Introduction to AI Governance
AI governance introduction:
You worked hard to develop a collection of AI models and now you are ready to deploy them.
- How do we sense and interpret a poorly working AI system? Presumably, you have developed a high-quality decision-making engine.
- Your AI system is deployed in a 24X7 environment where users are using it round the clock. What is your mechanism to fix the problems without tearing down the system. Can you develop a continuous learning module of your system which constantly learns from use, adjusts the model and provides sufficient logs to your contributing experts to monitor its learning.
- The model starts to learn and now your experts are baffled. The learning is exactly not what they designed in the original system and now as they redesign their model, they have differences in opinion. How do you reconcile these differences and bring the best of your corporation.
- Is there a tool or a system for life cycle management and can you employ that tool to monitor your AI models like the way you developed process monitoring tools for your governed processes.
You may have many more questions. AI Governance is an evolving topic and not much material is available on this side.
AI governance introduction, you will learn
- Does an AI based solution meets its stated objectives?
- How to put control and guard-rails for an AI based solution
- Does a good AI-based solution engages and “Wows” the users?
- What skills and tools are available to address some of the governance challenges?
AI Governance - Components
-
1
IntroductionIn this section we will provide a brief overview of the course with an outline for each section of the course. We will define the anticipated audience, and the expected outcome from the course. We will also introduce the instructor for the AI Governance course.
-
2
Key Driver for AI Governance includeWhat are the Key Drivers for AI Governance
Continuous Learning Process
-
3
AI Governance - Components
You probably have learned, designed or participated in Data or Process Governance. For AI models, governance is as important if not more. The promise of AI is in its ability to learn over time. If the learning process is not properly governed, it will lead to a decay of model quality and effectiveness. Additionally, it can be sabotaged by bad actors. This section introduces AI Governance and its components in order to deliver a tamper-proof and orderly approach to AI model.
-
4
AI Governance - Decision Quality and Compliance component supports
What does AI Governance - Decision Quality and Compliance component support?
Model Measurements
-
5
Continuous Learning Process
Learning is a key capability of any AI system. How do we design for a good learning environment? In this section, we will introduce you to continuous learning process which uses feedback to improve the model. We will examine a couple of different situations for learning and examine design implications.
-
6
How does challenger beats a champion model
How does challenger beats a champion model
Model Management
-
7
Model Measurements
In this section, we will introduce you to measurements we can use for calibrating the quality of decision-making. In conjunction with the experiment design and continuous improvement from last section, these measurements provide us the mechanisms for evaluating and comparing models.
.
-
8
model precision
How do you measure precision
Collaboration Strategies
-
9
Model Management
You worked hard to develop a collection of AI models and now you are ready to deploy them. You have meticulously designed continuous improvement. You have instrumented it with a good set of measurements. Now how do you process through its production runs so it works perfectly.
Even more important is the situation where you inherited a set of AI models. They look good but not designed for production and governance. How would you bring them to your governed environment and monitor their progress?
Whether you are a data scientist or an expert, this is an important section for you. Without proper model management, your model may start with a bang, but decay over time, or get corrupted by poor learning.
-
10
Role of Librarian
What is the role of Model Librarian?
Summary
-
11
Collaboration Strategies
This is an important aspect of AI Governance as it addresses the role of organizations in decision-making as well as enables human organizations to scale well beyond people collaboration. In addition to the strategies, we will also explore how governance establishes the people roles and processes for the development of collaborative intelligence.
-
12
An example of Smarter Hierarchies
What is an example of Smarter Hierarchies?
Related Courses
