Sabriya Alam

{ Selected Projects }

For our final project, my team and I created a two player colored light pattern repetition game, similar to the classic game Simon.
To play, one player enters a sequenced pattern using colored LED buttons. The pattern replays for the second player, who re-enters the pattern from memory. Points are allocated based on whether the pattern entered matches correctly and how quickly the correct pattern is entered.
An LCD display indicates which player's turn it is to enter a pattern/guess, and shows the current score for each player after his or her turn. A count-up timer is displayed when a guess is being entered, and a counter indicates how many buttons have been pressed. A musical note plays with the push of each LED button.
The game is played for 3 rounds, and at conclusion, the display indicates which player wins or if there is a tie.
The outside of the game is a word cloud created from all the code that was written to program the project. Larger words are those that appeared most frequently in our code. :-)

Key Project Components:

Some children playing our game at the public design showcase we demoed at.

The PCB and all electrical components are housed within the box.

My semester project for this course was about adversarial training for fairness in machine learning. I learned about recent advancements in fair machine learning published in conferences such as NeurIPS, ICML, and IJCAI.
I demonstrated the outcomes of implementing some strategies for fair machine learning in this video presentation.
I initially trained a biased model that put certain groups based on sensitive attributes such as race and sex, at a disadvantage. I quantified this bias, and then used an adversarial training algorithm to counter the unfair predictions. This created a more equitable model, as evaluated through various metrics of fairness.

HONR 399 : Security - Technology & Society

In a semester project, my team and I analyzed the technological, ethical, and societal implications of autonomous weapon systems. As the development of machine learning and artificial intelligence technology increasingly finds applications in defense sectors, it is critical to consider the socio-technical aspects of related policy measures. Our paper seeks to outline and address some of these issues.

This paper resulted in a publication in the Student Papers in Public Policy journal, citation as follows.
Alam, Sabriya; Jimenez, Olivia; Taylor, Julia; and Jengelley, Dwaine (2020) "Unmanned and Autonomous Weapons Systems: Practices and Related Policy," Student Papers in Public Policy: Vol. 2 : Iss. 1, Article 7. Available at https://docs.lib.purdue.edu/sppp/vol2/iss1/7

I engaged in a 48 hour hackathon focused on developing technical solutions for economic mobility issues in Indianapolis, Indiana.
My team developed a web app to help recently incarcerated people ease their transition back to society and become engaged citizens of their communities.
Over the course of the hackathon, we interviewed people who had been formerly incarcerated to learn what their most pressing challenges were in the early days of their re-entry, and we sought to directly address those challenges.

In Purdue's largest hackathon event, my team and I developed an educational tool for sustainability awareness and advocacy.
We trained a machine learning model on a dataset of college students' responses to a survey about their knowledge and lifestyle in terms of sustainable practices. The model predicts whether providing additional information about sustainable practices is likely to be effective in changing a particular user's habits.
Based on this prediction, the web app displays a customized persuasive strategy based on the user's current environmental engagement. This includes a combination of information tailored to the person's lifestyle habits and a carbon footprint calculator for the user to interact with.
The goal of this tool is to help sustainability advocates identify candidates who will be most receptive to their message. This can be used to determine where to most effectively allocate resources and mobilize initiatives.

BoilerMake2020

As a sub-project of my research in Smart Health, a teammate and I created two sets of slides introducing the graduate students on our research team (who are from a different department) to the fundamentals of Python for data analytics and machine learning.

Python Presentation

Neural Network Predictive Modeling

Page updated

Google Sites

Report abuse