Scalable Algorithms for Large-Scale Data (ScaleX) Lab

About the Lab:

Recent advances in the  WWW, IoT, social networks, and e-commerce have led to an unprecedented growth in high-dimensional data. Deriving meaningful insights from such data is critical for decision-making across a wide range of real-world applications. Yet, analytics at this scale remains extremely challenging—and at times infeasible—even with access to powerful computational resources.

Our research group addresses this challenge by developing Scalable Algorithmic Techniques for Large-Scale data (ScaleX) that effectively “compress big data into tiny data” with provable mathematical guarantees. These methods ensure that analytics performed on the compressed data closely approximate the results obtained from the original datasets.

Situated at the intersection of theory and practice, our approaches offer strong guarantees of accuracy and efficiency, are computationally practical, and are readily deployable. We further apply these techniques to real-world problems in AI/ML and signal processing, bridging foundational research with impactful applications.

Our Team:

Research Areas:

• Randomized Numerical Linear Algebra,

• Sublinear Algorithms, 

• Compression algorithms for large-scale tensor datasets,

• Application in AI/ML and Signal Processing.

Recent Publications:

• **Efficient and Accurate Tensor Compression via Recursive Sketching. Amit Sharma, Mohammad Azhar Khan, Rameshwar Pratap. Accepted to the 29th Annual Conference on Artificial Intelligence and Statistics (AISTAT 2026).

• It’s All In The (Exponential) Family: An Equivalence Between Maximum Likelihood Estimation and Control Variates For Sketching Algorithms. Keegan Kang, Kerong Wang, Ding Zhang, Rameshwar Pratap, Bhisham Dev Verma, Benedict H. W. Wong. Accepted to the 29th Annual Conference on Artificial Intelligence and Statistics (AISTAT 2026). (Paper Link).

• **Sampling Based Multi-User Detector for Uplink Massive MIMO Communication Networks. Gopal  Chamarthi, Adarsh Patel, Rameshwar Pratap.  Accepted to the IEEE Global Communications Conference (Globecom), 2025. (Paper link).

• Faster and Space-Efficient Indexing for Locality Sensitive Hashing. Bhisham Dev Verma and Rameshwar Pratap. Accepted to the Theoretical Computer Science (TCS) journal, 2025. (Paper link).

• **Random Projection based Fast Multi-User Detectors in Uplink massive MIMO Communication Networks" Gopal Sai Krishna Chamarthi, Adarsh Patel, Rameshwar Pratap. Accepted to the IEEE Open Journal of the Communications Society, 2025. (Paper link).

• **Improving LSH using Tensorized Random Projection. Bhisham Dev Verma, Rameshwar Pratap. Accepted to Acta Informatica journal 2025. (Paper link).

Want to join our Lab:  We are looking for motivated PhD/M.Tech Candidates with a solid foundation in algorithm design and analysis, linear algebra,  probability theory and statistics.  We are especially looking for candidates with a Master's in Mathematics/Applied Mathematics/Statistics or related areas. Candidates with JRF/INSPIRE PhD fellowship in Computer Science, Mathematics, and Statistics (or related areas) are encouraged to reach out via email.