activity
20172026
most citedOn-Chip Communication Network for Efficient Training of Deep Convolutional Networks on Heterogeneous Manycore Systems

79 citations · 194 across the 39 of their papers we have counts for

collaborators
Showing 2024Show all

10 papers · 1 filter

cs.AR2024★ 2 cited

MFIT: Multi-Fidelity Thermal Modeling for 2.5D and 3D Multi-Chiplet Architectures

Lukas Pfromm, Alish Kanani, Harsh Sharma +6

Rapidly evolving artificial intelligence and machine learning applications require ever-increasing computational capabilities, while monolithic 2D design technologies approach thei…

cs.AR2024

HeTraX: Energy Efficient 3D Heterogeneous Manycore Architecture for Transformer Acceleration

Pratyush Dhingra, Janardhan Rao Doppa, Partha Pratim Pande

Transformers have revolutionized deep learning and generative modeling to enable unprecedented advancements in natural language processing tasks and beyond. However, designing hard…

cs.LG2024

Active Learning for Derivative-Based Global Sensitivity Analysis with Gaussian Processes

Syrine Belakaria, Benjamin Letham, Janardhan Rao Doppa +3

We consider the problem of active learning for global sensitivity analysis of expensive black-box functions. Our aim is to efficiently learn the importance of different input varia…

cs.LG2024★ 9 cited

Pareto Front-Diverse Batch Multi-Objective Bayesian Optimization

Alaleh Ahmadianshalchi, Syrine Belakaria, Janardhan Rao Doppa

We consider the problem of multi-objective optimization (MOO) of expensive black-box functions with the goal of discovering high-quality and diverse Pareto fronts where we are allo…

cs.LG2024

Conformal Prediction for Class-wise Coverage via Augmented Label Rank Calibration

Yuanjie Shi, Subhankar Ghosh, Taha Belkhouja +2

Conformal prediction (CP) is an emerging uncertainty quantification framework that allows us to construct a prediction set to cover the true label with a pre-specified marginal or…

cs.AR2024

Look-Up Table based Neural Network Hardware

Ovishake Sen, Chukwufumnanya Ogbogu, Peyman Dehghanzadeh +4

Traditional digital implementations of neural accelerators are limited by high power and area overheads, while analog and non-CMOS implementations suffer from noise, device mismatc…