# US 11,694,678 B2 · Signal processor employing neural network trained using evolutionary feature selection

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 11,694,678 B2
- Document ID: US11694678B2
- Status: Granted
- Kind: B2
- Application number: 17/064,807
- Filed: 2020-10-07
- Published: 2022-04-07
- Granted: 2023-07-04
- Earliest claimed: 2020-10-07
- Assignees: General Dynamics Mission Systems, Inc.
- Applicants: General Dynamics Mission Systems, Inc.
- Inventors: David Lee; Scott Blanchard; Nickolas Dodd
- Web page: https://patentgrep.com/patent/US11694678B2

## CPC

- G10L 15/16: using artificial neural networks
- G10L 25/78: Detection of presence or absence of voice signals (switching of direction of transmission by voice frequency in two-way loud-speaking telephone systems H04M9/10)
- G06F 18/2111: by using evolutionary computational techniques, e.g. genetic algorithms
- G06N 3/086: using evolutionary algorithms, e.g. genetic algorithms or genetic programming
- G10L 15/063: {Training}
- G06N 3/048: Activation functions
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

The evolutionary feature selection algorithm is combined with model evaluation during training to learn feature subsets that maximize speech/non-speech distribution distances. The technique enables ensembling of low-cost models over similar features subspaces increases classification accuracy and has similar computational complexity in practice. Prior to training the models, feature analysis is conducted via an evolutionary feature selection algorithm which measures fitness for each feature subset in the population by its k-fold cross validation score. PCA and LDA based eigen-features are computed for each subset and fitted with a Gaussian Mixture Model from which combinations of feature subsets with Maximum Mean Discrepancy scores are obtained. During inference, the resulting features are extracted from the input signal and given as input to the trained neural networks.

## Claims

### Claim 1

A method for using a neural network to process speech signals, comprising:
providing a corpus of data containing speech and non-speech;
providing a collection of feature sets, each representing a different way to express a signal to be processed in terms of a feature vector;
performing evolutionary feature selection using a genetic algorithm to evolve from the collection of feature sets, a subset collection of feature sets that maximizes a distribution distance between the speech and the non-speech;
extracting from the corpus of data, features corresponding to the subset collection of feature sets to develop training data;
applying the training data to a neural network to develop a learned model;
performing validation of the learned model and iteratively repeating the performing of the evolutionary feature selection if the learned model does not meet a predefined stopping threshold; and
deploying the learned model in a processing system configured to extract from an input signal features, corresponding to the subset collection of feature sets, to be processed by the neural network based on the learned model.

### Claim 2 (depends on 1)

The method of claim 1 wherein the extracting features to develop training data comprises using a plurality of different speech signal processing application specific models to develop the training data representing each of the plurality of different speech signal processing applications.

### Claim 3 (depends on 1)

The method of claim 1 wherein the performing validation is performed by evaluating k-fold cross validation scores of the learned model.