Microsoft technology licensing, llc (20240338559). CAUSAL DISCOVERY AND MISSING VALUE IMPUTATION simplified abstract
Contents
CAUSAL DISCOVERY AND MISSING VALUE IMPUTATION
Organization Name
microsoft technology licensing, llc
Inventor(s)
Miltiadis Allamanis of Cambridge (GB)
Simon Loftus Peyton Jones of Cambridge (GB)
Angus James Lamb of East Yorkshire (GB)
Pablo Morales- Álvarez of Granada (ES)
CAUSAL DISCOVERY AND MISSING VALUE IMPUTATION - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240338559 titled 'CAUSAL DISCOVERY AND MISSING VALUE IMPUTATION
The abstract describes a computer-implemented method that involves encoding input vector values into latent vectors using a neural network, determining an output vector using a graph neural network based on causal relationships between variables, and minimizing a loss function by tuning parameters of the neural networks and edge probabilities of the graph.
- Input vector values are encoded into latent vectors using a first neural network.
- The output vector is determined using a second neural network with a graph neural network component.
- The graph neural network is parametrized by a graph with edge probabilities indicating causal relationships between variables.
- The loss function is minimized by tuning the edge probabilities of the graph and parameters of the neural networks.
- The loss function includes a measure of difference between the input vector and the output vector.
Potential Applications: - Predictive modeling in various industries such as finance, healthcare, and marketing. - Anomaly detection in complex systems. - Pattern recognition in image and speech processing.
Problems Solved: - Efficient encoding and decoding of input vectors. - Incorporating causal relationships between variables in predictive modeling. - Minimizing loss function to improve accuracy of predictions.
Benefits: - Improved accuracy and efficiency in predictive modeling. - Better understanding of causal relationships between variables. - Enhanced anomaly detection capabilities.
Commercial Applications: Predictive modeling software for financial institutions to forecast market trends and risks.
Questions about the technology: 1. How does the use of edge probabilities in the graph neural network improve predictive modeling accuracy? 2. What are the potential limitations of this method in real-world applications?
Frequently Updated Research: Stay updated on advancements in graph neural networks and predictive modeling techniques for improved performance.
Original Abstract Submitted
a computer-implemented method comprising: receiving an input vector comprising values of variables; using a first neural network to encode the values of the variables of the input vector into a plurality of latent vectors; determining an output vector by inputting the plurality of latent vectors into a second neural network comprising a graph neural network, wherein the graph neural network is parametrized by a graph comprising edge probabilities indicating causal relationships between the variables; and minimising a loss function by tuning the edge probabilities of the graph, at least one parameter of the first neural network and at least one parameter of the second neural network, wherein the loss function comprises a function of the graph and a measure of difference between the input vector and the output vector
