Multi-Task Learning by Multi-Wave Optical Diffractive Network

Jing Su; Yafei Yuan; Chunmin Liu; Jing Lia

Multi-Task Learning by Multi-Wave Optical Diffractive Network

Jing Su * §

Contact

Yafei Yuan

Contact

Chunmin Liu

Contact

Jing Lia

Contact

Fudan University

$Multi-Task Learning by Multi-Wave Optical Diffractive Network$

Article Fingerprint

ReserarchID

CSTZ0B3B

Multi-Task Learning by Multi-Wave Optical Diffractive Network Banner

AI TAKEAWAY

Connecting with the Eternal Ground

Abstract

Recently, there has been tremendous researches in Optical neural networks that could complete comparatively complex computation by optical characteristic with much more fewer dissipation than electrical networks. Existed neural networks based on the optical circuit are structured with an optical grating platform with different diffractive phase at different diffractive points. In this study, it proposed a multi-wave deep diffractive network with approximately 106 synapses, and it is easy to make hardware implementation of neuromorphic networks. In the optical architecture, it can utilize optical diffractive characteristic and different wavelengths to perform different tasks. Different wavelengths and different tasks inputs are independent of each other. Moreover, we can utilize the characteristic of them to inference several tasks, simultaneously. The results of experiments were demonstrated that the network could get a comparable performance to single-wavelength single-task. Compared to the multi-network, single network can save the cost of fabrication with lithography. We train the network on MNIST and MNIST-FASHION which are two different datasets to perform classification of 32*32 inputs with 10 classes. Our method achieves competitive results across both of them. In particular, on the complex task MNIST-FASION, our framework obtains an excellent accuracy improvement with 3.2%. On the meanwhile, MNSIT also have the improvement with 1.15%.

References

23 Cites in Article

Reference Format

J Jewell,H Gibbs,S Tarng,A Gossard,W Wiegmann (1982). Regenerative pulsations from an intrinsic bistable optical device.
J Kraitl,H Ewald,H Gehring (2005). An optical device to measure blood components by a photoplethysmographic method.
Yingying Liu,Xianqing Lin,Cong Wei,Chuang Zhang,Jiannian Yao,Yong Zhao (2019). 3D-printed optical-electronic integrated devices.
Seyoon Kim,Sergey Menabde,Victor Brar,Min Jang (2019). Functional Mid‐Infrared Polaritonics in van der Waals Crystals.
M Spencer,J Eickholt,J Cheng (2014). A Deep Learning Network Approach to ab initio Protein Secondary Structure Prediction.
X Zhou,R Takayama,S Wang,T Hara,H Fujita (2017). Deep learning of the sectional appearances of 3D CT images for anatomical structure segmentation based on an FCN voting method.
Chengqian Che,Ruogu Lin,Xiangrui Zeng,Karim Elmaaroufi,John Galeotti,Min Xu (2018). Improved deep learning-based macromolecules structure classification from electron cryo-tomograms.
X,R Girshick,P Doll¨¢r,Z Tu,K He (2017). Aggregated Residual Transformations for Deep Neural Networks.
Hein Htet Aung (null). Deep iterative convolutional neural network for face image super-resolution.
D Li Deep learning: from speech recognition to language and multimodal processing.
Shashi Singh,Ajai Kumar,Hemant Darbari,Lenali Singh,Anshika Rastogi,Shikha Jain (2017). Machine translation using deep learning: An overview.
T Kwok,Y Kwok On the design, control, and use of a reconfigurable heterogeneous multicore system-on-a-chip.
T Cianciulli,J Lax,F Cerruti,G Gigena,H Redruello,M Orsi,J Gagliardi,A Dorelle,M Riccitelli,H Prezioso (2010). COMPLEMENTARY VALUE OF TRANSTHORACIC ECHOCARDIOGRAPHY & CINEFLUOROSCOPIC EVALUATION OF MECHANICAL HEART PROSTHETIC VALVES.
A Vincent,J Larroque,W Zhao,N Romdhane,O Bichler,C Gamrat,J-O Klein,S Galdin-Retailleau,D Querlioz (2014). Spin-transfer torque magnetic memory as a stochastic memristive synapse.
Alexander Tait,Mitchell Nahmias,Yue Tian,Bhavin Shastri,Paul Prucnal (2013). Photonic Neuromorphic Signal Processing and Computing.
Carlos Ríos,Matthias Stegmaier,Peiman Hosseini,Di Wang,Torsten Scherer,C Wright,Harish Bhaskaran,Wolfram Pernice (2015). Integrated all-photonic non-volatile multi-level memory.
Roelof Hut,N Mrosovsky,Serge Daan (1999). Nonphotic Entrainment in a Diurnal Mammal, the European Ground Squirrel (<i>Spermophilus citellus</i>).
K Voutsas,G Langner,J Adamy,M Ochse A brain-like neural network for periodicity analysis.
M Dülk,S Fischer,M Bitter,M Caraccia,W Vogt,E Gini,H Melchior,W Hunziker,A Buxens,H Poulsen Ultrafast all-optical demultiplexer based on monolithic Mach-Zehnder interferometer with integrated semiconductor optical amplifiers.
Xing Lin,Yair Rivenson,Nezih Yardimci,Muhammed Veli,Yi Luo,Mona Jarrahi,Aydogan Ozcan (2018). All-optical machine learning using diffractive deep neural networks.
A Campo,C Greiner (2007). SU-8: a photoresist for high-aspect-ratio and 3D submicron lithography.
Y Tian,P Luo,X Wang,X Tang,L Deng The MNIST Database of Handwritten Digit Images for Machine Learning Research [Best of the Web.
C Loan (2011). Computational Frameworks for the Fast Fourier Transform.

Download References

Funding

No external funding was declared for this work.

Conflict of Interest

The authors declare no conflict of interest.

Ethical Approval

No ethics committee approval was required for this article type.

Data Availability

Not applicable for this article.

How to Cite This Article

Jing Su, Yafei Yuan, Chunmin Liu, Jing Lia. 2021. "Multi-Task Learning by Multi-Wave Optical Diffractive Network". Global Journal of Computer Science and Technology, Global Journal of Computer Science and Technology - A: Hardware & Computation GJCST-A Volume 20 (GJCST Volume 20 Issue A1).

More Citation Formats

Select Citation Style:

Download Citation

GJCST Volume 20 Issue A1

Explore Journals Explore Volume Read This Issue

Journal Specifications

Crossref Journal DOI 10.17406/gjcst

Print ISSN 0975-4350

e-ISSN 0975-4172

Keywords

Not Found

Classification

GJCST-A Classification C.2.1

Submission ReceivedDecember 6, 2019
Peer Review Double Blind
Handling Editor
Accepted December 31, 2019
Published January 15, 2020

Version of record

v1.2

Issue date

January 9, 2021

Language

English

Experiance in AR

Explore published articles in an immersive Augmented Reality environment. Our platform converts research papers into interactive 3D books, allowing readers to view and interact with content using AR and VR compatible devices.

View in VR

Read in 3D

Your published article is automatically converted into a realistic 3D book. Flip through pages and read research papers in a more engaging and interactive format.

View in 3D

Article Matrices

Total Views: 1K

Total Downloads: 93

All Trends