International business machines corporation (20240135185). HIGH DIMENSIONAL SURROGATE MODELING FOR LEARNING UNCERTAINTY simplified abstract

HIGH DIMENSIONAL SURROGATE MODELING FOR LEARNING UNCERTAINTY

Organization Name

international business machines corporation

Inventor(s)

Shashanka Ubaru of Ossining NY (US)

Paz Fink Shustin of Tel-Aviv (IL)

Lior Horesh of North Salem NY (US)

Vasileios Kalantzis of White Plains NY (US)

Haim Avron of Tel Aviv (IL)

HIGH DIMENSIONAL SURROGATE MODELING FOR LEARNING UNCERTAINTY - A simplified explanation of the abstract

This abstract first appeared for US patent application 20240135185 titled 'HIGH DIMENSIONAL SURROGATE MODELING FOR LEARNING UNCERTAINTY

Simplified Explanation

The method described in the patent application involves using a variational autoencoder (VAE) to learn a low-dimensional latent space representation of high-dimensional data, and then using a polynomial chaos expansion to map new data samples in the latent space to the corresponding data output. This allows for estimation with high-dimensional datasets under uncertainty, such as missing values, by estimating the values using the set of distributions learned by the VAE.

• Variational autoencoder (VAE) used to learn a low-dimensional latent space representation of high-dimensional data
• Encoder part of the VAE outputs a set of distributions of the high-dimensional dataset in a latent space
• New data samples are sampled in the latent space using the set of distributions from the encoder
• Polynomial chaos expansion is used to map the new data samples in the latent space to the corresponding data output
• Estimation with high-dimensional datasets under uncertainty, such as missing values, is performed using the set of distributions learned by the VAE

Potential Applications

This technology could be applied in various fields such as:

• Finance for risk assessment and portfolio management
• Healthcare for medical image analysis and patient diagnosis
• Manufacturing for quality control and predictive maintenance

Problems Solved

This technology helps in:

• Dealing with high-dimensional datasets with missing values
• Performing estimation under uncertainty
• Learning latent space representations for complex data

Benefits

The benefits of this technology include:

• Improved accuracy in estimation with high-dimensional datasets
• Better handling of uncertainty and missing values
• Efficient learning of latent space representations

Potential Commercial Applications

This technology could be commercially applied in:

• Data analytics software for businesses
• Risk management tools for financial institutions
• Healthcare AI solutions for medical diagnosis

Possible Prior Art

One possible prior art for this technology could be the use of variational autoencoders in machine learning for dimensionality reduction and latent space learning.

Original Abstract Submitted

a method to determine data uncertainty is provided. the method receives a high dimensional data input and a corresponding data output. the method trains a variational autoencoder (vae) with the high dimensional data input to learn a low dimensional latent space representation of the high dimensional data input. an encoder part of the vae outputs a set of distributions of the high dimensional dataset in a latent space. the method samples new data samples in the latent space using the set of distributions outputs from the encoder part of the vae. the method learns a polynomial chaos expansion to map the new data samples in the latent space to the corresponding data output to learn the set of distributions and their relation to perform estimation with high-dimensional dataset under uncertainty such as missing values by estimating the values using the set of distributions.