UNSUPERVISED, SUPERVISED AND REINFORCED LEARNING VIA SPIKING COMPUTATION

• US 20140250039A1
• Filed: 08/16/2012
• Published: 09/04/2014
• Est. Priority Date: 09/16/2011
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Modha, Dharmendra S.
• Granted Patent: US 9,245,223 B2
• Time in Patent Office: Days
• US Class Current: 706/25
• CPC Class Codes: G06N 3/04 Architecture, e.g. interconnection topology
• CPC Class Codes: G06N 3/049 Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
• CPC Class Codes: G06N 3/063 using electronic means
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G06N 3/088 Non-supervised learning, e.g. competitive learning

Abstract

The present invention relates to unsupervised, supervised and reinforced learning via spiking computation. The neural network comprises a plurality of neural modules. Each neural module comprises multiple digital neurons such that each neuron in a neural module has a corresponding neuron in another neural module. An interconnection network comprising a plurality of edges interconnects the plurality of neural modules. Each edge interconnects a first neural module to a second neural module, and each edge comprises a weighted synaptic connection between every neuron in the first neural module and a corresponding neuron in the second neural module.

Claims

1. A method comprising:

producing spiking computation in a neural network comprising a plurality of neural modules interconnected via weighted synaptic connections in an interconnection network, wherein each neural module comprises multiple digital neurons such that every neuron in a first neural module is connected to a corresponding neuron in a second neural module via a weighted synaptic connection;

wherein said spiking computation comprises generating signals which define a set of time steps for operation of the neurons, and at each time step, each neuron based on its operational state determines whether to generate a firing event in response to firing events received as input signals from neurons in other neural modules, wherein each said input signal is weighted by the weighted synaptic connection communicating said input signal to said neuron.

2. The method of claim 1, further comprising:

at each time step, each neuron in a neural module updating an operational state of said neuron; and

adapting the synaptic weight of each weighted synaptic connection as a function of the firing events of the interconnected neurons.

3. The method of claim 2, further comprising:

updating learning rules associated with each weighted synaptic connection, wherein each weighted synaptic connection is reconfigurable for learning rules comprising one or more of;

Hebbian, anti-Hebbian, spike-timing-dependent-plasticity (STDP), and anti-STDP.

4. The method of claim 2, further comprising:

applying a learning rule for learning false negatives to a weighted synaptic connection when a neuron interconnected via the weighted synaptic connection generates a firing event.

5. The method of claim 4, further comprising:

applying a learning rule for unlearning false positives to a weighted synaptic connection when a neuron interconnected via the weighted synaptic connection generates a firing event.

6. The method of claim 5, further comprising:

a set of neurons receiving input events from an input interface to the neural network; and

a set of neurons generating output events to an output interface of the neural network.

7. The method of claim 5, further comprising:

a set of neurons generating output events to an evaluation module in the neural network, wherein, based on the type of the output events generated, the output events are fed to another set of neurons by the evaluation module.

8. The method of claim 5, further comprising:

updating the learning rules such that the neural network operates as one or more of;

an auto-associative system, a hetero-associative system, and a reinforcement learning system.

9. The method of claim 5, further comprising:

adaptively updating the weights of weighted synaptic connections in the interconnection network as a function of the firing events of the interconnected neurons.

10. The method of claim 5, further comprising:

for a first set of neurons in the first neural module, information flowing from each of said neuron in the first set to a corresponding neuron in the second neural module; and

for a second set of neurons in the second neural module, information flowing from each of said neuron in the second set to a corresponding neuron in the first neural module.