17808927. GLOBAL CONTEXT EXPLAINERS FOR ARTIFICIAL INTELLIGENCE (AI) SYSTEMS USING MULTIVARIATE TIMESERIES DATA simplified abstract (INTERNATIONAL BUSINESS MACHINES CORPORATION)
GLOBAL CONTEXT EXPLAINERS FOR ARTIFICIAL INTELLIGENCE (AI) SYSTEMS USING MULTIVARIATE TIMESERIES DATA
Organization Name
INTERNATIONAL BUSINESS MACHINES CORPORATION
Inventor(s)
Diptikalyan Saha of Bangalore (IN)
Amaresh Rajasekharan of Wappingers Falls NY (US)
Shengrong Tang of Dublin OH (US)
GLOBAL CONTEXT EXPLAINERS FOR ARTIFICIAL INTELLIGENCE (AI) SYSTEMS USING MULTIVARIATE TIMESERIES DATA - A simplified explanation of the abstract
This abstract first appeared for US patent application 17808927 titled 'GLOBAL CONTEXT EXPLAINERS FOR ARTIFICIAL INTELLIGENCE (AI) SYSTEMS USING MULTIVARIATE TIMESERIES DATA
Simplified Explanation
The patent application describes techniques for explaining the behavior of Artificial Intelligence systems that use multivariate timeseries data. Here are the key points:
- Predictions are made for multivariate timeseries data.
- Feature importance weights are generated using a feature-based local explainer.
- Each feature importance weight is associated with a time period and a corresponding data source of the timeseries data.
- A dataset is created using the feature importance weights, with labels indicating whether the weight is positive or negative for each time period and data source.
- One or more global explanations are generated using the dataset and a rule-based explainer.
- The global explanations show how the predictions change at specific times in the timeseries data based on values from the corresponding data source.
- An action can be taken based on the global explanations.
Potential applications of this technology:
- Understanding the behavior of AI systems that use multivariate timeseries data.
- Identifying important features and their impact on predictions.
- Debugging and improving the performance of AI systems.
- Enhancing transparency and interpretability of AI systems.
Problems solved by this technology:
- Lack of transparency in AI systems using multivariate timeseries data.
- Difficulty in understanding the factors influencing predictions.
- Limited interpretability of AI systems using complex data.
- Challenges in debugging and improving the performance of AI systems.
Benefits of this technology:
- Provides explanations for the behavior of AI systems using multivariate timeseries data.
- Helps identify important features and their impact on predictions.
- Enhances transparency and interpretability of AI systems.
- Facilitates debugging and improvement of AI system performance.
Original Abstract Submitted
Provided are techniques for global context explainers for Artificial Intelligence systems using multivariate timeseries data. Predictions for multivariate timeseries data are received. Feature importance weights are generated from the predictions using a feature-based local explainer, where each of the feature importance weights is associated with a time period and a corresponding data source of timeseries data of the multivariate timeseries data. A dataset is generated using the feature importance weights, where the dataset includes, for each time period and the corresponding data source, a label indicating whether the feature importance weight is one of positive and negative. One or more global explanations are generated using the dataset and a directly interpretable rule-based explainer, where the one or more global explanations indicate how the predictions change at particular times in the multivariate timeseries data based on values from the corresponding data source. An action based on the global explanations is performed.
