Aryan Seth

I am an undergraduate final year student majoring in Computer Science from BITS Pilani, Pilani Campus. At present I am working on my Bachelor's Thesis at DFKI Germany under Prof. Sebastian Vollmer on Information Sharing in Structural Causal Bandits.

I am also presently mentored by Prof. Pratik Narang at BITS, where I am working on Alignment and Fairness in Deep Learning models.

In the past, I have had the pleasure of working under Prof. Asim Tewari where I worked on Transfer Learning to mitigate Chatter in CNC Machines. I also had the opportunity to work under Prof. Arun K Somani where I worked on Training Free Neural Architecture Search for hybrid CNN-Transformer Architectures.

My interests lie in mitigating the risks of Modern Deep Learning. Through this motivation, I work in fairness, model-based privacy, and causal inference. I am also interested in information-theoretic foundations of Machine Learning, and spend time studying allied fields such as Grokking and Unlearning.

Email  /  CV  /  Google Scholar  /  Twitter  /  Github

profile photo

(Hover over the image to magnify it)

* denotes joint first authorship

SynthShield: Leveraging Synthetic Distributions to Enhance Privacy Against Membership Inference

Aryan Seth, Vishnu Hari

Under Review at Privacy Preserving Workshop at AAAI-25

InFER++: Real-World Indian Facial Expression Datase

Syed Sameen Ahmad Rizvi*, Aryan Seth*, Jagat Sesh Challa, Pratik Narang

IEEE Open Journal of the Computer Society, 2024

Paper

Balancing the Scales: Enhancing Fairness in Facial Expression Recognition with Latent Alignment

Syed Sameen Ahmad Rizvi*, Aryan Seth*, Pratik Narang

International Conference on Pattern Recognition-24

Paper

FAIR-FER: A Latent Alignment Approach for Mitigating Bias in Facial Expression Recognition (Student Abstract)

Syed Sameen Ahmad Rizvi*, Aryan Seth*, Pratik Narang

AAAI Conference on Artificial Intelligence

Paper


Website Template borrowed from Jon Barron and Chaitanya Kapoor.