Скачать или смотреть Google TensorFlow Quantum | MIT TorchQuantum; Medical QNNs

Google's TensorFlow Quantum and MIT's TorchQuantum are quantum machine learning open source libraries for prototyping of hybrid quantum-classical ML models. These developments are based on TensorFlow and PyTorch and maintain user-friendliness, but now allows for incorporation of quantum circuits/algorithms run on simulators or real hardware. Format = Fireside Chat.

TensorFlow Quantum by Google:
The 56 page introduction titled "TensorFlow Quantum: A Software Framework for Quantum Machine Learning" is supported by 25 organizations and includes image analysis opportunities for Quantum Neural Networks (QNNs). The authors state "Research in quantum algorithms and applications can leverage Google’s quantum computing frameworks, all from within TensorFlow." [1-2]

TorchQuantum by MIT:
Author notebooks are available to run and revise through the torchquantum github page. An MIT application of the library includes "TorchQuantum Case Study for Robust Quantum Circuits" by Wang, H. et al., 2022. The authors' page states the open source library allows for "Easy construction of parameterized quantum circuits such as Quantum Neural Networks in PyTorch" which includes "batch mode inference and training on GPU/CPU, [and] supports highly-parallelized training" [3-5]

Background:
Deep Learning (Deep Neural Networks) saw increased development with the advent of new computer hardware in the early 2000s, essentially allowing for more advanced forms of machine learning applications. A popular area in Deep Learning includes Convolutional Neural Networks (CNNs), which have proved valuable for medical image analysis. Shen, D., et al. 2017; Razzak, M.I., et al. 2017; Arterys Inc., FDA 510(k) 2017 (Deep Learning) [6-8]

Quantum Advancements:
Quantum Deep Learning: 1st Generation hybrid quantum-classical models for medical images have featured conventional CNNs with small supplementary quantum circuits run on simulators for state preparation and measurement. At least seven articles have reported an increase in binary or multi-class classification accuracy improvements of up to 5% using this model type. ChemicalQDevice, video 2023. Next generation hybrid neural networks will likely incorporate additional facets of quantum code to further improve performance and/or efficiency with emerging hardware that utilizes promising libraries such as Google's TensorFlow Quantum and MIT's TorchQuantum. [9-10]

References:
[1] TensorFlow Quantum https://lnkd.in/gdpB8azz
[2] TensorFlow Quantum https://lnkd.in/g4uHx6mT
[3] TorchQuantum https://lnkd.in/gTwpZE3h
[4] TorchQuantum https://lnkd.in/gWxs3iDy
[5] TorchQuantum https://qmlsys.mit.edu/
[6] CNN Medical Images https://bit.ly/3UU1dKr
[7] CNN Medical Images https://bit.ly/3mV9wsM
[8] FDA 510(k) Approval https://lnkd.in/gzVEVwnV
[9] Quantum Hybrids https://bit.ly/41rUI3L
[10] Medical Images Future https://bit.ly/3oq6xZE

#google #tensorflowquantum #tensorflow #mit #torchquantum #pytorch