Outcomes of the Bonseyes project

Scientific publications

Publications in peer-reviewed journals, conference proceedings, preprints, PhD theses, book chapters and books.
1.
Overexpression of α-Synuclein by Oligodendrocytes in Transgenic Mice Does Not Recapitulate the Fibrillar Aggregation Seen in Multiple System Atrophy.
Cells 9, 2371 (2020). doi:10.3390/cells9112371
2.
Robustness to adversarial examples can be improved with overfitting.
International Journal of Machine Learning and Cybernetics 11, 935–944 (2020). doi:10.1007/s13042-020-01097-4
3.
Performance Aware Convolutional Neural Network Channel Pruning for Embedded GPUs.
in 2019 Annual IEEE International Symposium on Workload Characterization (IISWC’19) (2020). Archive: arXiv.org
4.
BlockSwap: Fisher-guided Block Substitution for Network Compression on a Budget.
arXiv:1906.04113 [cs, stat] (2020). Archive: arXiv.org
5.
Performance-Oriented Neural Architecture Search.
arXiv:2001.02976 [cs] (2020). Archive: http://arxiv.org/
6.
Distributed Ledger for Provenance Tracking of Artificial Intelligence Assets.
in Privacy and Identity Management. Data for Better Living: AI and Privacy: 14th IFIP WG 9.2, 9.6/11.7, 11.6/SIG 9.2.2 International Summer School, Windisch, Switzerland, August 19–23, 2019, Revised Selected Papers (Friedewald, M., Önen, M., Lievens, E., Krenn, S. & Fricker, S.eds. ) (Springer International Publishing, 2020). Archive: arXiv.org
7.
IoT meets distributed AI - Deployment scenarios of Bonseyes AI applications on FIWARE.
in 2019 IEEE 38th International Performance Computing and Communications Conference (IPCCC) 1–2 (2019). doi:10.1109/IPCCC47392.2019.8958742. Archive: ArODES
8.
Framework for Analysis of Multi-party Collaboration.
in 2019 IEEE 27th International Requirements Engineering Conference Workshops (REW) 44–53 (2019). doi:10.1109/REW.2019.00013. Archive: DiVA
9.
A Closer Look at Structured Pruning for Neural Network Compression.
arXiv:1810.04622 [cs, stat] (2019). http://arxiv.org/abs/1810.04622.
10.
Separable Layers Enable Structured Efficient Linear Substitutions.
arXiv:1906.00859 [cs, stat] (2019). Archive: arXiv.org
11.
RecNets: Channel-wise Recurrent Convolutional Neural Networks.
arXiv:1905.11910 [cs, stat] (2019). Archive: arXiv.org
12.
On the Relation Between the Sharpest Directions of DNN Loss and the SGD Step Length.
Seventh International Conference on Learning Representations (2019). Archive: arXiv.org
13.
Distilling with Performance Enhanced Students.
(2019). Archive: arXiv.org
14.
Assume, Augment and Learn: Unsupervised Few-Shot Meta-Learning via Random Labels and Data Augmentation.
arXiv:1902.09884 [cs, stat] (2019). Archive: arXiv.org
15.
AI Pipeline - bringing AI to you. End-to-end integration of data, algorithms and deployment tools.
in Emerging Deep Learning Accelerators (EDLA) Workshop at HiPEAC 2019 (2019). Archive: arXiv.org
16.
Towards Secure Collaborative AI Service Chains.
(Blekinge Institute of Technology, Faculty of Computing, Department of Computer Science., 2019). Archive: DIVA
17.
Scalar Arithmetic Multiple Data: Customizable Precision for Deep Neural Networks.
2019 IEEE 26th Symposium on Computer Arithmetic (ARITH) 61–68 (2019). doi:10.1109/ARITH.2019.00018. Archive: http://arxiv.org/
18.
Learning to infer: RL-based search for DNN primitive selection on Heterogeneous Embedded Systems.
in Proceedings of Design, Automation and Test in Europe Conference, DATE 19. March 2019 (2019). doi:https://doi.org/10.23919/DATE.2019.8714959. Archive: arXiv.org
19.
Designing a Secure IoT System Architecture from a Virtual Premise for a Collaborative AI Lab.
in Proceedings of the Workshop on Decentralized IoT Systems and Security (DISS) (2019). Archive: DIVA
20.
Characterising Across-Stack Optimisations for Deep Convolutional Neural Networks.
in Proceedings of the Workload Characterization (IISWC), 2018 IEEE International Symposium 101–110 (IEEE, 2018). doi:10.1109/IISWC.2018.8573503. Archive: arXiv.org
21.
QUENN: QUantization Engine for low-power Neural Networks.
in CF ’18 Proceedings of the 15th ACM International Conference on Computing Frontiers (2018). doi:10.1145/3203217.3203282. Archive: arXiv.org
22.
How to train your MAML.
arXiv:1810.09502 [cs, stat] (2018). Archive:
23.
Augmenting Image Classifiers Using Data Augmentation Generative Adversarial Networks.
in Artificial Neural Networks and Machine Learning – ICANN 2018 594–603 (Springer International Publishing, 2018). doi:10.1007/978-3-030-01424-7_58. Archive: Edinburgh Research Explorer
24.
DNN’s Sharpest Directions Along the SGD Trajectory.
in Modern Trends in Nonconvex Optimization for Machine Learning workshop at International Conference on Machine Learning 2018 (2018). Archive: xarXiv.org
25.
Towards Privacy Requirements for Collaborative Development of AI Applications.
in 14th Swedish National Computer Networking Workshop (SNCNW), 2018 (2018). Archive: DIVA
26.
Optimal DNN primitive selection with partitioned boolean quadratic programming.
in Proceedings of the 2018 International Symposium on Code Generation and Optimization - CGO 2018 340–351 (ACM Press, 2018). doi:10.1145/3168805. Archive: arXiv.org
27.
Moonshine: Distilling with Cheap Convolutions.
in Thirty-second Conference on Neural Information Processing Systems (NIPS 2018) (2018). Archive: arXiv.org
28.
Width of Minima Reached by Stochastic Gradient Descent is Influenced by Learning Rate to Batch Size Ratio.
in Artificial Neural Networks and Machine Learning – ICANN 2018 (Kůrková, V., Manolopoulos, Y., Hammer, B., Iliadis, L. & Maglogiannis, I.eds. ) 11141, 392–402 (Springer International Publishing, 2018). doi:10.1007/978-3-030-01424-7_39. Archive: Edinburgh Research Explorer
29.
Three Factors Influencing Minima in SGD.
in International Conference on Artificial Neural Networks 2018 (2018). Archive: arXiv.org
30.
Accelerating Deep Neural Networks on Low Power Heterogeneous Architectures.
in 11th International Workshop on Programmability and Architectures for Heterogeneous Multicores (MULTIPROG-2018) (2018). Archive: Semantic Scholar
31.
Artifact Compatibility for Enabling Collaboration in the Artificial Intelligence Ecosystem.
in Software Business (Wnuk, K. & Brinkkemper, S.eds. ) 336, 56–71 (Springer International Publishing, 2018). doi:10.1007/978-3-030-04840-2_5. Archive: DIVA
32.
Privacy and DRM Requirements for Collaborative Development of AI Applications.
in Proceedings of the 13th International Conference on Availability, Reliability and Security - ARES 2018 1–8 (ACM Press, 2018). doi:10.1145/3230833.3233268. Archive: DIVA
33.
Low-memory GEMM-based convolution algorithms for deep neural networks.
arXiv:1709.03395 [cs] (2017). Archive: arXiv.org
34.
Pricing of Data Products in Data Marketplaces.
in Software Business 49–66 (Springer, Cham, 2017). doi:10.1007/978-3-319-69191-6_4. Archive: DIVA
35.
Performance Analysis and Optimization of Sparse Matrix-Vector Multiplication on Modern Multi- and Many-Core Processors.
in 2017 46th International Conference on Parallel Processing (ICPP) 292–301 (IEEE, 2017). doi:10.1109/ICPP.2017.38. Archive: arXiv.org
36.
BONSEYES: Platform for Open Development of Systems of Artificial Intelligence: Invited Paper.
in Proceedings of the Computing Frontiers Conference 299–304 (ACM, 2017). doi:10.1145/3075564.3076259
37.
Privacy and trust in cloud-based marketplaces for AI and data resources.
in IFIPTM: IFIP International Conference on Trust Management 223–225 (Springer New York LLC, 2017). doi:10.1007/978-3-319-59171-1
38.
Flexible Privacy and High Trust in the Next Generation Internet : The Use Case of a Cloud-based Marketplace for AI.
in SNCNW - Swedish National Computer Networking Workshop, Halmstad (Halmstad university, 2017). http://urn.kb.se/resolve?urn=urn:nbn:se:bth-14963.
39.
Parallel Multi Channel convolution using General Matrix Multiplication.
in 2017 IEEE 28th International Conference on Application-specific Systems, Architectures and Processors (ASAP) 19–24 (IEEE, 2017). doi:10.1109/ASAP.2017.7995254. Archive: arXiv.org

Public deliverables

Approved deliverables available to the public.
1.
D3.1 Initial Deep Learning Toolbox
2.
D3.5 Initial Platform Deployment Methods and Tools
3.
D5.1 Demonstrator Proof-of-Concept

Selected presentations

Presentations at major conferences and public events
May
18
"Data >< Intelligence" . Keynote at Zooming Innovation in Consumer Electronics International Conference 2018 (ZINC 2018), Novi Sad, Serbia, 30 Mai 2018.
Mar
18
"BONSEYES: The artificial intelligence marketplace Supporting Surgical Data Science". DGE-BV 2018, Munich, Germany, 17 March 2018.
Dec
17
"Bonseyes AI Marketplace for Secure and Distributed Artificial Intelligence". School of Computer Science and Engineering at the University of New South Wales, Sydney, Australia, 14 December 2017.
Dec
17
"The Hardware and Software that will bring Deep Learning Everywhere" . Manchester, UK EMiT@CIUK workshop, 13 December 2017.
Nov
17
"DRM and Privacy in Virtualised and Programmable Network Architectures and Functions – The Bonseyes Use Case". Keynote at the Fourth Workshop on Network Function Virtualization and Programmable Networks (co-located with the 2017 IEEE Conference on Network Function Virtualization and Software Defined Networks (IEEE NFV-SDN 2017), Berlin, 6 November 2017.
Sep
17
"Artificial Intelligence: Mysteries of Emotions". ICCE Berlin 2017, Germany, 5 September 2017.
Jun
17
"Hybrid and Flexible Computing Architectures for Deep Learning Systems". Keynote at Zooming Innovation in Consumer Electronics International Conference 2017 (ZINC 2017), Novi Sad, Serbia. 31 May – 1 June 2017.
May
17
"BONSEYES: Platform for Open Development of Systems of Artificial Intelligence". ACM International Conference on Computing Frontiers 2017. 15–17 May, 2017, Siena, Italy.

AI Marketplace Flyer

The AI Marketplace flyer is mainly intended as a printed product but its electronic version is available for download.
Please contact elena@nviso.ai for a printed version.

Project Flyer

The Bonseyes Project flyer is mainly intended as a printed product but its electronic version is available for download.

Videos

Introductory videos to present the project, its goals and its partners.
This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 732204 (Bonseyes). This work is supported by the Swiss State Secretariat for Education‚ Research and Innovation (SERI) under contract number 16.0159. The opinions expressed and arguments employed herein do not necessarily reflect the official views of these funding bodies.
© 2017 Bonseyes Project – Created by SCIPROM – Privacy Policy