12 mJ per Class On-Device Online Few-Shot Class-Incremental Learning
YE Wibowo, C Cioflan, TM Ingolfsson… - … , Automation & Test …, 2024 - ieeexplore.ieee.org
Few-Shot Class-Incremental Learning (FSCIL) enables machine learning systems to expand
their inference capabilities to new classes using only a few labeled examples, without …
their inference capabilities to new classes using only a few labeled examples, without …
Processing and learning deep neural networks on chip
GB Hacene - 2019 - theses.hal.science
In the field of machine learning, deep neural networks have become the
inescapablereference for a very large number of problems. These systems are made of an …
inescapablereference for a very large number of problems. These systems are made of an …
Efficient hardware implementation of incremental learning and inference on chip
In this paper, we tackle the problem of incrementally learning a classifier, one example at a
time, directly on chip. To this end we propose an efficient hardware implementation of a …
time, directly on chip. To this end we propose an efficient hardware implementation of a …
Online Training from Streaming Data with Concept Drift on FPGAs
E Roorda, SJE Wilton - 2023 24th International Symposium on …, 2023 - ieeexplore.ieee.org
In dynamic environments, the inputs to machine learning models may exhibit statistical
changes over time, through what is called concept drift. Incremental training can allow …
changes over time, through what is called concept drift. Incremental training can allow …
[PDF][PDF] Efficient Representations for Graph and Neural Network Signals
V Gripon - 2020 - hal.science
Interestingly, when I began my PhD back in 2008, neural networks were almost unanimously
considered obsolete. Their inaptitude to solve real world problems, added to the …
considered obsolete. Their inaptitude to solve real world problems, added to the …
[图书][B] Design Space Exploration and Architecture Design for Inference and Training Deep Neural Networks
Y Qi - 2021 - search.proquest.com
Abstract Deep Neural Networks (DNNs) are widely used in various application domains and
achieve remarkable results. However, DNNs require a large number of computations for …
achieve remarkable results. However, DNNs require a large number of computations for …