publications

2025

1. NeurIPS

   OmniDraft: A Cross-vocabulary, Online Adaptive Drafter for On-device Speculative Decoding

   Ramchalam Kinattinkara Ramakrishnan, Zhaocong Yuan, Shaojie Zhuo, and 4 more authors

   In Advances in Neural Information Processing Systems, 2025

   Abs PDF

   Speculative decoding generally dictates having a small, efficient draft model that is either pretrained or distilled offline to a particular target model series, for instance, Llama or Qwen models. However, within online deployment settings, there are two major challenges: 1) usage of a target model that is incompatible with the draft model; 2) expectation of latency improvements over usage and time. In this work, we propose OmniDraft, a unified framework that enables a single draft model to operate with any target model and adapt dynamically to user data. We introduce an online n-gram cache with hybrid distillation fine-tuning to address the cross-vocabulary mismatch across draft and target models; and further improve decoding speed by leveraging adaptive drafting techniques. OmniDraft is particularly suitable for on-device LLM applications where model cost, efficiency and user customization are the major points of contention. This further highlights the need to tackle the above challenges and motivates the \textit“one drafter for all” paradigm. We showcase the proficiency of the OmniDraft framework by performing online learning on math reasoning, coding and text generation tasks. Notably, OmniDraft enables a single Llama-68M model to pair with various target models including Vicuna-7B, Qwen2-7B and Llama3-8B models for speculative decoding; and additionally provides up to 1.5-2x speedup.

2. arXiv

   Edge-ASR: Towards Low-Bit Quantization of Automatic Speech Recognition Models

   Chen Feng, Yicheng Lin, Shaojie Zhuo, and 4 more authors

   2025

   PDF

2024

1. NeurIPS

   Stepping Forward on the Last Mile

   Chen Feng, Shaojie Zhuo, Ramchalam Kinattinkara Ramakrishnan, and 3 more authors

   In Advances in Neural Information Processing Systems, 2024

   PDF

2022

1. TinyML Symposium

   An Empirical Study of Low Precision Quantization for TinyML

   Shaojie Zhuo, Hongyu Chen, Ramchalam Kinattinkara Ramakrishnan, and 5 more authors

   2022

   PDF

2020

1. arXiv

   Differentiable Mask for Pruning Convolutional and Recurrent Networks

   Ramchalam Kinattinkara Ramakrishnan, Eyyüb Sari, and Vahid Partovi Nia

   2020

   PDF
2. arXiv

   Tensor train decompositions on recurrent networks

   Alejandro Murua, Ramchalam Kinattinkara Ramakrishnan, Xinlin Li, and 2 more authors

   CoRR, 2020

   PDF

2019

1. ICIAR

   Deep Demosaicing for Edge Implementation

   Ramchalam Kinattinkara Ramakrishnan, Jui Shangling, and Vahid Partovi Nia

   In Image Analysis and Recognition, 2019

   Abs PDF

   Most digital cameras use sensors coated with a Color Filter Array (CFA) to capture channel components at every pixel location, resulting in a mosaic image that does not contain pixel values in all channels. Current research on reconstructing these missing channels, also known as demosaicing, introduces many artifacts, such as zipper effect and false color. Many deep learning demosaicing techniques outperform other classical techniques in reducing the impact of artifacts. However, most of these models tend to be over-parametrized. Consequently, edge implementation of the state-of-the-art deep learning-based demosaicing algorithms on low-end edge devices is a major challenge. We provide an exhaustive search of deep neural network architectures and obtain a Pareto front of Color Peak Signal to Noise Ratio (CPSNR) as the performance criterion versus the number of parameters as the model complexity that outperforms the state-of-the-art. Architectures on the Pareto front can then be used to choose the best architecture for a variety of resource constraints. Simple architecture search methods such as exhaustive search and grid search requires some conditions of the loss function to converge to the optimum. We clarify these conditions in a brief theoretical study.

2018

1. 

   BIBE

   RLALIGN: A Reinforcement Learning Approach for Multiple Sequence Alignment

   Ramchalam Kinattinkara Ramakrishnan, Jaspal Singh, and Mathieu Blanchette

   In 2018 IEEE 18th International Conference on Bioinformatics and Bioengineering (BIBE), 2018

   PDF
2. Masters Thesis

   Exploring Reinforcement Learning Techniques in Multiple Sequence Alignment

   Ramchalam Kinattinkara Ramakrishnan

   2018

   PDF
3. 

   BIBE

   [Regular Paper] Detection of Errors in Multi-genome Alignments Using Machine Learning Approaches

   Jaspal Singh, Ramchalam Kinattinkara Ramakrishnan, and Mathieu Blanchette

   In 2018 IEEE 18th International Conference on Bioinformatics and Bioengineering (BIBE), 2018

   PDF