Enabling the Real Step #1 for
AI Software Development Life Cycle
AI systems tend to operate in the darkness of black boxes. Input data is transformed into decisions without much human-readable justification. The journey from data to insights requires a crucial data comprehension step, so the direction of an AI/ML model is fully explainable. Omitting this step can produce a result that may not be actionable, causing the AI project to dead-end.
AI-Rover™ is the full implementation of an explainable AI algorithm on a virtualized notebook, in concert with Zeblok's GPU-powered data lake, Edge connectivity and other components. AI-Rover™ discovers and visually explains patterns in complex datasets as well as the causal relationships underlying those patterns, supporting data analysts in the construction of trust-able decision-making AI models. Zeblok's accelerated data lake enables rapid querying of multiple disparate data sources. AI-Rover™ enables the crucial data comprehension step to properly target the direction of an AI project.
Say goodbye to tedious data exploration and instead gain quick access to the hidden information in your data within an easy to understand visual interface that enables users to intuitively derive subpopulations of data attributes, leading to better predictive insights.
About AI- & HPC-WorkStation
The Zeblok AI-WorkStation is a one stop shop. Whether you are looking to start a multi-user data science project using various open-source AI frameworks, or you are looking for a high-performance computing (HPC) environment to distribute your jobs across a large cluster of computational power, we have you covered.
Powerful Machine Learning / Artificial Intelligence Notebook
Model Training, with ML/DL Pipelines
Out-of-the-box CUDA Optimization on GPU for AI- Frameworks
Supports popular data science language bindings such as R, Scala, Docker and Python
We embrace open-source frameworks and tools
About AI-Data Lake
The Zeblok AI-Data Lake - A high-performance data store that allows you to import, filter, and instantly analyze objects. Our solution is designed for performance, scales up with your data, and can provide industry-grade SSL security and data-redundancy for high availability of data at SSD speeds.
The Zeblok AI-Data Lake handles the tasks of data pipe-lining, analysis, and propagation so that you can focus on what matters most. Easily automate your imports, with large files using various methods, such as the Zeblok Data Browser Portal, Zeblok magic commands, inbuilt S3 REST API, Zeblok Data Lake Desktop App, and HTTP source. Whether you are ingesting Image, video data files coming from IoT sensors at Edge locations through 4G/5G low latency network or uploading large csv files.
About Explainable AI
A common headache for data analysts is how to pick a manageable set of meaningful attributes and features from a list of hundreds or more. Our new approach to data analysis puts an end to this guessing game. Our Explainable AI software discovers and visually explains interesting patterns and causal relations in complex data, supporting data analysts in the construction of trustable decision-making AI models.
Pattern Miner automatically decomposes the data into a manageable set of statistically robust data patterns, each of which can be concisely described with just a few attributes.
Visual Causal Analyst distills correlations that exist among these patterns into a terse set of causal relationships. This process eliminates all spurious correlations and makes it easy to discern true interactions within the data. This is illustrated in Step 3 below.
Data Context Map
The Data Context Map visualizes both patterns and relationships in an integrated and intuitive fashion. Analysis can freely interact with this visual layout to explore new patterns and relationships, creating new subpopulations. Simplifies data analysis by combating information overload.
How It Works
Step 1: Data Ingestion
Enterprise users login through a web browser and access their pre-configured AI-Rover™ Notebook, which includes Zeblok's AI-WorkStation and our exclusive Explainable-AI algorithm.
Users then ingest their data into the AI-Data Lake using their choice of:
Data browser portal
ZBL magic commands
S3 REST API
Data Lake Desktop App
Step 2: Start the Automated Analysis and See the Data Patterns in the Context of Their Relevant Attributes
Our software decomposes high-dimensional input data into a set of independent data patterns. Each pattern consists of data items that behave similarly in terms of a given target variable and are succinctly defined by just a small set of attributes, making them easy to understand. This analysis typically only takes a few minutes.
After the analyst has selected data patterns that appear interesting, the system produces an easy-to-read visual layout of these patterns in the context of their relevant attributes. In the image below, the green and red disks are patterns with unusually high or low values in the target attribute (disk radius encodes pattern size) and the triangles nearby are the attributes that define them.
Step 3: Add the Causal Interactions Among the Data Patterns
Many data analytics packages compute correlations among attributes to indicate relationships. But correlation does not necessarily imply causation. Our causal inference engine establishes true causal relationships, which imply a direction. For example, smoking causes cancer but not vice versa. Our visualization interface conveys the intricate web of causal relationships in a single intuitive map.
Step 4: Use the Identified Patterns & Relationships in Your Application
Our software is available as a Jupyter Notebook plug-in and can be easily integrated into an existing data analyst’s workflow. Simply select one (or more) data pattern in the visualization interface and export it as a data frame into your downstream application, such as training a neural network, decision tree, etc.
More Analytical Tools
See How Patterns Change Over Time to Aid in Predictive Analysis, Form Hypotheses Using Our Pattern and Feature Engineering Suite
Patterns often change over time. Our immigration visualization interface allows analysts to observe these changes and gain valuable insights for predictive analysis. This interface is further boosted into our temporal causal analysis engine.
Use our interactive pattern sculpting interface to view, edit and specify new patterns of interest. Then quickly test the effectiveness of new attributes, and features derived from them by examining their ability to form new patterns and causal relationships.