Research

I research how networked systems (cloud infrastructure, social media platforms, and AI-based technologies such as large language models) impact society through three interconnected lenses:

• Equitable Digital Development - Examining how technology design and implementation can advance inclusive growth in underserved communities, creating sustainable pathways for economic and social advancement while bridging persistent digital divides.

• Information Integrity - Investigating the psychological, social, and technical factors driving misinformation, while developing integrated solutions from evidence-based media literacy programs to advanced deepfake detection systems, all evaluated through robust benchmark datasets.

• Human-Centered Digital Health - Designing systems that enhance rather than replace human elements of healthcare, prioritizing patient agency, healthcare provider capabilities, and health equity through thoughtful integration of technical innovation with policy considerations.

At the Internet, Data, and Society Lab (IDSL) at LUMS, I lead initiatives combining approaches like generative AI and privacy-preserving machine learning with rigorous social science methods, including randomized controlled trials. By bridging computational and human-centered perspectives, my work aims to develop digital technologies that serve diverse communities equitably while addressing critical considerations of accessibility, sustainability, and responsible innovation.

Digital Development & Digital Equity

Keywords: Affordable Web; Entry-level Smartphones; Digital Literacy; Causal Inference; Surveys; Mobile Video; Digital Equity

• [Affordability of Mobile YouTube Ads, TheWebConf 2024]

• [Affordable Web Framework, ACM SIGCOMM 2023]

• [Affordability of Public Service Websites, TheWebConf 2023]

• [Validated Measures of Digital Literacy, Development Engineering 2023]

• [Impact of Memory Pressure on Mobile Video Quality of Experience (QoE), ACM CoNEXT 2022]

• [Causal Impact of Android Go, ACM IMC 2022]

• [Affordable Web Architecture, HotNets 2021]

• [Mobile Web Browsing Under Memory Pressure, SIGCOMM CCR 2020]

• [Google Web Light, WWW 2020]

• [MissIt, CHI 2020]

• [Mobile Video, SIGCOMM 2019p, IMC 2018p]

• [Mobile Web, IMC 2018p]

• [Facebook Free Basics, SIGCOMM CCR 2017 – Best Paper Award]

• [Facebook Free Basics, SIGMETRICS 2017p]

• [Mobile Phone Characteristics in Developing Regions, IMC 2016]

• [Long Distance WiFi, INFOCOM 2014p]

Countering Fake News

Keywords: Misinformation; Deepfakes; Digital (Media) Literacy; Cognitive  Reflection; Causal Inference; Randomized Control Trials; Deep Learning; Surveys

• [Deepfake Defense, ACL 2024]

• [Countering Misinformation through Educational Interventions (RCT in Pakistan), JDE 2023]

• [Dataset Pruning for Efficient Audio Deepfake Detection, INTERSPEECH 2023]

• [Validated Measures of Digital Literacy, Development Engineering 2023]

• [Digital Literacy & Misinformation, Journal of Quantitative Description: Digital Media 2022]

• [Cognitive Reflection & COVID-19 Misinformation, HKS Misinformation Review 2022]

• [Audio Deepfake Detection Under Resource-Constrained Environments, INTERSPEECH 2022]

• [Audio Deepfake Detection Using Attention DNNs, ASRU 2021]

• [Robust Detection of Audio Deepfakes, INTERSPEECH 2021]

• [Randomized Control Trial on Countering Misinformation on Social Media, arXiv 2021]

• [Educational Interventions for Combating Fake News, AEA 2020]

Online Safety, Internet Censorship & Internet Centralization

Keywords: Online Privacy; Censorship Measurements; Online Advertising; Ethics; Privacy-Preserving Communication

• [Government Hosting & Internet Centralization, IMC 2024]

• [Ad Exposure in Child-Oriented Content on YouTube, TheWebConf 2024]

• [Hasaas, ECML PKDD 2023]

• [Censorship Economics, AEA 2020]

• [C-Saw, SIGCOMM 2018, HotNets 2015]

• [Ethics, NSDI 2018p]

• [Advention, HotNets 2017]

AI/Machine Learning & Cloud/Edge Computing

Keywords: Data Pruning & Model Pruning; Active Learning; LLMs; Privacy-Preserving Machine Learning; Flow Scheduling; Tail Latency; Load Balancing; Software-Defined Networking

• [Hybrid Active Learning, COLING 2025]

• [Deepfake Defense, ACL 2024]

• [Data Pruning for Efficient Fine-Tuning, NeurIPS ENLSP 2023]

• [Dataset Pruning for Efficient Audio Deepfake Detection, INTERSPEECH 2023]

• [Data Pruning for Efficient Model Pruning, EMNLP 2023]

• [Hasaas, ECML PKDD 2023]

• [Audio Deepfake Detection Under Resource-Constrained Environments, INTERSPEECH 2022]

• [Audio Deepfake Detection Using Attention DNNs, ASRU 2021]

• [Robust Detection of Audio Deepfakes, INTERSPEECH 2021]

• [FedPrune, arXiv 2021]

• [DAS, CoNEXT 2019]

• [2D, CoNEXT 2018]

• [DRIBS, NSDI 2017p]

• [SAPS, INFOCOM 2017]

• [RANS, HotNets 2016]

• [eSDN, SIGCOMM 2015p]

• [MulBuff, SIGCOMM 2014p]

• [PASE, SIGCOMM 2014] (PASE ns2 simulation code, patch applies to ns-2.34)

• [L2DCT, INFOCOM 2013] (L2DCT ns2 simulation code, patch applies to ns-2.35)

• [Coexistence of Datacenter Transports, ICC 2014]

• [Incast Congestion, Globecom 2015]

• [RACS, ICC 2013]