The Department of Computer Science at Ashoka University fosters a research-driven approach to both foundational and applied areas of computing. Its unique positioning within a liberal arts environment provides a distinctive opportunity to advance core computer science research while also exploring interdisciplinary applications across the natural and social sciences. Our department values the cross-pollination of ideas from diverse fields, not only to solve existing problems but to ask novel research questions that deepen and expand the scope of computer science. Through this integration of foundational knowledge and interdisciplinary engagement, Ashoka’s CS department contributes meaningfully to both the theoretical and societal dimensions of the field.
Centres and Labs affiliated with Department of Computer Science:
The center leverages Ashoka’s interdisciplinary strengths in computer sciences, mathematics, economics, physical and social sciences to foster interdisciplinary…
Adjunct Professor of Computer Science, Ashoka University
Ph.D. University of Utah
Subhashis Banerjee
Head of the Department & Professor of Computer Science, Ashoka University
Ph.D. Indian Institute of Science
Santanu Chaudhury
Dean, Vachani School of Advanced Computing and Professor of Computer Science, Ashoka University
PhD, IIT Kharagpur
Partha Pratim Das
Professor of Computer Science, Ashoka University
PhD IIT Kharagpur
Lipika Dey
Professor of Computer Science, Ashoka University
Ph.D. IIT Kharagpur
Debayan Gupta
Assistant Professor of Computer Science, Ashoka University
Ph.D. Yale University
Mahavir Jhawar
Associate Professor of Computer Science, Ashoka University
Ph.D. Indian Statistical Institute, Kolkata
Sandeep Juneja
Professor of Computer Science, Director, Safexpress Centre for Data, Learning and Decision SciencesAshoka University Visiting ProfessorDepartment of Electrical Engineering, IIT Bombay
Anirban Sen
Assistant Professor of Computer Science, Ashoka University
Ph.D. IIT Delhi
Sandeep Sen
Dean, Faculty and Professor of Computer Science, Ashoka University
Ph.D. Duke University
Aalok Thakkar
Assistant Professor of Computer Science, Ashoka University
Abstract: We study the error rate of LLMs on tasks like arithmetic that require a deterministic output, and repetitive processing of tokens drawn from a small set of alternatives. By analyzing the accumulation of errors in the attention mechanism, we theoretically derive a quantitative two-parameter relationship between the accuracy and the complexity of the task. We empirically verify our formula across a range of tasks and state-of-the art LLMs find excellent agreement between the predicted and observed accuracy in many cases. We also identify deviations in some cases that lead us to interesting insights about the functioning of models. We show how this understanding helps to construct prompts to reduce the error rate.
About the speaker: Suvrat Raju is an Indian physicist whose research focuses on quantum gravity and quantum field theory. Suvrat has worked on black holes, focusing on the ‘information paradox’ around them. He has formulated the Papadodimas-Raju proposal for black holes. Suvrat studied physics at St. Stephen’s college in Delhi University and went on to complete his PhD at the Harvard University. He is currently a professor at the International Centre for Theoretical Sciences of the Tata Institute of Fundamental Research (TIFR).
Abstract: We study the error rate of LLMs on tasks like arithmetic that require a deterministic output, and repetitive processing of tokens drawn from a small set of alternatives. By analyzing the accumulation of errors in the attention mechanism, we theoretically derive a quantitative two-parameter relationship between the accuracy and the complexity of the task. We empirically verify our formula across a range of tasks and state-of-the art LLMs find excellent agreement between the predicted and observed accuracy in many cases. We also identify deviations in some cases that lead us to interesting insights about the functioning of models. We show how this understanding helps to construct prompts to reduce the error rate.
About the speaker: Suvrat Raju is an Indian physicist whose research focuses on quantum gravity and quantum field theory. Suvrat has worked on black holes, focusing on the ‘information paradox’ around them. He has formulated the Papadodimas-Raju proposal for black holes. Suvrat studied physics at St. Stephen’s college in Delhi University and went on to complete his PhD at the Harvard University. He is currently a professor at the International Centre for Theoretical Sciences of the Tata Institute of Fundamental Research (TIFR).