Deep AI Engine for E&P Data Mining and Domain Intelligence
Datawiz offers a comprehensive solution for data analytics with multiple data sources, enabling automated sample construction from Datapool. It harnesses over 30 standard machine learning and deep learning algorithms, featuring an automated, customizable training workflow.
Datawiz excels in data analytics for rapid production prediction, fracturing optimization, and sweet spot identification. The mission is to significantly lower the technical barriers for reservoir modeling and forecasting, via correlating exhaustive G&G&E data with well performance systematically.
Pre-trained and large-language models become an important foundation in Datawiz compensating to generic AI algorithms. Reservoir simulation models and their surrogates could be built intelligently and much more efficiently when complete reservoir data reside in Datapool.
Data Fusion and Sampling
Artificial General Intelligence
Artificial General Intelligence
- Multi-resolution and multi-frequency data fusion with use-specified length scale and time window
- Construction of temporal-spatial sample set of specified granularity for integrated data mining
- Observation and physical loss construction for PINN
Artificial General Intelligence
Artificial General Intelligence
Artificial General Intelligence
- 30+ classification, clustering, and regression algorithms with automated model selection and hyper-parameter optimization (AutoML)
- Built-in LLM for augmented contents retrieval and instant, ready-to-use table query
- Convenient application to E&P data analytics
Domain Intelligence
Smart Reservoir Agent
Smart Reservoir Agent
- Reservoir simulation surrogate modeling using a deep CNN/GNN + Transformer architecture
- Data space inversion and Bayesian evidential learning for uncertainty quantification and reduction
- PINN implementation for fast simulation and calibration
Smart Reservoir Agent
Smart Reservoir Agent
Smart Reservoir Agent
- Auto reporting and diagnosis of reservoir performance
- Multi-agent environment to utilize LLM’s capability in combination with domain algorithm APIs
- Reinforcement learning agents for assisted geosteering, fracturing optimization, and artificial lift recommendations