Azra Abtahi Fahliani: “Sweden is a welcoming environment for international researchers”

Meet the WASP Postdocs

Azra Abtahi Fahliani holds a particular interest in Machine Learning and Internet of Things (IoT) due to their potential to address real-world problems, particularly in healthcare challenges. Her choice of research path aligns with her goal of leveraging technology to create positive impacts on health, environmental sustainability, and overall quality of life.

Sweden is also a welcoming environment, she believes, with its family-friendly policies and open-minded and polite people, where many are proficient in English.

Could you highlight a specific project you are working on within Machine Learning?

One of the projects I am currently working on is Interpretable Machine Learning, which has particular significance in various applications, especially in the health and medical domains. In these areas, understanding how a machine learning model arrives at its decisions is as crucial as the decisions themselves. In the realm of Interpretable Machine Learning, my focus is on seizure detection/prediction, water contaminant analysis, and diagnosis, prognosis, and treatment of colon cancer (Florence EU project).

The EU-funded Florance project sounds interesting. Can you describe it further?

The EU-funded Florance project is a collaboration with DTU – Technical University of Denmark, Oslo University Hospital, Zealand University Hospital, and Netherlands Comprehensive Cancer Organization (IKNL). In this project, a tool based on artificial intelligence is being developed to enhance the diagnosis, prognosis, and treatment of patients with colon cancer.  In this project, I am serving as a Project Manager and Lund University is in the charge of Interpreting and Robustness Evaluation of the tool.  In the realm of interpretability, we are working on providing significant insight into how the model works and how each feature impacts the output.

What do you do within IoT?

Within the scope of IoT, I am immersed in two projects. The first project centers on energy efficient machine learning for distributed IoT sensors and wearable devices. Its goal is to extend battery life and improve overall energy efficiency of distributed IoT sensors and wearable devices. This project involves two key components: firstly, the development of energy efficient machine learning models for distributed IoT sensors and wearable devices, and secondly, the development of strategies for dividing and distributing computation tasks among interconnected IoT devices. The second project in the IoT domain focuses on contact tracing using Bluetooth Low Energy (BLE) IoT devices. The primary aim of this project is to develop protocols that ensure privacy while effectively tracing and managing contacts during critical public health situations such as the COVID pandemic.

Do you have any significant results or findings from your recent research projects?

Yes, there are a substantial number of exciting results from our projects! One of our key results from using interpretable machine learning was shedding light on intricate relationship between cardiac function and epileptic seizures. Another set of interesting findings relate to exploiting interpretable machine learning for water contaminant analysis, where we gained a deeper understanding of the dynamics of Pentachlorophenol—a common, recalcitrant, and toxic groundwater contaminant. Our research illuminated its transport properties in the subsurface and identified the key hydrochemical and hydrogeological drivers affecting its transport and fate.

For the contact tracing using BLE IoT devices-project, we developed new protocols that enhance the identification of high-risk contacts. One protocol, in particular, demonstrated that allowing conservative users—those who do not want to broadcast BLE beacons—to participate in digital automatic contact tracing significantly improves the high-risk contact identification.

Furthermore, our work in energy efficient ML for distributed IoT sensors and wearable devices has resulted in the development of several interesting and innovative models and schemes.

What drives your interest in the specific research fields of Machine Learning and IoT?

I am interested in machine learning and IoT due to the potential they hold for addressing real-world challenges, especially in healthcare challenges. My postdoc projects align with my broader goal of leveraging technology to create positive impacts on health, environmental sustainability, and overall quality of life.

How is your research important?

My research holds significance in addressing contemporary challenges. In the field of IoT, my work on energy-efficient machine learning for distributed IoT sensors and wearable devices and contact tracing using BLE IoT devices, can enhance the healthcare systems. Moreover, in various applications of machine learning, including biomedical application, interpretability is pivotal to ensure that the decisions made by the models are transparent and explainable. This not only fosters trust in the technology but also facilitates its adoption in critical decision-making processes.

How do you experience it to be part of WASP?

Being a part of WASP is a rewarding experience, as it provides me with the resources and collaborative opportunities essential for making meaningful contributions in my research areas. For example, with WASP support, I have had the opportunity to participate in conferences such as Privacy Enhancing Technologies (PETS) and International Conference on Acoustics, Speech, and Signal Processing (ICASSP). This support has enabled me to share my findings with a wider audience, connect with leading experts, and foster valuable collaborations, all of which have significantly advanced my work.

How would you summarize Sweden as a research arena?

A rich tradition of investing in research alongside with a strong emphasis on interdisciplinary collaboration has made Sweden a thriving research arena.

Is there anything specific you like about Sweden, as a country/culture?

The emphasis on work-life balance, parental leave policies, and social support systems make Sweden a family-friendly country, which is what I particularly like most about this country. Moreover, Swedish culture is often characterized by a strong emphasis on equality and inclusivity and the Swedish people are friendly, polite, and proficient in English, making it a welcoming environment for international researchers.