Hi! I’m Prakhar. I’m a Masters of Computing (AI specialization) student at NUS, under Prof. Reza Shokri. I did my BTech in Computer Science and Engineering from IIT Delhi in 2019 and then worked at Advanced Digital Sciences Center, Singapore as a Research Engineer for two years. My research interests involve studying the challenges that arise in machine learning when moving from the sandbox to the real world, including fairness, privacy, adversarial robustness, compression, etc., by uncovering the learning dynamics of neural models. My long term goal is a career in research and academia in the field of AI.

News!

  • [May 2022] I’ll be attending ACL 2022 in Dublin, Ireland! I’ll be presenting our work “Compressing Large-Scale Transformer-Based Models : A Case Study on BERT”.
  • [Jan 2022] I’ll be attending WACV 2022 virtually! I’ll be presenting our work “YOLO-ReT: Towards High Accuracy Real-time Object Detection on Mobile GPUs”.
  • [Nov 2021] We have updated the latest version of our work on Context Adjustment in Co-Saliency Detection. Checkout the paper on arxiv.
  • [Oct 2021] Our paper “YOLO-ReT: Towards High Accuracy Real-time Object Detection on Mobile GPUs” got accepted in WACV 2022.
  • [Sep 2021] Our paper “Compressing Large-Scale Transformer-Based Models : A Case Study on BERT” is now available online at The MIT Press. [link] [pdf]
  • [Sep 2021] Our paper “HiKonv: High Throughput Quantized Convolution With Novel Bit-wise Management And Computation” got accepted in ASP DAC 2022.
  • [June 2021] I’ll be joining NUS MComp (AI specialization) program in August 2021.
  • [June 2021] Our paper “Learning-Based Simultaneous Detection and Characterization of Time Delay Attack in Cyber-Physical Systems” is now available online at IEEE Explore. [link] [pdf]
  • [Oct 2020] We have updated the latest version of our work on Time Series Analysis in High Frequency Trading. Checkout the paper on arxiv.
  • [Sep 2020] We have updated the latest version of our work on Zero-shot Abstractive Summarization for Conversations. Checkout the paper on arxiv.