Posts by Collection

portfolio

Portfolio item number 1

Short description of portfolio item number 1

Portfolio item number 2

Short description of portfolio item number 2

publications

Software and firmware co-development using high-level synthesis

Published in Topical Workshop on Electronics for Particle Physics (TWEPP), 2016, 2016

This paper presents a case study of using high-level synthesis tools to streamline the firmware development process of CMS Endcap Muon Level-1 Track finder (EMTF) at CERN’s Large Hadron Collider.

Download here

Compiling KB-sized machine learning models to tiny IoT devices

Published in ACM SIGPLAN Conference on Programming Language Design and Implementation (PLDI), 2019

This paper presents SeeDot, a domain-specific language/compiler for efficiently translating ML inference into fixed-point code for execution on resource-constrained microcontrollers and FPGAs.

Download here

MAFIA: Machine Learning Acceleration on FPGAs for IoT Applications

Published in International Conference on Field-Programmable Logic and Applications (FPL), 2021

This paper describes a novel compilation tool that enables efficient ML inference on small form-factor FPGAs for IoT applications.

Download here

T-RecX: Tiny-Resource Efficient Convolutional neural networks with early-eXit

Published in International Conference on Computing Frontiers (CF23) - BEST PAPER AWARD , 2023

This paper describes a early-exit network architecture optimized for tinyML models, and presents a novel resource-efficient method to mitigate overthinking in tiny neural networks. This work recieved the ACM Best paper award.

Download here

Enabling Risk Management of Machine Learning Predictions for FPGA Routability

Published in International Symposium on Machine Learning for CAD, 2024

This paper studies the risks of inaccurate predictions of ML models in FPGA CAD algorithms.

Download here

DEBUG-HD: Debugging TinyML models on-device using Hyper-Dimensional computing

Published in Machine Learning for Systems Workshop at NeurIPS, 2024

This paper studies explores on-device debug of field-deployed TinyML models using Hyper-dimensional computing.

Download here

QUTE: Quantifying Uncertainty in TinyML with Early-exit-assisted ensembles for model-monitoring

Published in Under Review, 2024

This paper describes a resource-efficient method to practically quantify uncertainty for tinyML model-monitoring while meeting the strict tinyML constraints.

Download here