JPH0373058A - Rationalizing method for data processor - Google Patents

Abstract

PURPOSE:To prevent a drop into a local minimum in a learning process by forcedly changing the weight of a neuron generating a significant output at a certain time point so as to be temporarily decreased to a minimum value and afterwards to be increased to the value at this time point, and adjusting the weight during the period of this change. CONSTITUTION:Plural neural layers NL are provided with plural neurons N parallelly provided and the neural layer is constituted so that the output of a certain neural layer can be the input of the neural layer in the next step. Threshold values theta of the neurons in the same neural layer are made same and the threshold value theta is increased toward the neural layer in the rear step and changed like a step. The weight of the neuron generating the significant output at a certain time point is forcedly changed so as to be temporarily decreased to a minimum value and afterwards to be increased to the value at this time point. Then, while applying fixed input data during the period for this forced change of the weight, the weight is corrected based on the evaluated result of the output. Thus, the drop into the local minimum can be prevented.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention provides a plurality of neurons that output data according to a comparison result between the sum of products obtained by multiplying input data by a predetermined weight and a threshold value. The present invention relates to a method for optimizing the weights in a data processing device.

[Prior Art] In the learning process of this type of data processing device, when a certain degree of appropriate input-output correlation occurs and that correlation is strengthened,
Thereafter, the data processing device may not be able to escape from this state, and optimization may become impossible.

This is expressed as a fall to a local minimum, for example, a Boltzmann machine (David H, Ackley,
Geoffrey E, Hinton+and T
errence J, Sejnowski: Ale
Arning algorithm for Bolz
man machines: Cognitive 5c
ieence 9.1989.147-169), the energy is temporarily jumped to a predetermined high level in order to escape from the energy equation of the neural network and reach the global local minimum. However, the optimum correlation between input and output does not necessarily match the state of the global local minimum of energy. In addition, the depth of the local minimum is generally unknown, and there is no guarantee that when the energy level is increased by a predetermined value, it will be possible to escape from the local minimum, and the global local minimum of energy will be reached. There is no guarantee that it will even be possible. Furthermore, a temporary jump in energy itself is inconsistent with the functioning of biological systems.

[Problem to be solved by the invention]

The present invention was devised to solve the above-mentioned conventional problems, and it is an object of the present invention to provide a method for optimizing a data processing device that can reliably prevent a drop to a local minimum during the learning process.

[Means to solve the problem]

A method for optimizing a data processing device according to the present invention is to temporarily reduce the weight of a neuron that has produced a significant output at a certain point in time to a minimum value while giving a constant input to the data processing device, and then The weight is forcibly changed so as to increase to the value , and the weight is adjusted during the period of this forced change of the weight.

[Operation] According to the method for optimizing a data processing device according to the present invention, it is possible to reliably escape from a local minimum value using a mechanism that conforms to the function of the biological system.

〔Example] The present invention will be explained below with reference to illustrated embodiments.

Here, as an indicator showing the effectiveness of learning, the following standard - maturity T
Let us define P.

However, here, n: number of input events i, j: number of input events Pt, degree of coincidence between the neural network output due to the jsi-th input event and the neural network output due to the j-th input event (number of matching bits) Standard-graded TP is an index of the discrimination ability of input events in a neural network, and the lower the TP, the higher the discrimination ability. In a neural network that has not been trained at all, TP = 100%.
, and TP monotonically decreases with subsequent learning.

(Figure 1) However, if the learning method is inappropriate, it may actually impede the learning effect, and as shown in Figure 2, the TP that once decreased may rise again. .

When a drop to the local minimum occurs as described above, TP becomes saturated at a high value as shown in FIG.

This local drop to the local minimum is unrelated to the energy equation (1). For example, the weight of the features of a certain event may be biased, and as a result of emphasizing a certain feature, confusion of multiple events may occur. This can be thought of as the result of learning to affirm what will happen.

Expressing this schematically, as shown in Figure 4, because the influence of a certain input bit 4 is strong, the firing distribution of neurons becomes biased, and this state can be considered as a state that continues to be strengthened. . In this figure, neurons that have fired are indicated by black circles.

In order to escape from this state, other input bits 11 to I
There is a need to strengthen neuronal involvement in i,Is.

Here, the inventors focused on the absolute refractory period and relative refractory period in neural networks of biological systems.

These refractory periods are said to contribute to suppressing and controlling the propagation of excitation waves in the neural networks of biological systems. (Toshi Seri: Mathematics of Neural Networks - Information Processing Style of the Brain -; Sangyo Tosho) On the other hand, the inventors believe that these refractory periods temporarily balance the effects of each feature of the input data. Considering this as a changing factor, we changed the weight of a neuron firing in response to a certain input as shown in Figure 5, giving the same effect as the absolute refractory period and the relative refractory period.

That is, the respective weights w1 to w of the neurons fired by the association at a certain point to in the learning process are then reduced to a development small value Wain (for example, 0) for a certain period of time tl.

This corresponds to an absolute refractory period where the threshold value becomes infinite. Even during this absolute refractory period, learning continues and changes the weights of neurons. The neuron to which the weight change is applied at this time is a neuron that did not fire in response to that input before to. After the period t, each weight W~w
7 gradually increases to the value at time to.

This period of gradual increase corresponds to a relative refractory period in which the threshold value θ gradually decreases, and this period is designated as tl.

In the absolute refractory period, only the neurons that did not fire at time to have their weight strengthened, and learning is performed about the biased neuron distribution, but in the relative refractory period,
The neurons that fired at that point also gradually start firing, and balanced corrective learning is gradually performed for the entire neural network.

As a result, an escape from the so-called local minimum is realized, and learning is performed to bring the standard-political change to the optimal value (global local minimum).

Note that, as shown in FIG. 6, by applying the weight change cycle a plurality of times, escape from the local minimum becomes more reliable. However, if the number of weight change cycles is too large, the convergence of learning will be hindered, and the convergence of learning will also be hindered when a weight change cycle is given at the end of learning.

In other words, the weight change cycle needs to be given an appropriate number of times at a relatively early stage of learning. Also, the absolute refractory period t
a, the period t2 of the relative refractory period, and the balance between the two also need to be optimized.

Next, a data processing apparatus in which the above optimization method is effectively executed will be described.

In FIG. 7, the data processing device has a plurality of neural layers NL each having a plurality of neurons N arranged in parallel, and the neural layers are configured such that the output of one neural layer becomes the input of the next neural layer. has been done.

In such a tank bottom, the topology of the neural network can be defined, and the coordinates of each neuron can be specified as shown in FIG.

Here, the X-axis is taken in the direction from the input side to the output side, and the Y-axis is taken in the width direction of each neural layer.

Then, the threshold distribution of neurons in the initial state of the data processing device is set as shown in FIG. That is, the threshold value θ of neurons in the same neural layer is set to be the same, and the threshold value θ is increased as the neural layer becomes later. The gradient of increase in the threshold value is assumed to change stepwise.

That is, the threshold value has, for example, the same magnitude for three consecutive neural layers and increases for every third neural layer. The rate of increase may be uniform over the entire neural network, but it may be larger at later stages as shown in FIG.

In the normal learning process, the weight of synapses with significant synaptic input is increased in response to correct input, but in the threshold distribution shown in Figure 8, initially only neurons closer to the input side fire, and gradually later neurons fire. Neurons have a wider firing range.

In this process, introducing the weight minimization process described above has the effect of increasing the number of firing neurons for a given input in each neural layer, and by the time the final neural layer starts firing, many neurons are fired. Neurons become involved in data processing and do not converge on a biased firing pattern. Therefore, it is possible to prevent a drop to the local minimum.

FIG. 9 shows an initial threshold distribution of the second embodiment, in which the threshold of the central neuron of the data processing device is increased, and the threshold is decreased stepwise as the distance from the central neuron increases. In such a data processing device, a firing pattern occurs around the base of the threshold value in the initial stage of learning. However, when the weights are minimized as described above, there are neurons with relatively low thresholds adjacent to the firing pattern, so the distribution of firing neurons gradually expands toward the top of the threshold. As a result, many neurons become involved in data processing at the end of learning, and the firing pattern does not converge to a biased one. Therefore, it is possible to prevent a drop to the local minimum.

FIG. 10 shows a third example of the initial threshold distribution, in which the threshold of the neuron at the median Y coordinate is set to the lowest, and the threshold is increased in a step-like gradient toward the maximum and minimum values of the Y coordinate. There is. In such a data processing device, a firing pattern that passes through the vicinity of the median Y coordinate occurs in the initial stage of learning. However, when the weights are minimized as described above, there are neurons with relatively low thresholds adjacent to the firing pattern, so the distribution of firing neurons expands toward both the maximum and minimum values of the Y coordinate. go. As a result, many neurons are involved in data processing at the end of learning, and the firing pattern does not converge to a biased one. Therefore, falling to the local minimum can be prevented.

FIG. 11 shows a fourth embodiment. This embodiment differs from the first embodiment shown in FIG. 8 in that at the same -Y coordinate, the thresholds of neurons in all neural layers are the same, while in the same neural layer, as the Y coordinate increases, the neuron thresholds are the same. The threshold value θ is increased. The gradient of increase in this threshold value is stepwise. In the threshold distribution shown in FIG. i1, initially only neurons with large Y coordinates fire, and the firing range gradually expands to neurons with small Y coordinates. Here, when the weights are minimized as described above, there are neurons with relatively low thresholds adjacent to the firing pattern, so the distribution of firing neurons spreads even more smoothly. As a result, many neurons will be involved in data processing at the end of learning, preventing convergence to a biased firing pattern and preventing a drop to a local minimum.

FIG. 12 shows a fifth embodiment, in which the thresholds of neurons in all neural layers are the same at the same Y coordinate, while at a specific Y coordinate (for example, the center) in the same neural layer. The threshold value θ in M takes the maximum value, and the threshold value θ decreases stepwise as the Y coordinate becomes further away from this value. In the threshold distribution shown in FIG. 12, neurons on the maximum and minimum Y coordinate side initially fire, and the range of firing gradually expands to neurons on the Y coordinate M side. Here, if the aforementioned weight minimization is performed, since there are neurons with relatively low thresholds adjacent to the firing pattern, the distribution of firing neurons will spread even more smoothly, and a drop to the local minimum will be prevented. Ru.

FIG. 13 shows a sixth embodiment. In this example, the threshold value θ of the neuron decreases as the X and Y coordinates become smaller, and increases in a stepwise manner as the X and Y coordinates increase.In this threshold distribution, initially the
Neurons with smaller coordinates fire, and the firing range gradually expands to neurons with larger X and Y coordinates. Here, when the weights are minimized as described above, there are neurons with relatively low thresholds adjacent to the firing pattern, so the distribution of firing neurons spreads even more smoothly, preventing a drop to the local bias. be done.

FIG. 14 shows a seventh embodiment, in which the initial threshold distribution is such that the threshold of the central neuron is the minimum value, and the threshold increases as the distance from the central neuron increases. In this example, a firing pattern is generated from the central neuron at the beginning of learning. Here, when the aforementioned weight minimization is performed, since there are neurons with relatively low thresholds adjacent to the firing pattern, the distribution of firing neurons gradually expands toward surrounding neurons. As a result, at the end of learning, many neurons become involved in data processing, preventing convergence to a biased firing pattern.
This will prevent you from becoming depressed about your local lifestyle.

FIG. 15 shows the eighth embodiment, where Y is the initial threshold distribution.
Regarding the coordinates, for example, the center has a high threshold value, and regarding the X coordinate, for example, the center has a low threshold value. That is, the threshold distribution has a saddle shape, and the change in the distribution is step-like. This embodiment also provides the same effects as those of the above embodiments.

〔Effect of the invention〕

As described above, the method for optimizing a data processing device according to the present invention has the excellent effect of reliably escaping the local minimum.

[Brief explanation of drawings]

Figure 1 is a graph showing the change in standard-achievement when appropriate learning is performed. Figure 2 is a graph showing the change in standard-achievement when inappropriate learning is performed. Figure 3. is a graph showing the change in standard-to-adult when it falls to a local minimum during the learning process, Figure 4 is a conceptual diagram showing an example of the firing state of a neural network that falls to a local minimum, and Figure 5 is A graph showing weight changes according to one embodiment of the method of the present invention, FIG. 6 a graph showing weight changes according to another embodiment, FIG. 7 a conceptual diagram showing an example of a neural network of a data processing device, and FIG. 9 is a graph showing the initial threshold distribution of the second embodiment. FIG. 10 is a graph showing the initial threshold distribution of the third embodiment. FIG. 12 is a graph showing the initial threshold distribution of the fifth embodiment. FIG. 13 is a graph showing the initial threshold distribution of the sixth embodiment. FIG. 14 is the seventh embodiment. FIG. 15 is a graph showing the initial threshold distribution of the eighth embodiment.

Claims (1)

[Claims] (1) Data processing in which multiple neurons are provided that output data according to the comparison result of the sum of input data multiplied by a predetermined weight and a threshold value, and the threshold values of the neurons have a distribution that changes stepwise in the initial state. A method for optimizing a device in which the weight of a neuron that has produced a significant output at a certain point is forcibly changed so as to temporarily decrease to a minimum value and then increase to the value at that point. A method for optimizing a data processing apparatus, characterized in that the weights are corrected based on output evaluation results while giving constant input data within a period of forced change.