Hardware design and the competency awareness of a neural network

Abstract

The ability to estimate the uncertainty of predictions made by a neural network is essential when applying neural networks to tasks such as medical diagnosis and autonomous vehicles. The approach is of particular relevance when deploying the networks on devices with limited hardware resources, but existing competency-aware neural networks largely ignore any resource constraints. Here we examine the relationship between hardware platforms and the competency awareness of a neural network. We highlight the impact of two key areas of hardware development — increasing memory size of accelerator architectures and device-to-device variation in the emerging devices typically used in in-memory computing — on uncertainty estimation quality. We also consider the challenges that developments in uncertainty estimation methods impose on hardware designs. Finally, we explore the innovations required in terms of hardware, software, and hardware–software co-design in order to build future competency-aware neural networks.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.

from$8.99

All prices are NET prices.

Fig. 1: Competency awareness of neural networks.
Fig. 2: Trends in memory capacity and memory window.
Fig. 3: Uncertainty-related performance with respect to memory footprint, quantization and device-to-device variation.
Fig. 4: Timeline of developments in uncertainty estimation methods.

Data availability

The data that support the findings of this study are available from the corresponding author upon request.

Code availability

The code that support the findings of this study are available from the corresponding author upon request.

References

  1. 1.

    Zhu, J.-Y., Park, T., Isola, P. & Efros, A.A. Unpaired image-to-image translation using cycle-consistent adversarial networks. In 2017 IEEE Int. Conference on Computer Vision (ICCV) 2223–2232 (IEEE, 2017).

  2. 2.

    Oord, A. V. D. et al. Wavenet: A generative model for raw audio. Preprint at https://arxiv.org/abs/1609.03499 (2016).

  3. 3.

    Wu, Y. et al. Google’s neural machine translation system: Bridging the gap between human and machine translation. Preprint at https://arxiv.org/abs/1609.08144 (2016).

  4. 4.

    Johnson, J., Alahi, A. & Li, F.-F. Perceptual losses for real-time style transfer and super-resolution. In Proc. European Conference on Computer Vision 694–711 (Springer, 2016).

  5. 5.

    He, Y. et al. Streaming end-to-end speech recognition for mobile devices. In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 6381–6385 (IEEE, 2019).

  6. 6.

    Ali Eslami, S. M. et al. Neural scene representation and rendering. Science 360, 1204–1210 (2018).

    Article  Google Scholar 

  7. 7.

    Anderson, P. et al. Vision-and-language navigation: Interpreting visually-grounded navigation instructions in real environments. In Proc. IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 3674–3683 (IEEE, 2018).

  8. 8.

    Yi, K. et al. Neural-symbolic VQA: Disentangling reasoning from vision and language understanding. In Advances in Neural Information Processing Systems 31 (NIPS 2018) 1031–1042 (MIT Press, 2018).

  9. 9.

    Cloud TPU (Google, 2020); https://cloud.google.com/tpu

  10. 10.

    Intel Movidius Myriad X Vision Processing Unit Technical Specifications (Intel, 2020); https://www.intel.com/content/www/us/en/products/processors/movidius-vpu/movidius-myriad-x.html.

  11. 11.

    Chen, Y.-H., Yang, T.-J., Emer, J. & Sze, V. Eyeriss v2: A flexible accelerator for emerging deep neural networks on mobile devices. IEEE J. Em. Sel. Top. Circuits Syst. 9, 292–308 (2019).

    Article  Google Scholar 

  12. 12.

    Guo, K. et al. Angel-eye: A complete design flow for mapping CNN onto embedded fpga. IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst. 37, 35–47 (2017).

    Article  Google Scholar 

  13. 13.

    Valavi, H., Ramadge, P. J., Nestler, E. & Verma, N. A 64-tile 2.4-mb in-memory-computing CNN accelerator employing charge-domain compute. IEEE J. Solid-State Circuits 54, 1789–1799 (2019).

    Article  Google Scholar 

  14. 14.

    Xu et al. Scaling for edge inference of deep neural networks. Nat. Electron. 1, 216–222 (2018).

    Article  Google Scholar 

  15. 15.

    Zhang, S. et al. Cambricon-x: An accelerator for sparse neural networks. In 49th Annual IEEE/ACM Int. Symposium on Microarchitecture https://doi.org/10.1109/MICRO.2016.7783723 (IEEE, 2016).

  16. 16.

    Holzinger, A., Biemann, C., Pattichis, C. S. & Kell, D. B. What do we need to build explainable ai systems for the medical domain? Preprint at https://arxiv.org/abs/1712.09923 (2017).

  17. 17.

    Vovk, V., Gammerman, A. & Shafer, G. Algorithmic Learning in a Random World (Springer, 2005).

  18. 18.

    Papadopoulos, H., Vovk, V. & Gammermam, A. Conformal prediction with neural networks. In 19th IEEE Int. Conference on Tools with Artificial Intelligence (ICTAI 2007) 2, 388–395 (IEEE, 2007).

  19. 19.

    DeVries, T. & Taylor, G.W. Leveraging uncertainty estimates for predicting segmentation quality. Preprint at https://arxiv.org/abs/1807.00502 (2018).

  20. 20.

    DeVries, T. & Taylor, G.W. Learning confidence for out-of-distribution detection in neural networks. Preprint at https://arxiv.org/abs/1802.04865 (2018).

  21. 21.

    Guo, C., Pleiss, G., Sun, Y. & Weinberger, K.Q. On calibration of modern neural networks. In Proc. 34th International Conference on Machine Learning 70, 1321–1330 (ACM, 2017).

  22. 22.

    Malinin, A. & Gales, M. Predictive uncertainty estimation via prior networks. In Advances in Neural Information Processing Systems 31 (NIPS 2018) 7047–7058 (MIT Press, 2018).

  23. 23.

    Lakshminarayanan, B., Pritzel, A. & Blundell, C. Simple and scalable predictive uncertainty estimation using deep ensembles. In Advances in Neural Information Processing Systems 30 (NIPS 2017) 6402–6413 (MIT Press, 2017).

  24. 24.

    Geifman, Y., Uziel, G. & El-Yaniv, R. Bias-reduced uncertainty estimation for deep neural classifiers. In Proc. 7th Int. Conference on Learning Representations (ICLR) (ICLR, 2019).

  25. 25.

    Ding, Y., Liu, J., Xiong, J. & Shi, Y. Revisiting the evaluation of uncertainty estimation and its application to explore model complexity-uncertainty trade-off. In CVPR workshop on Fair, Data Efficient and Trusted Computer Vision 4–5 (IEEE, 2020).

  26. 26.

    Roy, A. G., Conjeti, S., Navab, N. & Wachinger, C. Inherent brain segmentation quality control from fully convnet Monte Carlo sampling. In Int. Conference on Medical Image Computing and Computer-Assisted Intervention 664–672 (Springer, 2018).

  27. 27.

    Su, H., Yin, Z., Huh, S., Kanade, T. & Zhu, J. Interactive cell segmentation based on active and semi-supervised learning. IEEE Trans. Med. Imaging 35, 762–777 (2015).

    Article  Google Scholar 

  28. 28.

    McAllister, R. et al. Concrete problems for autonomous vehicle safety: advantages of Bayesian deep learning. In Int. Joint Conferences on Artificial Intelligence, 4745–4753 (IJCAI, 2017).

  29. 29.

    Gasser, U. & Almeida, V. A. A layered model for ai governance. IEEE Internet Comput. 21, 58–62 (2017).

    Article  Google Scholar 

  30. 30.

    O’sullivan, S. et al. Legal, regulatory, and ethical frameworks for development of standards in artificial intelligence (AI) and autonomous robotic surgery. Int. J. Med. Robot. Comput. Assist. Surg. 15, e1968 (2019).

    Article  Google Scholar 

  31. 31.

    Shih, P.-J. Ethical guidelines for artificial intelligence (AI) development and the new “trust” between humans and machines. Int. J. Autom. Smart Technol. 9, 41–43 (2019).

    Google Scholar 

  32. 32.

    Liang, S., Li, Y. & Srikant, R. Enhancing the reliability of out-of-distribution image detection in neural networks. In Proc. 6th Int. Conference on Learning Representations (ICLR) (ICLR, 2018).

  33. 33.

    Kumar, A., Sarawagi, S. & Jain, U. Trainable calibration measures for neural networks from kernel mean embeddings. In Int. Conference on Machine Learning 2810–2819 (MLR, 2018).

  34. 34.

    Naeini, M. P., Cooper, G. & Hauskrecht, M. Obtaining well calibrated probabilities using Bayesian binning. In 29th AAAI Conference on Artificial Intelligence 2901–2907 (AAAI, 2015).

  35. 35.

    Kiureghian, A. D. & Ditlevsen, O. Aleatory or epistemic? Does it matter? Struct, Saf. 31, 105–112 (2009).

    Article  Google Scholar 

  36. 36.

    Kendall, A. & Gal, Y. What uncertainties do we need in bayesian deep learning for computer vision? In Advances in Neural Information Processing Systems 30 (NIPS 2017) 5574–5584 (MIT Press, 2017).

  37. 37.

    Shalev, G., Adi, Y. & Keshet, J. Out-of-distribution detection using multiple semantic label representations. In Advances in Neural Information Processing Systems 31 (NIPS 2018) 7375–7385 (MIT Press, 2018).

  38. 38.

    Lee, K., Lee, K., Lee, K. and Shin, J. Training confidence-calibrated classifiers for detecting out-of-distribution samples. In Proc. 6th Int. Conference on Learning Representations (ICLR) (ICLR, 2018).

  39. 39.

    Hendrycks, D. & Gimpel, K. A baseline for detecting misclassified and out-of-distribution examples in neural networks. In Proc. 5th Int. Conference on Learning Representations (ICLR) (ICLR, 2017).

  40. 40.

    Brock, A., Donahue, J. & Simonyan, K. Large scale GAN training for high fidelity natural image synthesis. In Proc. 7th International Conference on Learning Representations (ICLR) (ICLR, 2019).

  41. 41.

    Devlin, J. et al. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. In Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies 4171–4186 (MIT Press, 2019).

  42. 42.

    Sandler, M. et al. Mobilenetv2: Inverted residuals and linear bottlenecks. In Proc. IEEE Conference on Computer Vision and Pattern Recognition, 4510–4520 (IEEE, 2018).

  43. 43.

    Chen, Y.-H., Emer, J. Sze, V. Eyeriss: a spatial architecture for energy-efficient dataflow for convolutional neural networks. In Proc. 43rd International Symposium on Computer Architecture, 367–379 (IEEE, 2016).

  44. 44.

    Gao, M., Ayers, G. & Kozyrakis, C. Practical near-data processing for in-memory analytics frameworks. In 2015 International Conference on Parallel Architecture and Compilation (PACT) 113–124 (IEEE, 2015).

  45. 45.

    Xue, C.-X. et al. 24.1 a 1Mb multibit ReRAM computing-in-memory macro with 14.6 ns parallel MAC computing time for CNN based AI edge processors. In 2019 IEEE International Solid-State Circuits Conference (ISSCC) 388–390 (IEEE, 2019).

  46. 46.

    Jiang, W., Xie, B. & Liu, C. et al. Integrating memristors and CMOS for better AI. Nat. Electron. 2, 376–377 (2019).

    Article  Google Scholar 

  47. 47.

    Huang, G., Liu, Z., Maaten, L.V.D. & Weinberger, K.Q. Densely connected convolutional networks. In Proc. IEEE Conference on Computer Vision and Pattern Recognition 4700–4708 (IEEE, 2017).

  48. 48.

    Zagoruyko, S. & Komodakis, N. Wide residual networks. In Proc. British Machine Vision Conference (BMVC) 87.1–87.12 (BMVA, 2016).

  49. 49.

    Ni, K. et al. Fundamental understanding and control of device-to-device variation in deeply scaled ferroelectric FETs. In Proc. 2019 Symposium on VLSI Technology 40–41 (IEEE, 2019).

  50. 50.

    Jerry, M. et al. Ferroelectric FET analog synapse for acceleration of deep neural network training. In Proc. 2017 IEEE International Electron Devices Meeting (IEDM) 6.2.1–6.2.4 (IEEE, 2017).

  51. 51.

    Zhao, M. et al. Investigation of statistical retention of filamentary analog RRAM for neuromophic computing. In Proc. 2017 IEEE International Electron Devices Meeting (IEDM) 39.4.1–39.4.4 (IEEE, 2017).

  52. 52.

    Chou, T., Tang, W., Botimer, J. & Zhang, Z. CASCADE: Connecting RRAMs to Extend Analog Dataflow In An End-To-End In-Memory Processing Paradigm. In Proc. 52nd Annual IEEE/ACM International Symposium on Microarchitecture 114–125 (IEEE, 2019).

  53. 53.

    MacKay, D. J. A practical Bayesian framework for backpropagation networks. Neural Comput. 4, 448–472 (1992).

    Article  Google Scholar 

  54. 54.

    Neal, R. M. Bayesian Learning for Neural Networks 118 (Springer, 2012).

  55. 55.

    Blundell, C., Cornebise, J., Kavukcuoglu, K. & Wierstra, D. Weight uncertainty in neural network. In Proc. International Conference on Machine Learning 1613–1622 (ACM, 2015).

  56. 56.

    Graves, A. Practical variational inference for neural networks. In Advances in Neural Information Processing Systems 24 (NIPS 2011) 2348–2356 (MIT Press, 2011).

  57. 57.

    Louizos C. & Welling, M. Multiplicative normalizing flows for variational Bayesian neural networks. In Proc. 34th International Conference on Machine Learning https://doi.org/10.5555/3305890.3305910 (ACM, 2017).

  58. 58.

    Nair, T., Precup, D., Arnold, D. L. & Arbel, T. Exploring uncertainty measures in deep networks for multiple sclerosis lesion detection and segmentation. In Proc. Int. Conference on Medical Image Computing and Computer-Assisted Intervention 655–663 (Springer, 2018).

  59. 59.

    Ovadia, Y. et al. Can You Trust Your Model’s Uncertainty? Evaluating Predictive Uncertainty Under Dataset Shift. In Advances in Neural Information Processing Systems 32 (NIPS 2019) 13991–14002 (MIT Press, 2019).

  60. 60.

    Dhamija, A. R., Günther, M. & Boult, T. Reducing network agnostophobia. In Advances in Neural Information Processing Systems 31 (NIPS 2018) 9175–9186 (MIT Press, 2018).

  61. 61.

    Hein, M., Andriushchenko, M. & Bitterwolf, J. Why ReLU networks yield high-confidence predictions far away from the training data and how to mitigate the problem. In Proc. Conference on Computer Vision and Pattern Recognition (CVPR) 41–50 (IEEE, 2018).

  62. 62.

    Alexandari, A., Kundaje, A. & Shrikumar, A. Maximum likelihood with bias-corrected calibration is hard-to-beat at label shift adaptation. Preprint at https://arxiv.org/abs/1901.06852 (2019).

  63. 63.

    Chen, T., Navrátil, J., Iyengar, V. & Shanmugam, K. Confidence scoring using whitebox meta-models with linear classifier probes. In Proc. 22nd International Conference on Artificial Intelligence and Statistics 1467–1475 (PMLR, 2019).

  64. 64.

    Mandelbaum, A. & Weinshall, D. Distance-based confidence score for neural network classifiers. Preprint at https://arxiv.org/abs/1709.09844 (2017).

  65. 65.

    Oberdiek, P., Rottmann, M. & Gottschalk, H. Classification uncertainty of deep neural networks based on gradient information. In IAPR Workshop on Artificial Neural Networks in Pattern Recognition 113–125 (Springer, 2018).

  66. 66.

    Teerapittayanon, S., McDanel, B. & Kung, H.-T. Branchynet: Fast inference via early exiting from deep neural networks. In 2016 23rd Int. Conference on Pattern Recognition (ICPR) 2464–2469 (Springer, 2016).

  67. 67.

    Wang, X. et al. Idk cascades: fast deep learning by learning not to overthink. In Proc. Conference on Uncertainty in Artificial Intelligence 580–590 (UAIA, 2018).

  68. 68.

    Sze, V., Chen, Y.-H., Yang, T.-J. & Emer, J. S. Efficient processing of deep neural networks: a tutorial and survey. Proc. IEEE 105, 2295–2329 (2017).

    Article  Google Scholar 

  69. 69.

    Geifman, Y. & El-Yaniv, R. Selectivenet: a deep neural network with an integrated reject option. In Proc. Int. Conference on Machine Learning 2151–2159 (MLR, 2019).

  70. 70.

    Song, C., Liu, B., Wen, W., Li, H. & Chen, Y. A quantization-aware regularized learning method in multilevel memristor-based neuromorphic computing system. In 2017 IEEE 6th Non-Volatile Memory Systems and Applications Symposium (NVMSA) https://doi.org/10.1109/NVMSA.2017.8064465 (IEEE, 2017).

  71. 71.

    He, K., Zhang, X., Ren, S. & Sun, J. Deep residual learning for image recognition. In Proc. IEEE conference on Computer Vision and Pattern Recognition 770–778 (IEEE, 2016).

  72. 72.

    Gaier, A. & Ha, D. Weight agnostic neural networks. In Advances in Neural Information Processing Systems 32 (NIPS 2019) 5364–5378 (MIT Press, 2019).

  73. 73.

    Nguyen-Phuoc, T., Li, C., Theis, L. Richardt, C. & Yang, Y.-L. Hologan: Unsupervised learning of 3D representations from natural images. In Proc. IEEE Int. Conference on Computer Vision 7588–7597 (IEEE, 2019).

  74. 74.

    Nalisnick, E. et al. Hybrid models with deep and invertible features. In Proc. Int. Conference on Machine Learning 4723–4732 (MLR, 2019).

  75. 75.

    Wu, C.-J. et al. Machine learning at Facebook: understanding inference at the edge. In 2019 IEEE International Symposium on High Performance Computer Architecture (HPCA) 331–344 (IEEE, 2019).

  76. 76.

    Gupta, U. et al. The architectural implications of Facebook’s DNN-based personalized recommendation. In 2020 IEEE Int. Symposium on High Performance Computer Architecture (HPCA) 488–501 (IEEE, 2020).

  77. 77.

    Krizhevsky, A., Sutskever, I. & Hinton, G.E. Imagenet classification with deep convolutional neural networks. In Advances in Neural Information Processing Systems 25 (NIPS 2012) 1097–1105 (MIT Press, 2012).

  78. 78.

    Zoph, B. & Le, Q.V. Neural architecture search with reinforcement learning. In Proc. 5th Int. Conference on Learning Representations (ICLR) (ICLR, 2017).

  79. 79.

    Tan, M. et al. Mnasnet: Platform-aware neural architecture search for mobile. In Proc. IEEE Conference on Computer Vision and Pattern Recognition 2820–2828 (IEEE, 2019).

  80. 80.

    Yang, L. et al. Co-exploring neural architecture and network-on-chip design for real-time artificial intelligence. In Proc. Asia and South Pacific Design Automation Conference (ASP-DAC) 85–90 (2020).

  81. 81.

    Liu, H., Simonyan, K. & Yang, Y. DARTS: Differentiable architecture search. In Proc. 7th Int. Conference on Learning Representations (ICLR) (ICLR, 2019).

  82. 82.

    Hendrycks, D. & Dietterich, T. Benchmarking neural network robustness to common corruptions and perturbations. In Proc. 7th Int. Conference on Learning Representations (ICLR) (ICLR, 2019).

  83. 83.

    Huang, X., Kwiatkowska, M., Wang, S. & Wu, M. Safety verification of deep neural networks. In Proc. Int. Conference on Computer Aided Verification 3–29 (Springer, 2017).

  84. 84.

    Papernot, N. et al. Practical black-box attacks against machine learning. In Proc. 2017 ACM Asia Conference on Computer and Communications Security 506–519 (ACM, 2017).

  85. 85.

    Abadi, M. et al. Deep learning with differential privacy. In Proc. 2016 ACM SIGSAC Conference on Computer and Communications Security 308–318 (ACM, 2016).

  86. 86.

    Papernot, N. et al. Practical black-box attacks against machine learning. In Proc. 2017 ACM Asia Conference on Computer and Communications security 506–519 (ACM, 2017).

  87. 87.

    Gal, Y. & Ghahramani, Z. Dropout as a bayesian approximation: representing model uncertainty in deep learning. In Proc Int. Conference on Machine Learning 1050–1059 (MLR, 2016).

  88. 88.

    Liu, S. et al. Cambricon: An instruction set architecture for neural networks. In 2016 ACM/IEEE 43rd Annual International Symposium on Computer Architecture (ISCA) 393–405 (IEEE, 2016).

  89. 89.

    Chen, Y.-H., Krishna, T., Emer, J. S. & Sze, V. Eyeriss: An energy-efficient reconfigurable accelerator for deep convolutional neural networks. IEEE J. Solid-State Circ. 52, 127–138 (2016).

    Article  Google Scholar 

  90. 90.

    Chen, T. et al. Diannao: A small-footprint high-throughput accelerator for ubiquitous machine-learning. ACM SIGARCH Comput. Archit. News 42, 269–284 (ACM, 2014).

  91. 91.

    Chen, Y. et al. Dadiannao: A machine-learning supercomputer. In 2014 47th Annual IEEE/ACM Int. Symposium on Microarchitecture 609–622 (IEEE, 2014).

  92. 92.

    Jouppi, N. P. et al. In-datacenter performance analysis of a tensor processing unit. In Proc. 44th Annual Int. Symposium on Computer Architecture https://doi.org/10.1145/3140659.3080246 (ACM, 2017).

  93. 93.

    Han, S. et al. EIE: efficient inference engine on compressed deep neural network. ACM SIGARCH Comput. Archit. News 44, 243–254 (2016).

    Article  Google Scholar 

  94. 94.

    Farabet, C. et al. Neuflow: A runtime reconfigurable dataflow processor for vision. In CVPR2011 Workshops 109–116 (IEEE, 2011).

  95. 95.

    Yoo, H.-J. et al. A 1.93 tops/w scalable deep learning/inference processor with tetra-parallel MIMD architecture for big data applications. In IEEE Int. Solid-State Circuits Conference 80–81 (IEEE, 2015).

  96. 96.

    Du, Z. et al. ShiDianNao: Shifting vision processing closer to the sensor. In Proc. 42nd Annual International Symposium on Computer Architecture 92–104 (IEEE, 2015).

  97. 97.

    Moons, B. & Verhelst, M. A 0.3–2.6 TOPS/W precision-scalable processor for real-time large-scale ConvNets. In Proc. 2016 IEEE Symposium on VLSI Circuits (VLSI-Circuits) https://doi.org/10.1109/VLSIC.2016.7573525 (IEEE, 2016).

  98. 98.

    Whatmough, P. N. et al. 14.3 A 28nm SoC with a 1.2 GHz 568nJ/prediction sparse deep-neural-network engine with >0.1 timing error rate tolerance for IoT applications. In Proc. 2017 IEEE Int. Solid-State Circuits Conference (ISSCC) 242–243 (IEEE, 2017).

  99. 99.

    Zhou, X. et al. Cambricon-s: Addressing irregularity in sparse neural networks through a cooperative software/hardware approach. In Proc. 2018 51st Annual IEEE/ACM Int. Symposium on Microarchitecture (MICRO) 15–28 (IEEE, 2018).

  100. 100.

    Song, J. et al. 7.1 An 11.5 TOPS/W 1024-MAC butterfly structure dual-core sparsity-aware neural processing unit in 8nm flagship mobile SoC. In 2019 IEEE Int. Solid-State Circuits Conference-(ISSCC) 130–132 (IEEE, 2019).

  101. 101.

    Desoli, G. et al. 14.1 a 2.9 TOPS/W deep convolutional neural network SOC in FD-SOI 28nm for intelligent embedded systems. In 2017 IEEE Int. Solid-State Circuits Conference (ISSCC) 238–239 (IEEE, 2017).

  102. 102.

    Lee, J. et al. UNPU: A 50.6 TOPS/W unified deep neural network accelerator with 1b-to-16b fully-variable weight bit-precision. In 2018 IEEE Int. Solid-State Circuits Conference-(ISSCC) 218-220 (IEEE, 2018).

  103. 103.

    Park, E., Kim, D. & Yoo, S. Energy-efficient neural network accelerator based on outlier-aware low-precision computation. In 2018 ACM/IEEE 45th Annual Int. Symposium on Computer Architecture (ISCA) 688–698 (IEEE, 2018).

  104. 104.

    Judd, P., Albericio, J., Hetherington, T., Aamodt, T. M. & Moshovos, A. Stripes: Bit-serial deep neural network computing. In 2016 49th Annual IEEE/ACM Int. Symposium on Microarchitecture (MICRO) https://doi.org/10.1109/MICRO.2016.7783722. (IEEE, 2016).

  105. 105.

    Sharma, H. et al. Bit fusion: Bit-level dynamically composable architecture for accelerating deep neural network. In 2018 ACM/IEEE 45th Annual Int. Symposium on Computer Architecture (ISCA) 764–775 (IEEE, 2018).

  106. 106.

    Aimar, A. et al. Nullhop: A flexible convolutional neural network accelerator based on sparse representations of feature maps. IEEE Trans. Neural Netw. Learn. Syst. 30, 644–656 (2018).

    Article  Google Scholar 

  107. 107.

    Parashar, A. et al. Scnn: An accelerator for compressed-sparse convolutional neural networks. ACM SIGARCH Comput. Archit. News 45, 27–40 (2017).

    Article  Google Scholar 

  108. 108.

    Moloney, D. et al. Myriad 2: Eye of the computational vision storm. In 2014 IEEE Hot Chips 26 Symposium (HCS) https://doi.org/10.1109/HOTCHIPS.2014.7478823 (IEEE, 2014).

  109. 109.

    Intel Agilex FPGAs and SOCs (Intel, 2020); https://www.intel.com/content/www/us/en/products/programmable/fpga/agilex.html

  110. 110.

    List of Graphics Processing Units (Wikipedia, 2020); https://en.wikipedia.org/wiki/List_of_Nvidia_graphics_processing_units

  111. 111.

    Kumbhare, P. et al. A selectorless RRAM with record memory window and nonlinearity based on trap filled limit mechanism. In 2015 15th Non-Volatile Memory Technology Symposium (NVMTS) https://doi.org/10.1109/NVMTS.2015.7457491 (IEEE, 2015).

  112. 112.

    Larcher, L. et al. A compact model of program window in HfO x RRAM devices for conductive filament characteristics analysis. IEEE Trans. Electron Dev. 61, 2668–2673 (2014).

    Article  Google Scholar 

  113. 113.

    Lee, S. et al. Engineering oxygen vacancy of tunnel barrier and switching layer for both selectivity and reliability of selector-less ReRAM. IEEE Electron Dev. Lett. 35, 1022–1024 (2014).

    Article  Google Scholar 

  114. 114.

    Lee, S. et al. Selector-less ReRAM with an excellent non-linearity and reliability by the band-gap engineered multi-layer titanium oxide and triangular shaped AC pulse. In 2013 IEEE Int. Electron Devices Meeting 10.6.1-10.6.4 (IEEE, 2013).

  115. 115.

    Woo, J. et al. Selector-less RRAM with non-linearity of device for cross-point array applications. Microelectron. Eng. 109, 360–363 (2013).

    Article  Google Scholar 

  116. 116.

    Lee, S. et al. Effect of AC pulse overshoot on nonlinearity and reliability of selectorless resistive random access memory in AC pulse operation. Solid-State Electron. 104, 70–74 (2015).

    Article  Google Scholar 

  117. 117.

    Dongale, T. D. et al. Effect of write voltage and frequency on the reliability aspects of memristor-based RRAM. Int. Nano Lett. 7, 209–216 (2017).

    Article  Google Scholar 

  118. 118.

    Gismatulin, A., Volodin, V., Gritsenko, V. & Chin, A. All nonmetal resistive random access memory. Sci. Rep. 9, 6144 (2019).

    Article  Google Scholar 

  119. 119.

    Grossi, A. et al. Experimental investigation of 4-kb RRAM arrays programming conditions suitable for TCAM. IEEE Trans. VLSI Syst. 26, 2599–2607 (2018).

    Article  Google Scholar 

  120. 120.

    Mulaosmanovic, H. et al. Evidence of single domain switching in hafnium oxide based FeFETs: Enabler for multi-level FeFET memory cells. In Proc. 2015 IEEE Int. Electron Devices Meeting (IEDM) 26.8.1–26.8.3 (IEEE, 2015).

  121. 121.

    Ni, K., Li, X, Smith, J. A., Jerry, M. & Datta, S. Write disturb in ferroelectric FETs and its implication for 1T-FeFET AND memory arrays. IEEE Electron Device Lett. 39, 1656–1659 (2018)

  122. 122.

    Zhang, Z., Dalca, A. V. & Sabuncu, M. R. Confidence calibration for convolutional neural networks using structured dropout. Preprint at https://arxiv.org/abs/1906.09551 (2019).

  123. 123.

    Atanov, A., Ashukha, A., Molchanov, D., Neklyudov, K. & Vetrov, D. Uncertainty estimation via stochastic batch normalization. In Proc. Int. Symposium on Neural Networks 261–269 (Springer, 2019).

  124. 124.

    Liu, Z. et al. Deep gamblers: learning to abstain with portfolio theory. In Advances in Neural Information Processing Systems 32 (NIPS 2019) 10622–10632 (MIT Press, 2019).

  125. 125.

    Qiu, X., Meyerson, E. & Miikkulainen, R. Quantifying point-prediction uncertainty in neural networks via residual estimation with an I/O kernel. In Proc. 8th Int. Conference on Learning Representations (ICLR) (ICLR, 2020).

Download references

Author information

Affiliations

Authors

Contributions

Y.D. contributed to all aspects of the project. X.X., and W.J. contributed to data collection and discussion. J.L., Q.L., J.X., and X.H. contributed to discussion and writing. Y.S. contributed to project planning, development, discussion, and writing.

Corresponding authors

Correspondence to Xiaowei Xu or Yiyu Shi.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Experiment setup and supplementary results.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Ding, Y., Jiang, W., Lou, Q. et al. Hardware design and the competency awareness of a neural network. Nat Electron 3, 514–523 (2020). https://doi.org/10.1038/s41928-020-00476-7

Download citation

Further reading

Search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing