Daim Armaghan

I am undergradate CS student in the Department of Computer Science at the Lahore University of Management Sciences, Pakistan. My research interests include Artificial Intelliegnce, Software Engineering, Computer Vision, Natural Language Processing, and Deep Learning. I am also a member of the Technology of People Initiative advised by Dr. Murtaza Taj since Summer 2022.

Previously, I was a member of University of Illinois Urbana Champaign's Intelligent CAT Lab, where I was working with Dr. Reyhaneh Jabbarvand in fields of Deep learning and Machine Learning for Software Engineering (ML4SE). Moreover, I worked with Dr. Agha Ali Raza at’C-Salt Lab in various domains (i.e., Deep Learning, Speech Processing and Machine Learning).

Highlights

Oct 26, 2022	Publication Submitted to IEEE, ICASSP 2023
Sep 12, 2022	Bloodlink Mobile Application is Live
August 14, 2022	Research Internship at University of Illinois Urbana-Champaign

Awards and Honors

Jan 14, 2022	Stacks Pakistan Award
Dec 12, 2021	Dean's Honor List 2020-21
Dec 14, 2020	Dean's Honor List 2019-20
Sep 3, 2019	2nd Position in BISE 2017-19
Sep 3, 2016	Top Ranked at High School
Sep 3, 2014	97th percentile in International Kangaroo Math Contest

Teaching Experience

CS-360 Software Engineering, Teaching Assistant, LUMSSpring 2023 ’
CS-100 Computational Problem Solving , Teaching Assistant, LUMS Fall 2021’

There have been many recent developments in the use of Deep Learning Neural Networks for fire detection. In this paper, we explore an early warning system for the detection of forest fires. Due to the lack of sizeable datasets and models tuned for this task, existing methods suffer from missed detection. In this work, we first propose a dataset for the early identification of forest fires through visual analysis. Unlike existing image corpora that contain images of widespread fire, our dataset consists of multiple instances of smoke plumes and fire that indicate the initiation of fire. We obtained this dataset synthetically by utilizing game simulators such as Red Dead Redemption 2. We also combined our dataset with already published images to obtain a more comprehensive set. Finally, we compared image classification and localization methods on the proposed dataset. More specifically we used YOLOv7 (You Only Look Once) and different models of detection transformers

Pre-trained models for programming language have achieved dramatic empirical improvements on a variety of code-related tasks such as code search, code completion, code summarization, etc. However, existing pre-trained models regard a code snippet as a sequence of tokens, while ignoring the inherent structure of code, which provides crucial code semantics and would enhance the code understanding process. The research mainly focuses on using Deep learning models like code2seq, code2vec, CodeBERT etc., to utilize the potential of information inside syntactical structure of code.

I researched on Employing Deep learning models for keyword spotting (KWS), in low resource languages like Urdu, the National Language of Pakistan. My research revealed that existing keyword spotting methods are ineffective for languages with smaller resources. This requires use of few-shot and zero-shot learning approaches, which need minimum data. Previously with the same group, I worked on devising ways for semantic analysis and emotion tagging on lower resource languages.

Daim Armaghan

Highlights

Awards and Honors

Teaching Experience

Research

Industry

Projects

Manuscripts under review