EP3987452A1 - Scalable integrated circuit with synaptic electronics and cmos integrated memristors - Google Patents
Scalable integrated circuit with synaptic electronics and cmos integrated memristorsInfo
- Publication number
- EP3987452A1 EP3987452A1 EP20827694.9A EP20827694A EP3987452A1 EP 3987452 A1 EP3987452 A1 EP 3987452A1 EP 20827694 A EP20827694 A EP 20827694A EP 3987452 A1 EP3987452 A1 EP 3987452A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- circuit
- single physical
- processing node
- memory
- synapse
- 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
- 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
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- 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
- G06N3/065—Analogue 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
- G06N3/088—Non-supervised learning, e.g. competitive learning
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C11/00—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor
- G11C11/54—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using elements simulating biological cells, e.g. neuron
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C11/00—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor
- G11C11/56—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using storage elements with more than two stable states represented by steps, e.g. of voltage, current, phase, frequency
- G11C11/5685—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using storage elements with more than two stable states represented by steps, e.g. of voltage, current, phase, frequency using storage elements comprising metal oxide memory material, e.g. perovskites
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C13/00—Digital stores characterised by the use of storage elements not covered by groups G11C11/00, G11C23/00, or G11C25/00
- G11C13/0002—Digital stores characterised by the use of storage elements not covered by groups G11C11/00, G11C23/00, or G11C25/00 using resistive RAM [RRAM] elements
- G11C13/0007—Digital stores characterised by the use of storage elements not covered by groups G11C11/00, G11C23/00, or G11C25/00 using resistive RAM [RRAM] elements comprising metal oxide memory material, e.g. perovskites
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C13/00—Digital stores characterised by the use of storage elements not covered by groups G11C11/00, G11C23/00, or G11C25/00
- G11C13/0002—Digital stores characterised by the use of storage elements not covered by groups G11C11/00, G11C23/00, or G11C25/00 using resistive RAM [RRAM] elements
- G11C13/0021—Auxiliary circuits
- G11C13/0023—Address circuits or decoders
Definitions
- This disclosure relates to neural networks .
- reference [3] listed below, a memristor array integrated with CMOS is shown. However no neural circuits, synapses, STDP or interconnect fabric are used in reference [3] .
- an array comprises a plurality of processing nodes, wherein each processing node comprises a single physical neuron circuit, the single physical neuron circuit having only one input and having an output, a single physical synapse circuit, the single physical synapse circuit having a presynaptic input, and the single physical synapse circuit having a single physical output coupled to the only one input of the single physical neuron circuit, a weight memory for storing N synaptic conductance value or weights, the weight memory having an output coupled to the single physical synapse circuit, wherein N is an integer greater than 1, a single physical spike timing dependent plasticity (STDP) circuit, the single physical STDP circuit having an output coupled to an input of the weight memory, a first input coupled to the output of the single physical neuron circuit, and a second input coupled to the presynaptic input, and interconnect circuitry connected to the presynaptic input and connected to the output of the single physical neuron circuit,
- STDP physical spike timing dependent plasticity
- a method of providing an array comprises a plurality of processing nodes, wherein each processing node comprises a single physical neuron circuit, the single physical neuron circuit having only one input and having an output, a single physical synapse circuit, the single physical synapse circuit having a presynaptic input, and the single physical synapse circuit having a single physical output coupled to the only one input of the single physical neuron circuit, a weight memory for storing N synaptic conductance value or weights, the weight memory having an output coupled to the single physical synapse circuit, wherein N is an integer greater than 1, a single physical spike timing dependent plasticity (STDP) circuit, the single physical STDP circuit having an output coupled to an input of the weight memory, a first input coupled to the output of the single physical neuron circuit, and a second input coupled to the
- presynaptic input and interconnect circuitry connected to the presynaptic input and connected to the output of the single physical neuron circuit, wherein the only one input of the single physical neuron circuit is only connected to the output of the single physical synapse circuit, wherein the single physical synapse circuit is time multiplexed to implement N virtual synapse circuits, and wherein the single physical STDP circuit is time multiplexed to implement N virtual STDP circuits, and wherein the
- interconnect circuitry in each respective processing node in the array is coupled to the interconnect circuitry in each other processing node in the array.
- FIG. 1A shows a diagram of a reconfigurable neural network having an array of processing nodes
- FIG. IB shows a diagram of a processing node with a processing core, a memristor memory and an interconnect fabric in accordance with the present disclosure
- FIG. 2A shows components of a processing core of a node and shows the interaction between a neuron circuit, a synapse circuit, an spike timing dependent plasticity (STDP) circuit and a weight memory in accordance with the present disclosure ;
- STDP spike timing dependent plasticity
- FIG. 2B shows an example array of processing nodes and shows that each processing node has a single physical neuron circuit, a single physical synapse circuit, a single physical STDP circuit, a weight memory, a
- FIG. 3 shows one embodiment of a neuron in accordance with the present disclosure
- FIGs . 4A and 4B show timing diagrams for time multiplexing in accordance with the present disclosure
- FIG. 5A shows a diagram of a memristor array inside one processing node with row and column access circuits
- FIG. 5B shows circuitry connected to rows of the memristor array
- FIG. 5C shows circuitry connected to the columns of the memristor array
- FIG. 5D shows an example of a typical I-V characteristic for a memristor
- FIG. 5E shows typical values of memristor currents when biased at 0.4V and correspondence to a synapse weight code in accordance with the present disclosure
- FIG. 6A shows the interconnect fabric of one processing node
- FIG. 6B shows a switch with on/off control
- FIG. 6C shows a bi-directional switch with control
- FIG. 6D shows a detail of memory to store the connectivity controls in accordance with the present disclosure
- FIG. 7A shows an example for simulation of a neural network with the outputs of a single time shared synapse circuit represented by synapse outputs 1-16 connected to a neuron circuit
- FIG.7B shows a memory with 16 rows corresponding to time slots and 34 columns corresponding to switches that are set in the interconnect fabric, where black represents an OFF state and white represents an ON state in accordance with the present disclosure
- FIG. 8A shows on the top of FIG. 8A the time shared synapse spike inputs to the neuron circuit and on the bottom of FIG. 8A the neuron circuit output, and FIG.
- 8B shows example synaptic conductance values over time coupled to the single synapse circuit in accordance with the present disclosure ;
- FIG. 9A shows a voltage waveform for a memristor read operation
- FIG. 9B shows a current waveform for a memristor read operation in accordance with the present disclosure
- FIG. 10A shows a voltage waveform for a memristor write operation
- FIG. 10B shows a current waveform for a memristor write operation in accordance with the present disclosure
- FIG. 11A shows a neural network with 10 neurons and 16 synapses
- FIG. 11B shows snap shots of switch states stored in memory for the network with 16 rows corresponding to the time slots and 34 columns corresponding to the switches that are set in the interconnect fabric for each neuron type, where black represents an OFF state and white represents an ON state
- FIG. 11C shows a simulation of the output neuron C in FIG. 11A
- FIG. 11D shows synaptic conductance values over time for the 16 synapses and shows convergence to the correct states in accordance with the present disclosure
- FIG. 12A shows a simulation of presynaptic inputs of the 16 synapses of FIG. 11A
- FIG. 12B shows the
- FIG. 12C shows the weights of the 16 synapses of FIG. 11A during a 3 second test in accordance with the present disclosure.
- a scalable neuromorphic integrated circuit is described with spiking neurons, synapses, and spike timing dependent plasticity (STDP), in which the connections between the neurons and synapses are not fixed but can be programmed.
- the synapses, STDP circuits and the interconnect routing between the neurons and synapses are time multiplexed.
- the integrated circuit includes memories to store both synaptic weights and interconnect routing information.
- the circuit includes memristor memories to achieve high density, and CMOS circuitry to write and read memristors.
- each node in an array is composed of a neuron circuit, a time multiplexed synapse circuit, a time multiplexed STDP circuit, a weight
- An object of the present disclosure is a
- disclosure is the ability to scale to high density while having the flexibility to be able to implement different neural networks with different topologies.
- FIG. 1A shows a top level diagram of a
- reconfigurable neural circuit 10 having an array 14 of processing nodes 12.
- the array 14 is a two dimensional array, and may be a square array.
- FIG. 1A an example array 14 of 10x10 processing nodes 12 is shown. Larger arrays may be built, and an integrated circuit with an array of 24x24 nodes has been implemented using 90nm CMOS technology .
- the processing nodes 12 are arranged as a two dimensional array 14. This arrangement allows the number of nodes 12 to scale with the area of the integrated circuit. As the length of the horizontal or vertical side of the integrated circuit increases the number of nodes 12 increases by the square of the length.
- FIG. IB a diagram of a processing node 12 is shown.
- the processing node 12 has a processing core 20 with a single physical neuron circuit 22, which may be an integrate and fire neuron circuit, a single physical synapse circuit 24, a single physical STDP (spike timing dependent plasticity) element 26, a memory 28 to store synaptic weights, and a memory 30 to store interconnect routing connectivity.
- the memories 28 and 30 may be CMOS.
- the STDP element 26 is an adaption element or circuit that adjusts the weight or gain of the synapse 24 according to a biologically inspired spike timing dependent plasticity (STDP) learning rule.
- STDP biologically inspired spike timing dependent plasticity
- the processing node 12 may include a memristor memory 32. As shown in FIG. 2, the memristor memory 32 has a memristor array 34 of memristors 35, DAC/DeMUX (digital to analog converter/demultiplexer) circuitry 33, and
- ADC/MUX analog to digital converter/multiplexer circuitry 36.
- the memristor memory 32 may be used to store synaptic weights. In one example embodiment there may be 128 memristors per node. As shown in FIG. 2, memory 28 in the processing core 20 may also store synaptic weights.
- the processing node 12 also includes interconnect circuitry or fabric 38, shown in FIGs. IB and 6A, that is used to enable communication between processing nodes 12.
- the interconnect circuitry 38 in each processing node is coupled to the presynaptic input 40 in each other processing node 12 and coupled to the output 42 of the single physical neuron circuit 22 in each other processing node 12.
- the interconnect fabric 38 may also be used to provide the connections between the neuron circuit 22, the synapse circuit 24, and the STDP circuit 26 in each processing node, as shown in FIG. 6A.
- the interconnect fabric 38 is composed of wire segments 83 and switches 84 and 85, which are shown in FIG. 6A. As shown in FIG. 6A, the switches 84 of the
- interconnect fabric 38 may provide the connections between the neuron circuit 22, the synapse circuit 24, and the STDP circuit 26 in a processing node.
- the array of nodes 12 in the reconfigurable neural circuit 10 is modular, and each node 12 may be directly abutted to its neighboring nodes 12. All the processing nodes 12 have the same processing core 20, the same memristor memory 32, and the same interconnect fabric 38. However, the operation of each node 12 may be
- Each node 12 may be programmed to support communication between nodes 12 that are both near and far away in the node array 14.
- FIG. 2A shows a diagram showing the interaction between the neuron circuit 22, which may be an integrate and fire neuron circuit, the synapse circuit 24, the STDP circuit 26, and the memristor memory 32 or the memory 28.
- the neuron circuit 22 There is one neuron circuit 22 per node 12, one synapse circuit per node, one STDP circuit per node, and a
- memristor memory 32 or a memory 28 Shown in FIG. 3 is an integrate-and-fire type neuron circuit 22, which integrates the input 25, shown in FIG. 2, in an internal accumulator
- the processing core 20 of each node 12 has only a single physical neuron circuit 22, a single physical synapse circuit 24 and a single physical STDP circuit 26.
- the single physical synapse circuit 24 has a single physical output 25 connected to the input of the single physical neuron circuit 22.
- the single physical neuron circuit 22 in the node has only one physical input 25 that is connected only to the single physical synapse circuit 24 in the node.
- the single physical synapse circuit 24 may be implemented in a number of ways and one way is the synapse circuit shown in Fig. 3b of Reference [2] above, which is incorporated herein by reference, and which shows that the synapse has a presynaptic input and a weight w.
- the boundary of the single physical synapse circuit 24 is the same as the portion of the synapse inside the circle of Fig. 3b with a Vpre input and i s output and a w weight, and the synapse core shown in FIGs . 5a and 5b.
- the synapse of Fig. 3b is not time shared.
- the single physical synapse circuit 24 is time multiplexed to implement N virtual synapse circuits, which reduces the amount of needed circuitry, which may be CMOS circuitry. Having a single physical synapse circuit 24 in a node 12 also reduces the number of needed interconnections in the node 12.
- the single physical STDP circuit 26 in the node 12 may be implemented in a number of ways and one way is the STDP circuit shown in Fig. 5b of Reference [2] above.
- the boundary of the single physical synapse circuit 24 is the same as the portion of Fig. 5a and 5b that shows the STDP circuit with Vpre and Vpost inputs and a w output to the synapse core.
- the synapse 24 is time multiplexed as weights are read from the memristor memory 32 or the memory 28 to implement N virtual synapses.
- output 42 of the neuron 22 changes at each time interval, so the inputs to the STDP circuit also change each time interval so that the one physical STDP circuit is time multiplexed to implement N virtual STDPs, which reduces the amount of needed circuitry and also reduces the number of needed
- the memory for storing synaptic weights stores N synaptic weights, the same number as the number N of virtual synapse circuits and virtual STDP circuits.
- the synaptic weights to implement N virtual synapse circuits may be stored in the memristor memory 32 or in memory 28. In one example embodiment N may be 128.
- the single physical STDP circuit 26 has two inputs, one input is presynaptic input 40 and a second input is fed back from the postsynaptic output 42 of the single physical neuron circuit 22.
- the STDP circuit 26 output 44 is used to access a synaptic weight 46 from the memristor memory 32 or from memory 28.
- the synaptic weight 46 is connected to the single physical synapse circuity 24, as shown in FIG. 2A.
- FIG. 2B shows an example array of processing nodes 12 having a two dimensional array of nine processing nodes 12 arranged in a 3 x 3 array.
- the size of the array and the number of processing nodes in an array may be any size and number.
- FIG. 2B shows that each processing node 12 has a single physical neuron circuit 22, a single physical synapse circuit 24, a single physical STDP circuit 26, a weight memory 37, a connectivity memory 30 and interconnect circuity 38.
- the weight memory 37 may be the memristor memory 32 or memory 28.
- the interconnect circuity 38 in a node 12 is connected to the interconnect circuitry in all the other nodes 12, as shown in FIG. 2B. [ 0041 ] FIG.
- FIG. 3 shows an embodiment of a neuron circuit 22 in the processing core 20.
- the neuron circuit 22 is a digital circuit.
- the input 52 shown in FIG. 2 shows multiple virtual synapse inputs to the neuron from the single physical synapse circuit 24. Each virtual synapse input is digital and the output of the accumulator 48 is a count. When the count of the accumulator 48 reaches the threshold value the comparator 50 outputs a spike on line 51 and resets the accumulator 48.
- the spike may be a 1-bit digital signal.
- the accumulator 48 integrates the input 52, which may be the output of one or more virtual synapse circuits 24, that can be implemented by the time multiplexed single physical synapse circuit 24, whose single physical output 25 is connected to the single physical neuron circuit 22, as shown in FIG. 2A and 2B.
- the comparator 50 compares the accumulator 48 state to a threshold voltage, Vth 54, which may be digital. If the accumulator 48 state is more than Vth 54, the comparator 50 produces an output spike, and resets the accumulator 48 state. For example, the
- the accumulator 48 state may be reset to a count corresponding to 0 volts.
- the accumulator 48 may be a 9 bit accumulator.
- FIG. 6A Another neuron implementation that may be used is shown in FIG. 6A in U.S. Patent Application
- FIGs . 4A and 4B show timing diagrams for synaptic time multiplexing.
- This synaptic time multiplexing may be performed by dividing the time consumed in a given STM cycle 56, which is the time required to cycle through N virtual synapse circuits and N virtual STDP circuits, into time slots 58, which may, in one example, be 100 ps duration each for a total cycle time of up to N*100 ps duration .
- the synapse circuit 24 is assigned to do the function of one given virtual synapse and one virtual STDP.
- N may be 128, so during a 12.8 ms cycle, which corresponds to 128 time slots 58, the synapse 24 may implement 128 different virtual synapses and STDPs. Time multiplexing requires the storage of one synaptic
- this storage may be provided by a memristor array 34 of 128 memristors 35.
- one memristor 35 is read to access a synaptic conductance for a synapse.
- the stored synaptic weight value in each memristor 35 in the memristor array 34 may be updated according to an update value provided by the STDP circuit 26. The update value is used to increment or decrement the currently stored synaptic conductance value in the
- the memristor 35 for the virtual synapse may be accessed in a fixed order.
- the respective memristor 35 corresponding to a virtual synapse may be accessed once for reading and, if needed, once for writing to increment or decrement the currently stored synaptic conductance value by the update value .
- the synaptic weights or conductance values may be stored with 3 bits of accuracy in either memory 28 or memristor memory 32.
- the memory 28 may be made of CMOS flip-flops, a SRAM (static random access memory) , or any other type of digital memory.
- the memristor array 34 of each node 12 interfaces to circuitry, which may be CMOS, to select a memristor 35 for a read or write operation.
- a symbolic diagram of the memristor array 34 is shown in FIG. 5A, and shows 128 memristors 35 with nanowires 60 and 61 arranged in 16 rows and 8 columns, respectively.
- the row circuit 64 and the column circuit 66 are used to select at any one time slot 58 one memristor 35 of the memristor array 34 to perform either a read or a write operation.
- FIG. 5B shows a diagram of row circuit 64, which has a buffer amplifier 70, an analog to digital converter (ADC) 72, and a de-multiplexer (DeMUX)74.
- a row address is input to the DeMUX 74 to connect the buffer amplifier 70 output Vsel row 71 to one of the row vias 62 connected to memristor nanowires 60.
- the other 15 unselected nanowires 60 may be connected to a bias voltage 76.
- the buffer amplifier 70 For reading a memristor value from a memristor 35, the buffer amplifier 70 is used to set a reading voltage on Vsel row 71.
- the amplifier 70 has an extra terminal 73 that provides a current equal to that flowing to the memristor.
- This current which is proportional to the value stored in the memristor, is digitized by the analog to digital converter 72 to produce a synaptic weight or synaptic conductance value 46, which may be a 3-bit (8- level) code.
- the synaptic weight or synaptic conductance value 46 is applied to the synapse 24 as shown in FIG. 2.
- FIG. 5C shows a diagram for the column circuit
- DeMUX demultiplexer
- Vsel col 79 connects Vsel col 79 to one of the column vias 63 connected to memristor nanowires 61, shown in FIG. 5A, in accordance with a column address 80.
- the other unselected nanowires are connected to other bias voltages, such as bias voltage 81. These bias voltages are used to minimize leakage paths.
- FIG. 5D A typical I-V characteristic of a memristor 35 is shown in FIG. 5D, and typical current levels during a memristor read operation at 0.4V are shown in FIG. 5E.
- the output spike signals produced by a neuron circuit 22 in a processing node 12 can be routed to the synapse circuit 24 of a different node 12 by interconnect fabric 38, which is shown in detail in FIGs. 6A, 6B and 6C .
- FIG. 6A shows the detail of the interconnect fabric or interconnect circuitry 38 associated with one node 12.
- the interconnect fabric 38 is composed of conductors or wires 83, uni-directional buffer-based switches 84, bi-directional buffer-based switches 85 and memory 30.
- the memory 30 stores connectivity data used by the interconnect fabric. In one embodiment this memory is implemented using CMOS technology.
- the states, ON or OFF, of each of the switches 84 and 85 are stored in memory 30. All the nodes 12 have the same hardware, but the switch states can be programmed independently in each node 12, for each time slot 58 of a STM cycle 56.
- a unidirectional switch 84 may be implemented by a buffer 84 as shown in FIG. 6B.
- the buffer 84 can be turned on or off according to a control line 86.
- the detail of a bi-directional switch 85 is shown in FIG. 60. This switch is composed of two buffers, but only one of the two buffers may be set ON at a given time, and the state of the bi-directional switch 85 is controlled by two control lines 87.
- the memory 30 inside the node 12 contains information about the control (ON or OFF) of all of the switches 84 and 85. All the switch states for a node 12 for all the time slots 58 in a STM cycle 56 may be the same for each STM cycle 56. So the memory 30 need only store the switch states of all the time slots of one STM cycle, which is the time required to cycle through N virtual synapses .
- the memory 30 stores the interconnect or routing configuration for all the time slots of a STM cycle as shown in FIG. 6D.
- the memory 30 has 34 columns and N rows per node.
- the data in a column of the memory 30 is used to generate control signals for a particular switch for all N time slots.
- the memory 30 is initialized with a user-defined network topology.
- a neuromorphic compiler can be used to initialize this memory. An example of compiler is described in Reference [5] above, which is incorporated herein by reference.
- the reconfigurable neural circuit 10 can be programmed to implement different neural networks . Two simulations are described below. In each simulated case the integrated circuit implements a particular neural network topology.
- a neural network in a first simulation, shown in FIG. 7A, is composed of a neuron 22 and 16 synapses 24.
- the weight of each synapse is internally controlled by a STDP circuit 26.
- the network of FIG. 7A can distinguish if several of their inputs are correlated to each other.
- the simulation that implements the network of FIG. 7A begins by initializing the memory 30 for routing spikes between neurons by setting the switch states in the node as shown in FIG. 7B.
- the plot of FIG. 7B shows the contents of the memory 30 using the same format as shown in FIG. 6D.
- FIG. 7B shows the memory as an array of bits. Columns are associated with a particular switch of the node. Rows are associated with a time slot.
- the size of the memory 30 is 34 rows each with 16 bits. In this example a bit with value of "0" is shown in black and a bit with value of "1" is shown in white.
- the top plots of FIG. 8A show the inputs provided to the synapses 24 in the form of spike trains.
- a set of eight different input spike trains, not correlated to each other, are applied to synapses 9 through 16.
- One additional spike train signal, shown as Ini_ 3 in FIG. 8A is used as a common input to synapses 1 through 8.
- the network can be used to determine which inputs are correlated or uncorrelated to each other.
- the circuit of the invention can implement different neural networks (with different topologies) that could be used for different applications . To illustrate and simulate the operation of the circuit we have implemented, as a first example, a neural network that can be used in the application of distinguishing correlated inputs to uncorrelated inputs.
- the circuit of the invention can also implement other neural networks for other applications .
- FIG. 8A The bottom plot of FIG. 8A shows the output produced by the neuron 22 shown in FIG. 7A during the simulation.
- the presynaptic inputs 40 and the neuron output 42 are used by a STDP circuit 24 to generate updates to the synaptic conductance values .
- 16 synaptic conductance weights that are stored in 16 memristors.
- the time evolution of the 16 synaptic conductance weights, denoted as wpi to wi,i 6 is shown in FIG. 8B. They are stored in memristors Mpi through M I , I6 35 within one node.
- the memristors of a node are accessed cyclically for each time slot 58.
- Each access operation consists of a memristor read, calculation of a weight increment or decrement by the STDP circuit, and a memristor write to update the synaptic conductance weight in the memristor.
- the writing is performed only if there is a change different from a zero increment or decrement.
- the vertical axes of the plots of FIG. 8B represent the code for the synaptic conductance value 46.
- This code is produced by the ADC 72 shown in FIG. 5B that ranges from 0 to 7 in steps of 1. It can be observed, according to the simulation, that after 0.3 second the synaptic conductance values w lrl to wi 3 , which are associated with synapses 24 receiving correlated inputs, all tend to a high value. The weights w l 9 to wi ,i6 , associated with synapses 24 receiving uncorrelated inputs, all tend to a low value, which is the desired behavior.
- FIGs . 9A and 9B The details of a single memristor 35 read operation during this simulation are shown in FIGs . 9A and 9B .
- the waveform in FIG. 9A labeled as P 100 represents the voltage applied to a positive terminal of the memristor 35 and is 0.4 V in this simulation, however, the voltage can also be programmed to be a different value.
- the line labeled as N 102 represents the voltage applied to the negative terminal of the memristor 35. It is zero during a read operation.
- the control signal shown in FIG. 9A, enables the operation of the ADC circuit 72 to digitize the current of the memristor to one of eight possible synaptic conductance codes 46. In a typical embodiment the read operation lasts 4 ps.
- the current through the memristor 35 during a read operation is shown in FIG. 9B . In a typical embodiment, during a read operation, the currents range from 2 mA to 16 mA.
- FIGs. 10A and 10B The details of a typical write operation during the simulation are shown in FIGs. 10A and 10B.
- the write operation is used to increment or decrement the value of a memristor 35 by an update value.
- the STDP 26 calculates the required increase or decrease to the synaptic weight. Then it applies one pulse or a set of pulses in proportion to the magnitude of the change in synaptic conductance to one of the two terminals of the memristors.
- pulses are applied to the positive terminal of the memristor 35.
- the key voltages for the write operation are shown in FIG. 10A.
- the waveform labeled as P 104 represents the voltage applied to the positive terminal of the memristor 35. In this simulation the write voltage used is 1.4 volts, however, the voltage can also be programmed to be a different value.
- the line labeled as N 106 represents the voltage applied to the negative terminal of the memristor 35. It is approximately zero during a read operation.
- the dotted line 108 represents a control signal that sets the duration of the write pulse. For writing an increment to a synaptic weight, from 1 to 4 write pulses are applied.
- the number of pulses is determined by an on-chip control circuit that reads the memristor current just after each write pulse.
- the set of pulses is stopped when the target increment value is achieved.
- the target increment is achieved after two write pulses.
- the current through the memristor 35 during the writing sequence is shown in FIG. 10B.
- each write pulse currents of about 100 ps can flow through the memristor 35.
- the read currents measured in the 4 ps intervals after each write pulse are in the desired range of 2 pA to 16 pA. For a decrement operation, a similar process occurs when the pulses are applied to the negative terminal of the memristor 35.
- FIG. 11A The simulation of a reconfigurable neural circuit 10 implementing a more complex network with ten neurons 22 is shown in FIG. 11A. This simulation has an additional layer of 9 neurons 22 located between the inputs 90 and the output neuron 92, which has an output 93.
- the functional behavior of this network is similar to the simulation described above, in that this network can distinguish correlated inputs from uncorrelated inputs.
- the switch states in each node 12 are set as shown in FIG. 11B and are used to route the spikes between the various neurons in the network during each STM cycle 56. The process is repeated from the beginning after the completion of each STM cycle.
- the output 93 of the neuron 92 is shown in FIG. 11C.
- the time evolution of the 16 synaptic weights, denoted as wi ,i to wi ,i6 is shown in FIG. 11D.
- the weights are stored in memristors Mc , i to M I , I6 35 within a node 12. In one embodiment of the invention these memristors are part of the 8x16 array of memristors shown in FIG. 5A.
- FIG. 11A An integrated circuit implementing a reconfigurable neural circuit 10 was fabricated and tests were conducted.
- the reconfigurable neural circuit 10 was configured to implement the same network as shown in FIG. 11A, which as described can distinguish if several input signals are correlated to each other.
- the synaptic weights were stored in memory 28.
- the graph of FIG. 12A shows the presynaptic inputs of those 16 synapses. Eight of the inputs to synapses are the same and therefore correlated to each other. The other eight inputs to other synapses are uncorrelated to each other.
- the graph of FIG. 12B shows the postsynaptic spike produced by the output neuron 92.
- the graph of FIG. 12C shows the weights of 16 synapses during a 3 second test. At the end of the test, the weights diverge to either a high value or a low value. The weights reaching a high value correspond to synapses receiving correlated inputs, as desired. The weights reaching a low value correspond to synapses receiving uncorrelated inputs, as desired.
- a reconfigurable neural circuit includes an array of processing nodes.
- Each processing node includes a single physical neuron circuit having only one input and an output, a single physical synapse circuit having a
- presynaptic input and a single physical output coupled to the input of the neuron circuit
- a weight memory for storing N synaptic conductance value or weights having an output coupled to the single physical synapse circuit
- a single physical spike timing dependent plasticity (STDP) circuit having an output coupled to the weight memory, a first input coupled to the output of the neuron circuit, and a second input coupled to the presynaptic input
- interconnect circuitry connected to the presynaptic input and connected to the output of the single physical neuron circuit.
- the synapse circuit and the STDP circuit are each time multiplexed circuits.
- the interconnect circuitry in each respective processing node is coupled to the
- a reconfigurable neural circuit comprising: an array comprising a plurality of processing nodes; wherein each processing node comprises:
- the single physical neuron circuit having only one input and having an output
- the single physical synapse circuit having a presynaptic input, and the single physical synapse circuit having a single physical output coupled to the only one input of the single physical neuron circuit;
- a weight memory for storing N synaptic conductance value or weights, the weight memory having an output coupled to the single physical synapse circuit, wherein N is an integer greater than 1;
- STDP spike timing dependent plasticity
- interconnect circuitry connected to the presynaptic input and connected to the output of the single physical neuron circuit
- the single physical synapse circuit is time multiplexed to implement N virtual synapse circuits; and wherein the single physical STDP circuit is time multiplexed to implement N virtual STDP circuits; and wherein the interconnect circuitry in each respective processing node in the array is coupled to the interconnect circuitry in each other processing node in the array.
- each processing node is adapted to read a respective one of the N synaptic conductance values or weights from the weight memory for each respective time period of N time periods and wherein each respective one of the N synaptic conductance values or weights coupled to the single physical synapse circuit during a respective time period of the N time periods causes the single physical synapse circuit to operate as a respective virtual synapse according to the coupled conductance value or weight during the respective time period;
- each processing node is adapted to update the respective synaptic conductance value or weight read from the weight memory by writing the weight memory in the respective time period according to the output of the single physical spike timing dependent plasticity circuit during the respective time period.
- the neuron circuit comprises an integrate and fire circuit .
- the weight memory comprises a memristor memory, flip flops, a static random access memory or a digital memory.
- Concept 5 The reconfigurable neural circuit of concept 1, 2 , 3 or 4 :
- interconnect circuitry in each processing node is coupled to the presynaptic input in each other processing node and coupled to the output of the single physical neuron circuit in each other processing node.
- the interconnect circuitry comprises a plurality of switches for changing the interconnections to and from a respective processing node and other processing nodes in the array.
- the plurality of switches comprise a plurality of uni directional and bi-directional switches .
- the single physical STDP circuit comprises a
- STDP spike timing dependent plasticity
- the array comprises a two dimensional array of processing nodes.
- each processing node further comprises:
- each processing node comprises a plurality of switches
- the connectivity memory stores a plurality of interconnect circuitry routing controls for each respective switch of the plurality of switches
- the plurality of interconnect circuitry routing controls for a respective switch of the plurality of switches is configured to control the respective switch for each respective time period of N time periods.
- interconnect circuitry routing controls stored in the connectivity memory in each processing node determine whether the output of the single physical neuron circuit in a respective processing node is connected to a respective presynaptic input in another processing node.
- interconnect circuitry in a respective processing node provides connections between the single physical neuron circuit, the single physical synapse circuit, and the single physical STDP circuit in the respective processing node.
- the plurality of interconnect circuitry routing controls stored in the connectivity memory in a respective processing node is programmed independently for each respective processing node in the array.
- each column has N rows corresponding to the N time periods
- interconnect circuitry routing controls in a column for a respective switch comprise N controls for the respective switch, wherein each of the N controls is ON or OFF.
- time is divided into synaptic time multiplexing cycles
- each cycle comprises N time slots; and wherein during each of the N time slots the weight memory outputs a respective one of the N synaptic
- the single physical synapse circuit implements a respective one of the N virtual synapse circuits and outputs an output of the respective one of the N virtual synapse circuits to the single physical neuron circuit
- the single physical STDP circuit implements a respective one of the N virtual STDP circuits based on the output of the single physical neuron circuit during the time slot and the presynaptic input during the time slot, and update the respective one of the N synaptic conductance value or weights during the time slot .
- each processing node has a connectivity memory for controlling the interconnect circuitry in the
- the interconnect circuitry in a respective processing node is configured according to controls read from the connectivity memory in the respective processing node.
- a method for providing a reconfigurable neural circuit comprising:
- each processing node comprises:
- the single physical neuron circuit having only one input and having an output
- the single physical synapse circuit having a presynaptic input, and the single physical synapse circuit having a single physical output coupled to the only one input of the single physical neuron circuit;
- a weight memory for storing N synaptic conductance value or weights, the weight memory having an output coupled to the single physical synapse circuit, wherein N is an integer greater than 1;
- STDP spike timing dependent plasticity
- interconnect circuitry connected to the presynaptic input and connected to the output of the single physical neuron circuit
- the only one input of the single physical neuron circuit is only connected to the output of the single physical synapse circuit; wherein the single physical synapse circuit is time multiplexed to implement N virtual synapse circuits;
- the single physical STDP circuit is time multiplexed to implement N virtual STDP circuits; and wherein the interconnect circuitry in each respective processing node in the array is coupled to the interconnect circuitry in each other processing node in the array.
- each respective one of the N synaptic conductance values or weights coupled to the single physical synapse circuit during a respective time period of the N time periods causes the single physical synapse circuit to operate as a respective virtual synapse according to the coupled conductance value or weight during the respective time period;
- the neuron circuit comprises an integrate and fire circuit .
- the weight memory comprises a memristor memory, flip flops, a static random access memory or a digital memory.
- the interconnect circuitry comprises a plurality of switches for changing the interconnections to and from a respective processing node and other processing nodes in the array.
- the plurality of switches comprise a plurality of uni directional and bi-directional switches .
- the STDP circuit comprises a biologically inspired spike timing dependent plasticity (STDP) learning rule.
- STDP biologically inspired spike timing dependent plasticity
- providing the array comprises providing a two
- each processing node further comprises:
- the interconnect circuitry in each processing node comprises a plurality of switches; and storing in the connectivity memory a plurality of interconnect circuitry routing controls for each respective switch of the plurality of switches; and configuring the plurality of interconnect circuitry routing controls for a respective switch of the plurality of switches to control the respective switch for each respective time period of N time periods.
- interconnect circuity routing controls stored in the connectivity memory in each processing node determine whether the output of the single physical neuron circuit in a respective processing node is connected to a respective presynaptic input in another processing node.
- interconnect circuitry in a respective processing node provides connections between the single physical neuron circuit, the single physical synapse circuit, and the single physical STDP circuit in the respective processing node.
- programming is independent for each respective processing node in the array.
- the connectivity memory has a column
- each column has N rows corresponding to the N time periods
- interconnect circuitry routing controls in a column for a respective switch comprise N controls for the respective switch, wherein each of the N controls is ON or OFF.
- time is divided into synaptic time multiplexing cycles
- each cycle comprises N time slots; and wherein during each of the N time slots the weight memory outputs a respective one of the N synaptic
- the single physical synapse circuit implements a respective one of the N virtual synapse circuits and outputs an output of the respective one of the N virtual synapse circuits to the single physical neuron circuit
- the single physical STDP circuit implements a respective one of the N virtual STDP circuits based on the output of the single physical neuron circuit during the time slot and the presynaptic input during the time slot, and update the respective one of the N synaptic conductance value or weights during the time slot .
- each processing node has a connectivity memory for controlling the interconnect circuitry in the
- the interconnect circuitry in a respective processing node is configured according to controls read from the connectivity memory in the respective processing node.
Abstract
Description
Claims
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US16/447,210 US11501143B2 (en) | 2013-10-11 | 2019-06-20 | Scalable integrated circuit with synaptic electronics and CMOS integrated memristors |
PCT/US2020/037891 WO2020257168A1 (en) | 2019-06-20 | 2020-06-16 | Scalable integrated circuit with synaptic electronics and cmos integrated memristors |
Publications (2)
Publication Number | Publication Date |
---|---|
EP3987452A1 true EP3987452A1 (en) | 2022-04-27 |
EP3987452A4 EP3987452A4 (en) | 2023-06-28 |
Family
ID=74037186
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
EP20827694.9A Pending EP3987452A4 (en) | 2019-06-20 | 2020-06-16 | Scalable integrated circuit with synaptic electronics and cmos integrated memristors |
Country Status (3)
Country | Link |
---|---|
EP (1) | EP3987452A4 (en) |
CN (1) | CN113994346A (en) |
WO (1) | WO2020257168A1 (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113313240B (en) * | 2021-08-02 | 2021-10-15 | 成都时识科技有限公司 | Computing device and electronic device |
CN115831190B (en) * | 2023-02-16 | 2023-05-09 | 华中科技大学 | Self-write-stop operation circuit and self-write-stop operation method of memristor |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8712940B2 (en) * | 2011-05-31 | 2014-04-29 | International Business Machines Corporation | Structural plasticity in spiking neural networks with symmetric dual of an electronic neuron |
US20160364643A1 (en) * | 2012-03-08 | 2016-12-15 | Hrl Laboratories Llc | Scalable integrated circuit with synaptic electronics and cmos integrated memristors |
FR3007867B1 (en) * | 2013-06-26 | 2018-02-09 | Commissariat A L'energie Atomique Et Aux Energies Alternatives | ARTIFICIAL NEURONE COMPRISING A RESISTIVE MEMORY |
US20150278682A1 (en) * | 2014-04-01 | 2015-10-01 | Boise State University | Memory controlled circuit system and apparatus |
-
2020
- 2020-06-16 WO PCT/US2020/037891 patent/WO2020257168A1/en active Application Filing
- 2020-06-16 EP EP20827694.9A patent/EP3987452A4/en active Pending
- 2020-06-16 CN CN202080041609.7A patent/CN113994346A/en active Pending
Also Published As
Publication number | Publication date |
---|---|
CN113994346A (en) | 2022-01-28 |
EP3987452A4 (en) | 2023-06-28 |
WO2020257168A1 (en) | 2020-12-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20160364643A1 (en) | Scalable integrated circuit with synaptic electronics and cmos integrated memristors | |
US11501143B2 (en) | Scalable integrated circuit with synaptic electronics and CMOS integrated memristors | |
US10990651B2 (en) | Systems and methods for efficient matrix multiplication | |
TWI661428B (en) | Neuromorphic weight cell and method of forming the same and artificial neural network | |
US11074982B2 (en) | Memory configured to perform logic operations on values representative of sensed characteristics of data lines and a threshold data value | |
US11361215B2 (en) | Neural network circuits having non-volatile synapse arrays | |
Cruz-Albrecht et al. | A scalable neural chip with synaptic electronics using CMOS integrated memristors | |
CN110352436B (en) | Resistance processing unit with hysteresis update for neural network training | |
Kim et al. | A digital neuromorphic VLSI architecture with memristor crossbar synaptic array for machine learning | |
CA2817802C (en) | Electronic synapses for reinforcement learning | |
JP2021507349A (en) | A method for storing weights in a crosspoint device of a resistance processing unit array, its crosspoint device, a crosspoint array for performing a neural network, its system, and a method for performing a neural network. Method | |
US10366326B2 (en) | Unit having an artificial neuron and a memristor | |
US11157803B2 (en) | Neuromorphic device including a synapse having a variable resistor and a transistor connected in parallel with each other | |
US20130031040A1 (en) | Hierarchical routing for two-way information flow and structural plasticity in neural networks | |
JP2013546065A (en) | Methods, devices, and circuits for neuromorphic / synaptronic spiking neural networks with synaptic weights learned using simulation | |
EP3987452A1 (en) | Scalable integrated circuit with synaptic electronics and cmos integrated memristors | |
JP6865819B2 (en) | Memory cell structures, memory cell systems, and methods for manufacturing memory cell structures | |
US11227211B2 (en) | Neuromorphic device including a synapse having a plurality of synapse cells | |
KR20210151737A (en) | Synapse string and synapse string array for neural networks | |
KR20170080433A (en) | Methods of Reading-out Data from Synapses of Neuromorphic Device | |
Mitra et al. | Learning to classify complex patterns using a VLSI network of spiking neurons | |
CN114861902A (en) | Processing unit, operation method thereof and computing chip | |
KR20170117861A (en) | Neural Network Systems | |
Dalgaty et al. | The neuromorphic Mosaic: re-configurable in-memory small-world graphs | |
US11501150B2 (en) | Neural network architecture |
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 |
|
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: 20211203 |
|
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 |
|
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: 20230530 |
|
RIC1 | Information provided on ipc code assigned before grant |
Ipc: G06N 3/088 20230101ALI20230523BHEP Ipc: G06N 3/065 20230101ALI20230523BHEP Ipc: G06N 3/049 20230101ALI20230523BHEP Ipc: G11C 13/00 20060101ALI20230523BHEP Ipc: G11C 11/54 20060101ALI20230523BHEP Ipc: G06N 3/063 20060101AFI20230523BHEP |