in response to comparing the conditions to the sensor input data. ### Claim 9 (depends on 1) The method of claim 1, wherein the sensor input data further comprises at least one of data signals, audio data signals, vibration data signals, and electromagnetic data signals and combinations thereof. ### Claim 10 A computing apparatus, the computing apparatus comprising: a processor; and a memory storing instructions that, when executed by the processor, configure the apparatus to: receive sensor input data from a plurality of sensors and machine insights, the machine insights comprising one or more states, with a trainer, the trainer generating a selection signal; operate a selector with the selection signal to select a classification scheme, the classification scheme specifying a classifier; transform the machine insights and the sensor input data into training data utilizing the classifier; transform the sensor input data into a class prediction with the classifier; apply the machine insights and the sensor input data to a machine learner to generate a machine learner prediction; and compare the class prediction and the machine learner prediction with a tester and altering a machine state of one or more machines if the difference between the class prediction and the machine learner prediction is less than a threshold value and updating the machine learner if the difference between the class prediction and the machine learner prediction is not less than the threshold value. ### Claim 11 (depends on 10) The computing apparatus of claim 10, wherein the one or more machines is the trainer. ### Claim 12 (depends on 11) The computing apparatus of claim 11, wherein the trainer's altered machine state prompts the generation of an update selection signal to operate the selector to select a different classification scheme. ### Claim 13 (depends on 11) The computing apparatus of claim 11, wherein the trainer's altered machine state prompts the generation of an update signal to update the machine learner. ### Claim 14 (depends on 10) The computing apparatus of claim 10, wherein the machine learner further comprises a neural network. ### Claim 15 (depends on 10) The computing apparatus of claim 10, wherein the one or more machines is a message transmitter, the message transmitter sending a message comprising one of the one or more states as a current state. ### Claim 16 (depends on 10) The computing apparatus of claim 10, wherein the trainer transmits a correlation to the one or more machines if the trainer determines that the machine learner outputs correlate to the machine insights. ### Claim 17 (depends on 10) The computing apparatus of claim 10, wherein the trainer: selects one of the one or more states; determines conditions of the one or more states; compares the conditions to the sensor input data; and alters the machine state of the one or more machines in response to comparing the conditions to the sensor input data. ### Claim 18 (depends on 10) The computing apparatus of claim 10, wherein the sensor input data further comprises at least one of data signals, audio data signals, vibration data signals, and electromagnetic data signals and combinations thereof.