Methods and systems for treating a heterogeneous mixture of materials
AssigneeSixone Labs Ltd.
InventorsJin Suntivich, Christopher Wai
Life of this patent
Estimating the term…
Abstract
A system and method for classifying a material in a heterogeneous mixture of materials, and predicting a processing method for treating the material are disclosed. At least some of the heterogeneous mixture of material pieces comprise different chemical information such as different chemical composition and/or physical characteristics. Embodiments of the system and method of the present invention use machine learning to classify and predict a processing protocol for material pieces based on the chemical compositions or reactivities of each of the material pieces.
Claims28
A method for predicting a processing protocol for treating a heterogeneous mixture of desired materials, comprising:
- performing a plurality of experiments on a control set of materials, each one of the plurality of experiments comprising:
- acquiring a materials information about the control set of material by one or more sensors;
- treating the material using a processing protocol from a plurality of processing protocols that identifies conditions for processing materials;
- obtaining an output of the treatment;
- and
- analyzing the output to obtain training results of the experiments;
- creating a machine-learning model by training a machine-learning algorithm using the materials information, the processing protocol, and the training results from each of the plurality of experiments;
- receiving an input for a desired material for treating wherein the input comprises a materials information about the desired material obtained by the one or more sensors;
- creating, by the machine-learning model, a processing protocol for treating the desired material based on the input, wherein the creating of the processing protocol for treating the desired material comprises dynamically mapping the desired material to a newly created category to predict the processing protocol for treating the desired material;
- treating the desired material using the processing protocol for treating the desired material;
- obtaining an output from treating the desired material;
- analyzing the output to obtain further training results;
- and
- updating the machine-learning model by further training the machine-learning algorithm using the input, the processing protocol for treating the desired material, and the further training results.
A method for predicting a processing protocol for treating a heterogeneous mixture of desired materials, comprising:
- receiving an input for a desired material for treating, wherein the input comprises a materials information about the desired material obtained by one or more sensors;
- creating, by a machine-learning model, a processing protocol for treating the desired material based on the input,
- wherein the creating of the processing protocol for treating the desired material comprises dynamically mapping the desired material to a newly created category to predict the processing protocol for treating the desired material,
- wherein the machine-learning model is created by training a machine-learning algorithm using a materials information, a processing protocol, and training results obtained from analyzing an output of a treatment from a plurality of experiments performed on a control set of materials;
- treating the desired material using the processing protocol for treating the desired material;
- obtaining an output from treating the desired material;
- analyzing the output to obtain further training results;
- and
- updating the machine-learning model by further training the machine-learning algorithm using the input, the processing protocol for treating the desired material, and the further training results.
Dependencies are inferred from the claim text.