Artificial intelligence for robust drug dilution detection
Techniques are provided detecting diluted drugs using machine learning. Measurements and images corresponding to a product are obtained, wherein the product is formulated as a liquid, and wherein the measurements and images capture physical, spectral, optical, and/or chemical properties of the product. The measurements and images are provided to a machine learning model, wherein the machine learning model is trained using data generated from interactive learning modules (e.g., a generative adversarial network). The machine learning model detects whether the product or chemical is a real or counterfeit product. In addition, these techniques may be used by practitioners (e.g., medical personnel dispensing a prescribed dosage of a drug with a specific dilution level) to detect prescription errors at the point of administration.
1. A method of detecting drugs or chemicals that have been diluted using machine learning comprising:
obtaining measurements and images corresponding to a product, wherein the product is formulated as a liquid, and wherein the measurements and images capture properties of the product;
providing the measurements and images to a machine learning model, wherein the machine learning model is trained using data generated from a learning network comprising a plurality of interacting learning models with contradictory objectives, and wherein at least one of the plurality of interacting learning models is trained using information derived from images obtained from diluted drugs corresponding to actual drugs that have been diluted by a diluent at a plurality of dilution points; and
detecting whether the product is a real or counterfeit product.
2. The method of claim 1 , wherein the plurality of interacting learning models comprises a discriminator network and a generator network, and further comprising:
training the discriminator network using training data derived from real drugs, counterfeit drugs, and coarse grained images obtained from the diluted drugs;
generating counterfeit-dilution data by the generator network; and
when the generated counterfeit-dilution data meets specified classification criteria by the generator network, including the generated counterfeit-dilution data in the training data for the discriminator network.
3. The method of claim 2 , further comprising:
providing a noise vector to the generator network; and
injecting noise into data produced by the generator network based on the noise vector.
4. The method of claim 2 , further comprising:
training the discriminator network using training data derived from coarse grained data and fine grained data, wherein the fine grained data is obtained from extrapolating information from the coarse grained data.
5. The method of claim 2 , further comprising:
determining a counterfeit-dilution percentage value for a product;
determining a dosage or an efficacy based on the counterfeit-dilution percentage value; and
when a reference efficacy or a prescribed dosage is less than the determined efficacy or dosage, sending an alert that the product is counterfeit-diluted.
6. The method of claim 1 , further comprising:
performing progressive analysis by analyzing product characteristics corresponding to external non-intrusive data, determining features for internal non-intrusive data, and determining features for internal intrusive data; and
detecting whether the product is a real, counterfeit or counterfeit-diluted product based on the progressive analysis.
7. The method of claim 1 , wherein the data generated by the plurality of interactive learning models are used to train another machine learning model.
8. A system to detect drugs or chemicals that have been diluted using machine learning, the system comprising:
one or more computer processors;
one or more computer readable storage media;
program instructions stored on the one or more computer readable storage media for execution by at least one of the one or more computer processors, the program instructions comprising instructions to:
receive measurements and images corresponding to a product, wherein the product is formulated as a liquid, and wherein the measurements and images capture properties of the product;
provide the measurements and images to a machine learning model, wherein the machine learning model is trained using data generated from a learning network comprising a plurality of interacting learning models with contradictory objectives, and wherein at least one of the plurality of interacting learning models is trained using information derived from images obtained from diluted drugs corresponding to actual drugs that have been diluted by a diluent at a plurality of dilution points; and
detect whether the product is a real or counterfeit product.
9. The system of claim 8 , wherein the plurality of interactive learning models comprises a discriminator network and a generator network, and wherein the program instructions further comprise instructions to:
train the discriminator network using training data derived from real drugs, counterfeit drugs, and coarse grained images obtained from the diluted drugs;
generate counterfeit-dilution data by the generator network; and
when the generated counterfeit-dilution data meets specified classification criteria by the generator network, include the generated counterfeit-dilution data in the training data for the discriminator network.
10. The system of claim 9 , wherein the program instructions further comprise instructions to:
provide a noise vector to the generator network; and
inject noise into data produced by the generator network based on the noise vector.
11. The system of claim 9 , wherein the program instructions further comprise instructions to:
train the discriminator network using training data derived from coarse grained data and fine grained data, wherein the fine grained data is obtained from extrapolating information from the coarse grained data.
12. The system of claim 9 , wherein the program instructions further comprise instructions to:
determine a counterfeit-dilution percentage value for a product;
determine a dosage or an efficacy based on the counterfeit-dilution percentage value; and
when a reference efficacy or a prescribed dosage is less than the determined efficacy or dosage, send an alert that the product is counterfeit-diluted.
13. The system of claim 9 , wherein the program instructions further comprise instructions to:
perform progressive analysis by analyzing product characteristics corresponding to external non-intrusive data, determining features for internal non-intrusive data, and determining features for internal intrusive data; and
detect whether the product is a real, counterfeit or counterfeit-diluted product based on the progressive analysis.
14. The system of claim 8 , wherein the data generated by the plurality of interacting learning models is used to train another machine learning model.
15. A computer program product to detect drugs or chemicals that have been diluted using machine learning, the computer program product comprising one or more computer readable storage media collectively having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to:
receive measurements and images corresponding to a product, wherein the product is formulated as a liquid, and wherein the measurements and images capture properties of the product;
provide the measurements and images to a machine learning model, wherein the machine learning model is trained using data generated from a learning network comprising a plurality of interacting learning models with contradictory objectives, and wherein at least one of the plurality of interacting learning models is trained using information derived from images obtained from diluted drugs corresponding to actual drugs that have been diluted by a diluent at a plurality of dilution points; and
detect whether the product is a real or counterfeit product.
16. The computer program product of claim 15 , wherein the plurality of interactive learning models comprises a discriminator network and a generator network, and wherein the program instructions further comprise instructions to:
train the discriminator network using training data derived from real drugs, counterfeit drugs, and coarse grained images obtained from the diluted drugs;
generate counterfeit-dilution data by the generator network; and
when the generated counterfeit-dilution data meets specified classification criteria by the generator network, include the generated counterfeit-dilution data in the training data for the discriminator network.
17. The computer program product of claim 16 , wherein the program instructions further comprise instructions to:
provide a noise vector to the generator network; and
inject noise into data produced by the generator network based on the noise vector.
18. The computer program product of claim 16 , wherein the program instructions further comprise instructions to:
train the discriminator network using training data derived from coarse grained data and fine grained data, wherein the fine grained data is obtained from extrapolating information from the coarse grained data.
19. The computer program product of claim 16 , wherein the program instructions further comprise instructions to:
determine a counterfeit-dilution percentage value for a product;
determine a dosage or an efficacy based on the counterfeit-dilution percentage value; and
when a reference efficacy or a prescribed dosage is less than the determined efficacy or dosage, send an alert that the product is counterfeit-diluted.
20. The computer program product of claim 16 , wherein the program instructions further comprise instructions to:
perform progressive analysis by analyzing product characteristics corresponding to external non-intrusive data, determining features for internal non-intrusive data, and determining features for internal intrusive data; and
detect whether the product is a real, counterfeit or counterfeit-diluted product based on the progressive analysis.