Workshop Overview:

Explainable AI (XAI) is a critical aspect of building trust and transparency in AI systems. This workshop aims to provide participants with a basic understanding of Explainable AI, its importance, and various techniques to make AI models more interpretable.

Topics covered:

1. Introduction to Explainable AI
2. Techniques for Explainable AI – Hands-On Session
   • Inherent Model Approaches
   • Model-Agnostic Methods
     • Global Model-Agnostic Methods
       • Partial dependence plot
       • Accumulated local effect plots
       • Feature interaction (H-statistic)
       • Functional decomposition
       • Permutation feature importance
       • Global surrogate models
       • Prototypes and criticism
     • Local Model-Agnostic Methods
       • Individual conditional expectation curves
       • Local surrogate models (LIME)
       • Scoped rules (anchors)
       • Counterfactual explanations
       • Shapley values
       • SHAP
3. Challenges and Trade-offs of using XAI techniques
4. Case Studies and Best Practices in XAI
5. Future Trends and Open Questions in XAI

Prerequisites:

Basic knowledge in python and basic knowledge in Machine Learning

(NOTE: Having these prerequisites is mandatory for a productive experience. We do not recommend to register for the workshop if you do not have these requirements)

Software/hardware requirements:

Laptop or Desktop with internet connection.

Resource Person:

Mr. Oshada Senaweera
Lecturer
Department of Statistics, University of Colombo

First come first serve. Limited registration. Send in your application for registration.

Workshop Overview:

This workshop will be conducted in two sessions.

Session 1: Preserving Privacy in Machine Learning.

This session guides you through the essentials of preserving privacy in ML. Explore groundbreaking techniques like Federated Learning, a collaborative way to train models without compromising individual data integrity. We'll also dive into Differential Privacy, a sturdy framework ensuring individual data protection during analyses. This research area is not just intriguing but also offers ample opportunities for exploration, making it an excellent avenue for those keen on understanding and contributing to the evolving landscape of machine learning privacy.

Session 2: Future of Information Retrieval with Large Language Models

This session is designed to give participants a glimpse into information retrieval, focusing on how Large Language Models and vector databases are changing the game. By exploring real- world applications, participants will gain valuable insights on how these technologies are shaping the landscape of information retrieval.

Topics covered:

1. Preserving privacy in Machine Learning
   • Federated Learning
   • Differential Privacy
   • Multimodal Data Privacy
   • Energy Efficiency of the Data
2. Future of Information Retrieval with Large Language Models
   • Introduction to Large Language Models (LLMs)
   • Exploring the role of vector databases in storing and retrieving information.
   • Applications of Large Language Models in Information Retrieval

Resource Person:

Mr. Sachith Dilhara Abeywickrama
Associate Machine Learning Engineer
Rootcode AI

Mohammed Munsif
Associate Machine Learning Engineer,
Rootcode AI

First come first serve. Limited registration. Send in your application for registration.

Prerequisites:

Basic understanding of machine learning concepts.

(NOTE: Having these prerequisites is mandatory for a productive experience. We do not recommend to register for the workshop if you do not have these requirements)

Software/hardware requirements:

Laptop or Desktop with internet connection.