November 19.2022

Explainable AI

Presenters:

1. Havva Alizadeh

Ph.D. student Computer Engineering, Ferdowsi University of Mashhad, Member of WTLAB Laboratory.

2. Sahar Ghassabi

MSc student Computer Engineering, Ferdowsi University of Mashhad, Member of WTLAB Laboratory.

3. Abolfazl Mohajeri

MSc student Computer Engineering, Ferdowsi University of Mashhad, Member of WTLAB Laboratory.

Duration: 8:00 Am to 12 Pm (4 Hours)

Date: 2022-11-19

Abstract:

Explainable AI is artificial intelligence in which humans can understand the decisions or predictions made by the AI. In other words, XAI proposes creating a suite of ML techniques that 1) produce more explainable models while maintaining a high level of learning performance (e.g., prediction accuracy), and 2) enable humans to understand andappropriately trust.

This workshop aims to cover some of the most important topics in explainable AI (e.g. information source for explanations, model-intrinsic or model-agnostic approaches, evaluation of explainable models, LIME, SHAP and others) and their applications in important subject areas like recommender systems. In addition, an open source library called Captum will be introduced. Captum helps ML researchers more easily implement interpretability algorithms that can interact with PyTorch models.

In order to register please register in the website and go to services section.

In order to register guideline go to the link.

Dialogue System

Presenters:

1. Mohamed Sobhy

Ph.D. student in Computer Engineering, Amirkabir University of Technology.

2. Mahmoud Soltani

Ph.D. student Computer Engineering, Ferdowsi University of Mashhad, Member of WTLAB Laboratory.

Duration: 14:00 to 18:00 Pm (4 Hours)

Date: 2022-11-19

Abstract:

A dialogue system is a computer program that provides an interface between the user and a computer-based application that permits interaction with the application in a relatively natural manner. The goal of dialogue system is to facilitate user interaction with online services, products or companies. A Person can interact with this online service in natural language. In return, the dialogue agent on the other side of this channel, uses natural language understanding to understand and speak to humans. In general, these systems must understand the users' intention and provide answers related to the user's problem.

In this workshop, we introduce the different types of dialog systems (task-oriented and open-domain) and describe their components. Then we discuss the challenges, datasets used in different components and how to evaluate them. Industrial paradigms as well as the most state-of-the-art architectures in this field (introduced by Google and Facebook) are reviewed. For example, these architectures use sequence-to-sequence deep neural networks and transformers. In the end, in addition to how to implement an industrial chatbot, the implementation of various components of a complex dialogue system (such as natural language understanding: intent detection and slot filling) will be discussed.

In order to register guideline go to the link.

November 20.2022

Human Computing

Presenters:

1.Mohammad Allahbakhsh

Associate Professor at Computer Engineering Department, Ferdowsi University of Mashhad.

Honorary Associate Professor at Macquarie University, Sydney, Australia.

2.Sareh Sadeghianasl

Research Fellow at School of Information Systems, Queensland University of Technology, Brisbane, Australia.

Duration: 8:00 Am to 12 Pm (4 Hours)

Date: 2022-11-20

Abstract:

The presence of computers and artificial intelligence in all aspects of our daily life, from personal to professional and research aspects, is increasing day by day. Although the computer is able to solve complex problems, there are still some tasks that the computer is either not able to do at all, or has a very low accuracy level. Object identification in images, relevance check, and audio transcription are few examples of such tasks. Since human intelligence is used to accomplish these tasks, they are called Human Intelligence Tasks or HIT. The computational model of using human intelligence in doing tasks is called Human Computation.  Crowdsourcing and serious games are examples of human computation.

In this workshop, we will discuss the basic concepts of human computation. We will also study its most important aspects, especially in terms of work breakdown and aggregation techniques. Moreover, we will investigate quality control as the most critical concern in the application of human computation. Gamification will be studied as another important method of applying human computation.

In order to register guideline go to the link.

Computational Psychology

Presenters:

1. Mahsa Khorasani

Ph.D. student Computer Engineering, Ferdowsi University of Mashhad, Member of WTLAB Laboratory.

2. Reza Saeedi

Ph.D. student Computer Engineering, Ferdowsi University of Mashhad, Member of WTLAB Laboratory.

Duration: 4 Hours.

Duration: 14:00 to 18:00 (4 Hours)

Date: 2022-11-20

Abstract:

Computational psychology is an interdisciplinary field that bridges the gap between psychology and artificial intelligence. It aims to build representations of mental processes in people. A wide range of topics and methodologies are covered in the application of computational concepts to the understanding of human behavior. Among them, we can mention the increasing desire of researchers to use machine learning methods to study and predict mental disorders based on clinical or massive data sets such as real-time activity of social media users.

In order to register guideline go to the link.