안녕하세요, 인도 Hyderabad 대학 Siba K Udgata 교수의 세미나가 아래와 같이 예정되어 있습니다.

관심 있는 학생들의 많은 참여 바랍니다.

[세미나1] General Purpose Wi-Fi Sensing Application for Various Engineering Problems

1. 일시: 11월 2일(목) 15:00(진리관C 221호)

2. 참여(Zoom): https://yonsei.zoom.us/j/4894367796

3. Abstract: 

Device free Wi-Fi sensing has gained much attention due to its simplicity, low cost, and it requires no additional hardware. The main advantages of Wi-Fi sensing are that it is unobtrusive, can operate through walls, work without lighting, is ubiquitous, and does not require users to carry any additional wearable devices. CISCO estimates that there will be 543 million Wi-Fi hot spots in the whole world by end of 2022 which makes the wi-fi signal availability almost omnipresent. The traditional methods of video and sensor based system suffer from many short comings like acceptability, availability, affordability and moreover privacy concerns. The received Wi-Fi signal characteristics changes with change in the dielectric constant of the medium and other reflections and scattering. These change in the Wi-Fi signal patterns can be exploited to detect various events, environmental conditions in the wireless zone, identify the materials in the wall and also detect faults. Recently Wi-Fi sensing techniques are also being used for measuring the physiological parameters like heartbeat, breathing rate monitoring among a few other things.

This talk is an attempt to demonstrate a device free general purpose Wi-Fi sensing system to track events and recognize activities even through the wall and other materials using the Channel State Information (CSI) values extracted from the received Wi-Fi signals at the receiver end. The received signal characteristics changes with the presence of the human beings, and their activities affect the signal propagation, resulting from reflection and scattering. The activities can be recognized by analyzing the CSI values corresponding to different sub-carriers of the received signal. CSI values contain fine grain information such as amplitude and phase to achieve better sensing accuracy with a unique pattern that can be observed corresponding to each activity and material.

We will present our experience of developing the transmitter and receiver hardware modules together with the necessary software for capturing the CSI from Wi-Fi signals and conducted multiple experiments using the low power, low cost ESP-32 Wi-Fi module for human presence, activity detection, material detection, ambient condition in indoor environments. The results show the effectiveness of the proposed method with accurate event detection. 

[세미나2] Explainable AI(XAI) Models for Informed Decision Making

1. 일시: 11월 3일(금) 10:00(진리관C 221호)

2. 참여(Zoom): https://yonsei.zoom.us/j/4894367796

3. Abstract: 

The lack of transparency and interpretability in Artificial Intelligence (AI) models can hinder users' trust and dependency in these systems. Explainable AI (XAI) is an emerging field in machine learning that addresses how decisions are made in an AI system that otherwise considered as a black box. It has emerged as a solution to address this challenge by enhancing the understandability and trustworthiness of AI systems. This area inspects and tries to understand the steps and models involved in making decisions.

This talk will discuss the various issues involved in traditional AI system and also how different explainable AI models are proposed to address these issues. We will discuss mainly the LIME (Local Interpretable Model-Agnostic Explanations) and SHAP (SHapley Additive exPlanations) and how these models can be used to increase the transparency in the AI systems. We will also discuss the counter factual models for recommending remedial measures in stead of focusing on the cause only.

In conclusion, we will discuss how these models can be used for wearable sensor based human activity recognition system to improve the interpretability and also performance.