Deepak Gouda

selected publications

1. COMSNETS

   Design and Validation of BlockEval, A Blockchain Simulator

   Gouda, Deepak, Jolly, Shashwat, and Kapoor, Kalpesh

   In 2021 International Conference on COMmunication Systems NETworkS (COMSNETS) 2021

   Abs HTML Code Slides

   Blockchain technology is finding its application in a wide range of areas. Due to the highly decentralized and distributed nature of blockchain technology, it is essential to understand the behaviour of a system before its actual deployment. In this paper, we introduce the design and architecture of our blockchain simulator, BlockEval, which mimics the behaviour of concurrent operations that occur in a real-life blockchain system. We have established the correctness of our simulator by comparing it with an independent model that is built from the real Bitcoin-transactions data using deep-learning techniques. Unlike existing validation procedures which do not scale to large topologies, our method is scalable and efficient in terms of time and compute-infrastructure requirements. The observations made using our simulator are found to match with the results obtained from the model trained on real Bitcoin-transactions data.

2. arXiv

   Parameter-free version of Adaptive Gradient Methods for Strongly-Convex Functions

   Gouda, Deepak, Naveed, Hassan, and Kamath, Salil

   arXiv pre-print Jun 2023

   Abs

   The optimal learning rate for adaptive gradient methods applied to λ-strongly convex functions relies on the parameters λ and learning rate η. In this paper, we adapt a universal algorithm along the lines of Metagrad, to get rid of this dependence on λ and η. The main idea is to concurrently run multiple experts and combine their predictions to a master algorithm. This master enjoys O(d log T) regret bounds.

3. HOTMOBILE ’24

   Demo: On-Device Video Analysis with LLMs

   Jaganathan, Vishnu,  Gouda, Deepak, Arora, Kriti, Aggarwal, Mohit, and Zhang, Chao

   In Proceedings of the 25th International Workshop on Mobile Computing Systems and Applications Feb 2024

   Abs HTML

   We present a new on-device pipeline that efficiently summarizes lecture videos and provides relevant answers directly from a smartphone. We utilize widely accessible tools like OCR and Vosk speech-to-text, coupled with powerful large language models (LLMs), to identify crucial sentences and generate summaries. By harnessing the capabilities of LLMs and the computational power of mobile devices, we fine-tune and quantize BERT and GPT-2 to achieve efficient lecture video summarization and question answering on consumer-grade smartphones like the Pixel 8 Pro. Notably, this approach eliminates the need for cloud APIs, ensuring enhanced user privacy and minimal mobile data usage.https://www.youtube.com/shorts/zwGdONlKays