EP3807822A1 - Neuromorphic system for authorized user detection - Google Patents
Neuromorphic system for authorized user detectionInfo
- Publication number
- EP3807822A1 EP3807822A1 EP19820220.2A EP19820220A EP3807822A1 EP 3807822 A1 EP3807822 A1 EP 3807822A1 EP 19820220 A EP19820220 A EP 19820220A EP 3807822 A1 EP3807822 A1 EP 3807822A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- neuromorphic
- spikes
- sensor data
- rates
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/70—Protecting specific internal or peripheral components, in which the protection of a component leads to protection of the entire computer
- G06F21/88—Detecting or preventing theft or loss
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/30—Authentication, i.e. establishing the identity or authorisation of security principals
- G06F21/31—User authentication
- G06F21/316—User authentication by observing the pattern of computer usage, e.g. typical user behaviour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/06—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
- G06N3/063—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using electronic means
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Definitions
- the present invention relates to an anomaly detection system and, more
- Anomaly detection systems are often employed to classify sensor data or otherwise identify a change in system dynamics that can be designated as anomalous behavior.
- Conventional automated detection systems operate on top of conventional electronics and, as such, use a significant amount of energy to solve sophisticated classification problems.
- Such systems traditionally operate on large scale processors or otherwise relatively large computing systems.
- the implementation of anomaly detection in a variety of mobile and field applications requires a small- sized system that can efficiently operate on complex problems.
- detection or classification systems have not been incorporated into small scale chips that can operate on complex problems with little computational overhead.
- This disclosure is directed to a neuromorphic system for authorized user
- the system includes a neuromorphic electronic component for embedding in or attached to a client device.
- the neuromorphic electronic component having a neuromorphic chip operable for continuously monitoring streaming sensor data from a client device and generating out-spikes based on the streaming sensor data.
- the neuromorphic system further includes a client device comprising an input processing component, an output processing component, and a plurality of sensor types for providing the streaming sensor data.
- the output processing component classifies the streaming sensor data based on the out-spikes to detect a user-transition.
- the input processing component is configured to further perform operations of:
- the neuromorphic electronics component generates the out- spikes based on the in-spikes.
- the neuromorphic electronics component generates the out- spikes based on the in-spikes using a randomly connected excitatory-inhibitory spiking network.
- the output processing component further performs
- the output processing component further performs at least one operation of:
- the present invention also includes a computer program product and a computer implemented method.
- the computer program product includes computer-readable instructions stored on a non-transitory computer-readable medium that are executable by a computer having one or more processors, such that upon execution of the instructions, the one or more processors perform the operations listed herein.
- the computer implemented method includes an act of causing a computer to execute such instructions and perform the resulting operations.
- FIG. 1 is an illustration of a neuromorphic system for authorized user
- FIG. 2 is a block diagram depicting the components of input and output
- FIG. 3 is an illustration of a computer program product according to some embodiments of the present disclosure.
- FIG. 4 provides a flowchart depicting processing flow within the input
- FIG. 5 is a graphical representation of paths that an electronic signal may travel through a neuromorphic electronics component according to some embodiments of the present disclosure
- FIG. 6 illustrates a process flow for the output processing component
- FIG. 7 is a flowchart illustrating an example of a tiered threat detection system in which the neuromorphic system of the present disclosure was implemented
- FIG. 8 shows a schematic of an example mobile device and neuromorphic electronics in which the neuromorphic detection system of the present disclosure was implemented
- FIG. 9 is an illustration of an example mobile device in which the
- FIG. 10 is a graph illustrating a 5-user exponentially smoothed readout signal from the neuromorphic platform, depicting detected user-transitions.
- the present invention relates to an anomaly detection system and, more
- any element in a claim that does not explicitly state“means for” performing a specified function, or“step for” performing a specific function, is not to be interpreted as a“means” or“step” clause as specified in 35 U.S.C.
- the first is a low-power neuromorphic system for authorized user detection.
- the system has three general components, an input processing component 102, a neuromorphic electronics component 104, and an output processing component 106.
- Both the input processing 102 and the output processing 106 components are implemented in software and/or hardware as a computer system that is owned by or otherwise within a“client” system (such as the memory and processing components, etc., with a mobile device, vehicle, etc.).
- the neuromorphic electronics component 104 is the component that is
- the second principal aspect is a method, typically in the form of software or other programming, operated using a data processing system (computer) and the neuromorphic hardware described herein.
- the third principal aspect is a computer program product.
- the computer program product generally represents computer-readable instructions stored on a non-transitory computer-readable medium such as an optical storage device, e.g., a compact disc (CD) or digital versatile disc (DVD), field programmable gate array (FPGA), or a magnetic storage device such as a floppy disk or magnetic tape.
- a non-transitory computer-readable medium such as an optical storage device, e.g., a compact disc (CD) or digital versatile disc (DVD), field programmable gate array (FPGA), or a magnetic storage device such as a floppy disk or magnetic tape.
- Other, non limiting examples of computer-readable media include hard disks, read-only memory (ROM), and flash-type memories.
- FIG. 2 provides a block diagram depicting a non-limiting example of a computer system 200 that can be implemented to operate as the input and/or output processing components (i.e., elements 102 and/or 106 of FIG. 1).
- a computer system 200 is configured to perform calculations, processes, operations, and/or functions associated with a program or algorithm.
- certain processes and steps discussed herein are realized as a series of instructions (e.g., software program) that reside within computer readable memory units and are executed by one or more processors of the computer system 200. When executed, the instructions cause the computer system 200 to perform specific actions and exhibit specific behavior as may be required to perform the processes described herein.
- the computer system 200 may include an address/data bus 202 that is
- processor 204 configured to communicate information. Additionally, one or more data processing units, such as a processor 204 (or processors), are coupled with the address/data bus 202.
- the processor 204 is configured to process information and instructions.
- the processor 204 is a microprocessor.
- the processor 204 may be a different type of processor such as a parallel processor, application-specific integrated circuit (ASIC), programmable logic array (PLA), complex programmable logic device (CPLD), or a field
- FPGA programmable gate array
- the computer system 200 is configured to utilize one or more data storage units.
- the computer system 200 may include a volatile memory unit 206 (e.g., random access memory (“RAM”), static RAM, dynamic RAM, etc.) coupled with the address/data bus 202, wherein a volatile memory unit 206 is configured to store information and instructions for the processor 204.
- the computer system 200 further may include a non-volatile memory unit 208 (e.g., read-only memory (“ROM”), programmable ROM (“PROM”), erasable programmable ROM
- the computer system 200 may execute instructions retrieved from an online data storage unit such as in“Cloud” computing.
- the computer system 200 also may include one or more interfaces, such as an interface 210, coupled with the address/data bus 202, or other interfaces as described in further detail below (e.g., digital interface to the neuromorphic electronics). The one or more interfaces are configured to enable the computer system 200 to interface with other electronic devices and computer systems.
- the communication interfaces implemented by the one or more interfaces may include wireline (e.g., serial cables, modems, network adaptors, etc.) and/or wireless (e.g., wireless modems, wireless network adaptors, etc.) communication technology.
- wireline e.g., serial cables, modems, network adaptors, etc.
- wireless e.g., wireless modems, wireless network adaptors, etc.
- the computer system 200 may include an input device 212
- the input device 212 is coupled with the address/data bus 202, wherein the input device 212 is configured to communicate information and command selections to the processor 200.
- the input device 212 is an alphanumeric input device, such as a keyboard, that may include alphanumeric and/or function keys.
- the input device 212 may be an input device other than an alphanumeric input device.
- the computer system 200 may include a cursor control device 214 coupled with the address/data bus 202, wherein the cursor control device 214 is configured to communicate user input information and/or command selections to the processor 200.
- the cursor control device 214 is implemented using a device such as a mouse, a track-ball, a track pad, an optical tracking device, or a touch screen.
- the cursor control device 214 is directed and/or activated via input from the input device 212, such as in response to the use of special keys and key sequence commands associated with the input device 212.
- the cursor control device 214 is configured to be directed or guided by voice commands.
- the computer system 200 further may include one or more
- a storage device 216 coupled with the address/data bus 202.
- the storage device 216 is configured to store information and/or computer executable instructions.
- the storage device 216 is a storage device such as a magnetic or optical disk drive (e g., hard disk drive (“HDD”), floppy diskette, compact disk read only memory
- HDD hard disk drive
- floppy diskette compact disk read only memory
- a display device 218 is coupled with the address/data bus 202, wherein the display device 218 is configured to display video and/or graphics.
- the display device 218 may include a cathode ray tube (“CRT”), liquid crystal display (“LCD”), field emission display (“FED”), plasma display, or any other display device suitable for displaying video and/or graphic images and alphanumeric characters recognizable to a user.
- CTR cathode ray tube
- LCD liquid crystal display
- FED field emission display
- plasma display or any other display device suitable for displaying video and/or graphic images and alphanumeric characters recognizable to a user.
- the computer system 200 presented herein is an example computing
- the non-limiting example of the computer system 200 is not strictly limited to being a computer system.
- the computer system 200 represents a type of data processing analysis that may be used in accordance with various aspects described herein.
- other computing systems may also be implemented.
- the spirit and scope of the present technology is not limited to any single data processing environment.
- one or more operations of various aspects of the present technology are controlled or implemented using computer-executable instructions, such as program modules, being executed by a computer.
- program modules include routines, programs, objects, components and/or data structures that are configured to perform particular tasks or implement particular abstract data types.
- an aspect provides that one or more aspects of the present technology are implemented by utilizing one or more distributed computing environments, such as where tasks are performed by remote processing devices that are linked through a communications network, or such as where various program modules are located in both local and remote computer- storage media including memory- storage devices.
- FIG. 3 An illustrative diagram of a computer program product (i.e., storage device) embodying an aspect of present invention is depicted in FIG. 3.
- the computer program product is depicted as floppy disk 300 or an optical disk 302 such as a CD or DVD.
- the computer program product generally represents computer-readable instructions stored on any compatible non-transitory computer-readable medium.
- the term“instructions” as used with respect to this invention generally indicates a set of operations to be performed on a computer, and may represent pieces of a whole program or individual, separable, software modules.
- Non-limiting examples of“instruction” include computer program code (source or object code) and“hard-coded” electronics (i.e. computer operations coded into a computer chip, or in other aspects, a
- The“instruction” is stored on any non-transitory computer-readable medium, such as in the memory of a computer or on a floppy disk, a CD-ROM, FPGA and a flash drive.
- neuromorphic anomaly detection system 100 that identifies anomalies by detecting a change in classification by autonomously discovering significant differences between sensor data 108. Upon detection of such a change, the system then generates an output signal 110 as appropriate per the particular application.
- the output signal 110 can trigger any number of subsequent actions, such as computer processing tasks, transmission of information related to the anomaly or change in environment, and storage of data.
- the neuromorphic electronics 104 described herein can be implemented using any suitable neuromorphic hardware.
- the neuromorphic electronics 104 component is implemented using the neuromorphic hardware as described in U.S. Patent No. 8,977,578, the entirety of which is incorporated herein by reference.
- the low-power neuromorphic system of the present disclosure uses the neuromorphic hardware (such as that as described in the‘578 patent) together with additional data flow steps to produce a single output signal 110.
- neuromorphic hardware is tuned/configured in a specific way as to produce desired results.
- the system is unique in that it uses neuromorphic hardware (e.g., an amine, an amine
- the neuromorphic technology scales in size and performs fast enough for complex problems. Further, it uses machine learning techniques to discover differences in nominal and off-nominal conditions that can be applied practically to modern“client” systems that include electronic devices.
- the system applies a linear classifier to a liquid state machine with a coupling of electronics and software as sequenced in the figures submitted herewith. In doing so, the system offers a low power consumption processing of a complex problem that occurs on-board the“client” system and that enables the system to use less resources than the prior art.
- the implementation allows for a very low power (less than 100 Milliwatts) solution to complex classification problems with a quick response time (milliseconds to seconds depending on application) and a small size and weight footprint.
- the neuromorphic system 100 includes three serial
- the sensor input 108 includes various types of sensor data.
- sources of the sensor input 106 data include gyroscopes, accelerometers, altimeters, fuel levels, computer network traffic, etc., that feed the input processing component 102.
- both the input processing 102 and the output processing 106 components are implemented in software and run on a conventional computer processor and can be incorporated into any platform or client device that can receive streaming data.
- client devices include a mobile device (e.g., phone, ipad, etc.), an autonomous vehicle, computer, or any other platform/client device that uses a processor and receives streaming data.
- the neuromorphic electronics 104 component is the component that is
- the neuromorphic electronics 104 communicates with the other processing components 102 and 106 through a digital interface (e.g., a Serial Peripheral Interface (SPI)).
- the output processing 106 generates a binary output signal 110 which represents either a nominal status (e.g., authorized user) or an anomaly detected status (e.g., unauthorized user).
- FIG. 4 provides a flowchart depicting processing flow within the input processing 102 component.
- the input processing 102 component cleans (i.e., normalizes data 400) the sensor input 108 (i.e., streaming sensor data) and then maps the data into either the frequency domain (fd) or the time domain (td). All of the data (both td and fd) is then converted into rates and then into in-spikes 418, as described in further detail below.
- the sensor input 108 data is normalized 400 by mapping sensor-specific
- the data is combined 402 by grouping data types into a single scalar quantity per time instance. Thereafter, a grouping of samples are collected 404 into a queue of sample size appropriate to the application. The queues are transformed 406 into discrete one-dimensional (1D) frequency domain data.
- the 1D frequency domain data is then modified 408 by multiplying the data by a window function (such as a Hamming window or any other suitable window function) to reduce spectral leakage.
- a window function such as a Hamming window or any other suitable window function
- the data is filtered 410 by rejecting frequency bins outside of a particular application frequency range, resulting in scaled windowed frequency bins.
- a non-limiting example of such an application and frequency range includes selecting frequencies between 0.5 and 10 Hz in a gait application.
- the input processing 102 component then scales 412 all values (both the normalized time series and scaled windowed frequency bins) such that values are capped at the maximum execution time to generate in-rates. For example, the values are linearly mapped onto the range of spiking rates (e.g., 0-200 Hz, etc.) The in-rates are mapped 414 to a distribution function (e.g., Poisson probability distribution function (P)). Finally, in-spikes 418 are generated 416 for each input pads of the neuromorphic electronics component 104 and its neuromorphic chip 821.
- P Poisson probability distribution function
- the in-spikes 418 are binary values for each input pad that are generated based on a randomly generated number compared with the Poisson probability distribution (P) value associated with the in-rates value.
- the in-spikes 418 are transmitted in a structure equivalent to a 1 -dimensional binary array.
- the maximum number of in-spikes 418 is equal to the number of input pads that reside on the neuromorphic chip (i.e., neuromorphic electronics component with an electronically implemented Liquid State Machine with leaky integrate and fire neurons, such as that disclosed in U.S. Patent No. 8,977,578).
- FIG. 5 provides a graphical representation of a neuromorphic electronics 104 component and the paths that an electronic signal may travel from input pads 500 to output pads 502 for a particular configuration.
- Light squares 504 represent excitatory neurons and dark squares 506 represent inhibitory neurons.
- the neuromorphic electronics 104 component receives as an input the in-spikes 418 and generates out-spikes 508 as the output pad signal.
- the out-spikes 508 are also transmitted in a structure equivalent to a 1 -dimensional binary array.
- the maximum number of out-spikes 508 is equal to the number of output pads 502 that reside on the neuromorphic chip of the neuromorphic electronics 104.
- FIG. 6 illustrates a process flow for the output processing 106 component. As shown in FIG.
- the out-spikes 508 are smoothed 600 into firing rates or out-rates 602, which are a measure of the number of spikes per second.
- the rates 602 are then applied to the linear classifier 604 to compute 606 the readouts 608.
- the linear classifier 604 acts as a mapping from rates 602 to readouts 608.
- Linear classifiers are commonly understood by those skilled in the art (see, for example, “Linear classifier” as defined by Wikipedia®, the entirety of which is
- each readout 608 is a linear combination of rates 602.
- it is a simple matrix multiplication operation, where the learning process is what determines the matrix by which the rates 602 are multiplied to get the readouts 608.
- discriminative training 603 must occur on the linear classifier 604 before a reliable anomaly signal generation can occur.
- Training 603 is based on supervised machine learning techniques using ground truth values 614. Each training session is dedicated to one class at a time (there may be 2 or more classes). The differences between the classes (time domain and/or frequency domain) is what will cause the anomaly or anomalous signal.
- the frequency domain data is calculated in real-time and used in a sliding window manner. The amount of frequency domain output data size is equal to the queue sample size.
- Readouts 608 are a number of float values that correspond to each Class (2 or more). Readouts are then filtered 610 to remove noise and perform an anomaly detection process 612, which results in the final output signal 110 (e.g., the anomalous signal which specifies that there was an authorized or unauthorized user). The anomaly detection process 612 identifies an unauthorized user or authorized user by signaling that a change in users has occurred. The readout 608 signals become anomalous when there is a user-transition, and that anomaly (in the readout 608 signal) is what is detected by the system. [00065] (4) Example Implementations [00066] As can be appreciated by those skilled in the art, the neuromorphic anomaly detection system of the present disclosure has many applications.
- the system has been applied to detecting a change in users of a mobile device using biometric sensor data.
- the system was physically attached or embedded into the mobile device where power usage of the mobile device was the focus of resource savings.
- the mobile device can lock out the user from further using the device until an authorized user is detected (at which point the features/functions of the mobile device are unlocked and accessible).
- the system can be used in other applications where power consumption, feature size and or accessibility to the device is very limited.
- the system is also beneficial where a complicated anomaly detection would be most beneficial if performed directly on the Client System (as appose to off-board server / cloud solutions).
- the low-power neuromorphic system of the present disclosure was applied in a tiered anomaly classification solution, such as in Stage 1 box 700 of the tiered threat detection algorithm 702 (shown in FIG. 7) further described in U.S. Application No. 15/338,228.
- the process as described in U.S. Application No. 15/338,228 was modified to include the low- power neuromorphic system of the present disclosure as the Stage 1 box 700.
- the sensor data was generated from gyroscopes and accelerometers at 50 Hz.
- the queue size was 200 samples.
- the application frequency range of interest was from 0.3 Hz to 20.0Hz. Rates were capped at 200 Hz.
- Execution times ranged from 1 5ms to 5ms. About 25 input pads were actively used and about 50 output pads were actively used.
- the liquid state machine (LSM) was configured such that 300 excitatory neurons and 25 inhibitory neurons were activated. The network graph of neurons was connected at 1%. The system did successfully distinguish between multiple users operating the phone and a phone alarm sounded when a non- authenticated user walked with phone.
- FIG. 8 shows a schematic of an example mobile device 800 (e.g., phone) and neuromorphic electronics 104 (i.e., neuromorphic chip 821 communicating with an FPGA 823 or any other hardware or component as may be necessary to allow the neuromorphic chip 821 to operate) component.
- the FPGA 823 is responsible for neuromorphic chip 821 configuration (e.g., a non-limiting example of which includes loading a particular network, such as the one described above with respect to FIG. 5).
- Subcomponents that have been developed and that reside on the mobile device 800 include the sensor data collection app 802 and data re sampler 804 which converts all sensor streams from a non-uniform sensor data 803 to a uniform sampling rate (i.e., uniform data (binary) 805).
- the output of the re-sampler 804 goes directly to the local EWS app 806 and to the rate-based spike encoder 808 (which operate as the input processing component described above).
- a shielded cable 830 is illustrated for transmission; however, the invention is not limited to such a cable as the specific cable 830 as illustrated is provided as a non-limiting example of such a transmission medium.
- the sensor signals encoded as in-spikes 418, are sent to the neuromorphic electronics 104 component via the serial peripheral interface (SPI) connection 807 on the mobile device 800 and the corresponding line drivers and receivers 827.
- the alerts generated by the neuromorphic electronics 104 are encoded as out- spikes 518 and sent over the SPI connection 807 (and corresponding line drivers and receivers 827) to the mobile device 800, where they are decoded 809 and then read by the EWS app 806 (i.e., which operates as the output processing component described above) to determine the intent or classification, which is then broadcast 813 to an optional policy engine 815 that maintains policies as to what is acceptable intent.
- the policy engine 815 specifies that a
- the mobile device 800 such as an operation of locking all device access 819 until an authorized user is detected (e.g., an appropriate access code is input into the system by the authorized user).
- Other examples based on anomaly detection include starting a new processing task or executes a new logic branch of executable code, transmitting information associated with the anomaly, and saving information associated with the anomaly to memory storage.
- the above activities are terminated once or shortly after the signal transitions from an anomaly back to a nominal state.
- the device 800 of other features in the device can be unlocked.
- activity between devices can be correlated by an EWS 825.
- FIG. 9 illustrates an example mobile device 800 in which the neuromorphic detection system of the present disclosure was implemented, along with the included neuromorphic electronics 104 component on the interface backplate attached to the mobile device 800.
- the 5-user exponentially smoothed readout signal from the neuromorphic platform was input to the neuromorphic classifier and performed with an average 84.16% accuracy.
- the“filtered readouts” axis 1002 is a dimensionless number. They are the result of multiplying the out-rates with the linear classifier weight matrix and then applying a low-pass filter.
- The“Sample ID” axis 1004 refers to a data point index and so is also dimensionless. It should be understood that the x-axis can also be generated based on time since index and time are isomorphic in this context. [00072]
- the lowest per-user classification accuracy was 66.53%, and the highest reached 99.21%.
- the average true positive and negative rates were 89.67% and 60.32% respectively. As the chip was trained on 5 different subjects’ data, the chance classification rate is 20%.
- a consecutive alarm aggregation policy was enacted on the user-transition output and the time-series was divided into equal-length blocks from which ground-truth was determined.
- a user transition policy set a minimum margin of 8.2 seconds between detection of false alarms. Doing so prevents unwanted consecutive alarms from occurring for the same transition event.
- the total number of trials tested for each alarm was set to: (total time in sample set) / (2 * margin). If an alarm was not detected within +/- margin seconds of the truth time associated with an alarm, then the alarm was classified as a false-negative.
- the devices can be caused to perform a variety of automated actions, including ceasing operations, pulling an autonomous vehicle safely to the side of a road and turning off, locking out a user, etc.
Abstract
Description
Claims
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US201862684515P | 2018-06-13 | 2018-06-13 | |
PCT/US2019/026830 WO2019240868A1 (en) | 2018-06-13 | 2019-04-10 | Neuromorphic system for authorized user detection |
Publications (2)
Publication Number | Publication Date |
---|---|
EP3807822A1 true EP3807822A1 (en) | 2021-04-21 |
EP3807822A4 EP3807822A4 (en) | 2022-03-23 |
Family
ID=68842032
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
EP19820220.2A Pending EP3807822A4 (en) | 2018-06-13 | 2019-04-10 | Neuromorphic system for authorized user detection |
Country Status (3)
Country | Link |
---|---|
EP (1) | EP3807822A4 (en) |
CN (1) | CN112740234B (en) |
WO (1) | WO2019240868A1 (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111669514B (en) * | 2020-06-08 | 2021-02-26 | 北京大学 | High dynamic range imaging method and apparatus |
Family Cites Families (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8515885B2 (en) * | 2010-10-29 | 2013-08-20 | International Business Machines Corporation | Neuromorphic and synaptronic spiking neural network with synaptic weights learned using simulation |
US9460387B2 (en) * | 2011-09-21 | 2016-10-04 | Qualcomm Technologies Inc. | Apparatus and methods for implementing event-based updates in neuron networks |
US8977578B1 (en) * | 2012-06-27 | 2015-03-10 | Hrl Laboratories, Llc | Synaptic time multiplexing neuromorphic network that forms subsets of connections during different time slots |
BR112016002229A2 (en) * | 2013-08-09 | 2017-08-01 | Behavioral Recognition Sys Inc | cognitive neurolinguistic behavior recognition system for multisensor data fusion |
US10248675B2 (en) * | 2013-10-16 | 2019-04-02 | University Of Tennessee Research Foundation | Method and apparatus for providing real-time monitoring of an artifical neural network |
US10423879B2 (en) * | 2016-01-13 | 2019-09-24 | International Business Machines Corporation | Efficient generation of stochastic spike patterns in core-based neuromorphic systems |
US10157629B2 (en) * | 2016-02-05 | 2018-12-18 | Brainchip Inc. | Low power neuromorphic voice activation system and method |
US20170227995A1 (en) * | 2016-02-09 | 2017-08-10 | The Trustees Of Princeton University | Method and system for implicit authentication |
US10628568B2 (en) * | 2016-03-31 | 2020-04-21 | Fotonation Limited | Biometric recognition system |
US20170337469A1 (en) * | 2016-05-17 | 2017-11-23 | Agt International Gmbh | Anomaly detection using spiking neural networks |
US11003984B2 (en) * | 2016-05-31 | 2021-05-11 | Samsung Electronics Co., Ltd. | Timing sequence for digital STDP synapse and LIF neuron-based neuromorphic system |
US10671912B2 (en) * | 2016-09-13 | 2020-06-02 | Sap Se | Spatio-temporal spiking neural networks in neuromorphic hardware systems |
-
2019
- 2019-04-10 EP EP19820220.2A patent/EP3807822A4/en active Pending
- 2019-04-10 CN CN201980031921.5A patent/CN112740234B/en active Active
- 2019-04-10 WO PCT/US2019/026830 patent/WO2019240868A1/en active Search and Examination
Also Published As
Publication number | Publication date |
---|---|
CN112740234A (en) | 2021-04-30 |
WO2019240868A8 (en) | 2020-11-26 |
CN112740234B (en) | 2024-01-02 |
EP3807822A4 (en) | 2022-03-23 |
WO2019240868A1 (en) | 2019-12-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Koroniotis et al. | A new network forensic framework based on deep learning for Internet of Things networks: A particle deep framework | |
Abbas et al. | A new ensemble-based intrusion detection system for internet of things | |
Khraisat et al. | Survey of intrusion detection systems: techniques, datasets and challenges | |
Khan et al. | HML-IDS: A hybrid-multilevel anomaly prediction approach for intrusion detection in SCADA systems | |
Azmoodeh et al. | Detecting crypto-ransomware in IoT networks based on energy consumption footprint | |
Rao et al. | A hybrid intrusion detection system based on sparse autoencoder and deep neural network | |
Tomlinson et al. | Towards viable intrusion detection methods for the automotive controller area network | |
Hojjatinia et al. | Android botnet detection using convolutional neural networks | |
Elnakib et al. | EIDM: Deep learning model for IoT intrusion detection systems | |
Lee et al. | M-IDM: A Multi-Classification Based Intrusion Detection Model in Healthcare IoT. | |
Du et al. | Digital forensics as advanced ransomware pre-attack detection algorithm for endpoint data protection | |
Mohammed et al. | Intrusion detection system for NSL-KDD dataset based on deep learning and recursive feature elimination | |
Kuppa et al. | Finding rats in cats: Detecting stealthy attacks using group anomaly detection | |
Carmichael et al. | Unfooling perturbation-based post hoc explainers | |
Raymond et al. | Investigation of Android Malware with Machine Learning Classifiers using Enhanced PCA Algorithm. | |
US10902115B2 (en) | Neuromorphic system for authorized user detection | |
Jain et al. | Hybrid intelligent intrusion detection system for internet of things | |
WO2022115419A1 (en) | Method of detecting an anomaly in a system | |
Sangaiah et al. | Towards data security assessments using an IDS security model for cyber-physical smart cities | |
EP3807822A1 (en) | Neuromorphic system for authorized user detection | |
Khan et al. | DivaCAN: Detecting in-vehicle intrusion attacks on a controller area network using ensemble learning | |
US11256806B2 (en) | System and method for cyber attack detection based on rapid unsupervised recognition of recurring signal patterns | |
Sabir et al. | A Lightweight Deep Autoencoder Scheme for Cyberattack Detection in the Internet of Things. | |
EP3841504B1 (en) | System and method for cyber attack detection based on rapid unsupervised recognition of recurring signal patterns | |
Nasri et al. | Android malware detection system using machine learning |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE |
|
PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE |
|
17P | Request for examination filed |
Effective date: 20201117 |
|
AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR |
|
AX | Request for extension of the european patent |
Extension state: BA ME |
|
DAV | Request for validation of the european patent (deleted) | ||
DAX | Request for extension of the european patent (deleted) | ||
A4 | Supplementary search report drawn up and despatched |
Effective date: 20220223 |
|
RIC1 | Information provided on ipc code assigned before grant |
Ipc: G06N 3/08 20060101ALI20220217BHEP Ipc: G06N 3/04 20060101ALI20220217BHEP Ipc: G06F 21/31 20130101ALI20220217BHEP Ipc: G06F 21/88 20130101ALI20220217BHEP Ipc: G06N 3/063 20060101AFI20220217BHEP |
|
P01 | Opt-out of the competence of the unified patent court (upc) registered |
Effective date: 20230525 |