{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,30]],"date-time":"2026-05-30T02:08:25Z","timestamp":1780106905455,"version":"3.54.0"},"reference-count":58,"publisher":"Association for Computing Machinery (ACM)","issue":"5s","license":[{"start":{"date-parts":[[2026,8,10]],"date-time":"2026-08-10T00:00:00Z","timestamp":1786320000000},"content-version":"vor","delay-in-days":318,"URL":"http:\/\/www.acm.org\/publications\/policies\/copyright_policy#Background"}],"funder":[{"DOI":"10.13039\/100000001","name":"US National Science Foundation","doi-asserted-by":"crossref","award":["2235398 (sub-award from University of California at Riverside) and CSR-2308530"],"award-info":[{"award-number":["2235398 (sub-award from University of California at Riverside) and CSR-2308530"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/100000183","name":"Army Research Office","doi-asserted-by":"crossref","award":["ARO-W911NF-24-1-0240"],"award-info":[{"award-number":["ARO-W911NF-24-1-0240"]}],"id":[{"id":"10.13039\/100000183","id-type":"DOI","asserted-by":"crossref"}]},{"name":"Intel\u00a0CAD\u00a0SRS program"}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["ACM Trans. Embed. Comput. Syst."],"published-print":{"date-parts":[[2025,11,30]]},"abstract":"<jats:p>\n            Graph Neural Networks (GNNs) are made up of multiple layers, with each layer comprising of different compute kernels involving weight vectors and adjacency matrices of input graph dataset. These layers exhibit varying features such as sparsity, storage requirement, and impact on predictive accuracy. Non-volatile memory (NVM)-based 3D Processing-In-Memory (PIM) architectures offer a promising approach to accelerate GNN inferencing. However, NVM device-based crossbars suffer from various non-idealities that affect the overall predictive accuracy. In this work, we consider the problem of finding a suitable mapping of GNN layers to PIM-based processing elements (PEs) in a 3D manycore architecture such that the impact of crossbar non-idealities on predictive accuracy is minimized. We develop a framework called\n            <jats:italic toggle=\"yes\">GINA,<\/jats:italic>\n            which leverages low-cost, approximate Hessian-based methodology to automatically determine the GNN layers that are critical for accuracy and find a suitable GNN layer to PE mapping. To tackle non-idealities and to exploit sparsity at the crossbar level, a subset of the full crossbar is activated in a cycle, referred to as Operation Unit (OU). However, OU configurations vary with the above-mentioned GNN layer features, time-dependent conductance drift, and input graph dataset. GINA learns to optimize the OU configuration for unseen datasets as a function of GNN layer features and time-dependent conductance drift. Our experimental results demonstrate that GINA-enabled 3D PIM architecture reduces the latency and energy by 7.4\n            <jats:italic toggle=\"yes\">imes<\/jats:italic>\n            and 13\n            <jats:italic toggle=\"yes\">imes<\/jats:italic>\n            on an average, respectively, compared to state-of-the-art PIM architectures without compromising the predictive accuracy. Finally, we demonstrate the applicability of GINA to Convolutional Neural Networks (CNNs) and Vision Transformers.\n          <\/jats:p>","DOI":"10.1145\/3759918","type":"journal-article","created":{"date-parts":[[2025,8,11]],"date-time":"2025-08-11T04:00:46Z","timestamp":1754884846000},"page":"1-26","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":1,"title":["GINA: Exploiting Graph Neural Network Layer Features for Energy Efficient Inferencing in NVM-based PIM Accelerators"],"prefix":"10.1145","volume":"24","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-9517-1280","authenticated-orcid":false,"given":"Gaurav","family":"Narang","sequence":"first","affiliation":[{"name":"Arizona State University","place":["Tempe, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8170-1161","authenticated-orcid":false,"given":"Chukwufumnanya","family":"Ogbogu","sequence":"additional","affiliation":[{"name":"Washington State University","place":["Pullman, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7668-2824","authenticated-orcid":false,"given":"Biresh Kumar","family":"Joardar","sequence":"additional","affiliation":[{"name":"University of Houston","place":["Houston, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3848-5301","authenticated-orcid":false,"given":"Janardhan Rao","family":"Doppa","sequence":"additional","affiliation":[{"name":"Washington State University","place":["Pullman, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4475-6435","authenticated-orcid":false,"given":"Krishnendu","family":"Chakrabarty","sequence":"additional","affiliation":[{"name":"Arizona State University","place":["Tempe, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5930-8531","authenticated-orcid":false,"given":"Partha Pratim","family":"Pande","sequence":"additional","affiliation":[{"name":"Washington State University","place":["Pullman, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"320","published-online":{"date-parts":[[2025,9,26]]},"reference":[{"key":"e_1_3_1_2_2","volume-title":"Proceedings of the International Conference on Learning Representations (ICLR)","author":"Kipf T.","year":"2017","unstructured":"T. Kipf and M. Welling. 2017. Semi-supervised classification with graph convolutional networks. In Proceedings of the International Conference on Learning Representations (ICLR), 2017."},{"key":"e_1_3_1_3_2","volume-title":"Proceedings of the 34th Conference on Neural Information Processing Systems (NeurIPS)","author":"Hu W.","year":"2020","unstructured":"W. Hu, M. Fey, M. Zitnik, Y. Dong, H. Ren, B. Liu, M. Catasta, and J. Leskovec. 2020. Open graph benchmark: Datasets for machine learning on graphs. In Proceedings of the 34th Conference on Neural Information Processing Systems (NeurIPS), Vancouver, Canada, 2020."},{"key":"e_1_3_1_4_2","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA.2017.55"},{"key":"e_1_3_1_5_2","doi-asserted-by":"publisher","DOI":"10.1145\/3489517.3530479"},{"key":"e_1_3_1_6_2","unstructured":"S. Zhang A. Sohrabizadeh C. Wan Z. Huang Z. Hu Y. Wang Y. (Celine)Lin J. Cong and Y. Sun. 2023. A survey on graph neural network acceleration: Algorithms systems and customized hardware. arXiv preprint arXiv:2306.14052."},{"key":"e_1_3_1_7_2","first-page":"1029\u20131042","article-title":"Accelerating graph convolutional networks using crossbar-based processing-in-memory architectures","author":"Huang Y.","year":"2022","unstructured":"Y. Huang, L. Zheng, P. Yao, Q. Wang, X. Liao, H. Jin, and J. Xue. 2022. Accelerating graph convolutional networks using crossbar-based processing-in-memory architectures. In Proceedings of the 2022 IEEE International Symposium on High-Performance Computer Architecture (HPCA) (2022), 1029\u20131042.","journal-title":"Proceedings of the 2022 IEEE International Symposium on High-Performance Computer Architecture (HPCA)"},{"key":"e_1_3_1_8_2","first-page":"1\u20136","article-title":"Learning to predict IR drop with effective training for ReRAM-based neural network hardware","author":"Lee S.","year":"2020","unstructured":"S. Lee, G. Jung, M.-E. Fouda, J. Lee, A. Eltawil, and F. Kurdahi. 2020. Learning to predict IR drop with effective training for ReRAM-based neural network hardware. In Proceedings of the 2020 57th ACM\/IEEE Design Automation Conference (DAC) (2020), 1\u20136.","journal-title":"Proceedings of the 2020 57th ACM\/IEEE Design Automation Conference (DAC)"},{"key":"e_1_3_1_9_2","doi-asserted-by":"publisher","DOI":"10.1145\/3508352.3549345"},{"key":"e_1_3_1_10_2","first-page":"1\u20139","article-title":"A thermal-aware optimization framework for ReRAM-based deep neural network acceleration","author":"Shin H.","year":"2020","unstructured":"H. Shin, M. Kang, and L. Kim. 2020. A thermal-aware optimization framework for ReRAM-based deep neural network acceleration. In Proceedings of the 39th International Conference on Computer-Aided Design (ICCAD) (2020), 1\u20139.","journal-title":"Proceedings of the 39th International Conference on Computer-Aided Design (ICCAD)"},{"key":"e_1_3_1_11_2","doi-asserted-by":"publisher","DOI":"10.1109\/TED.2019.2894273"},{"key":"e_1_3_1_12_2","doi-asserted-by":"publisher","DOI":"10.1145\/3307650.3322271"},{"key":"e_1_3_1_13_2","doi-asserted-by":"publisher","DOI":"10.1145\/3508352.3549364"},{"issue":"6","key":"e_1_3_1_14_2","first-page":"2130\u20132143","article-title":"Energy-efficient DNN inferencing on ReRAM-based PIM accelerators using heterogeneous operation units","volume":"44","author":"Narang G.","year":"2024","unstructured":"G. Narang, J. R. Doppa, and P. P. Pande. 2024. Energy-efficient DNN inferencing on ReRAM-based PIM accelerators using heterogeneous operation units. IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems 44, 6 (2024), 2130\u20132143.","journal-title":"IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems"},{"key":"e_1_3_1_15_2","doi-asserted-by":"publisher","DOI":"10.1145\/3400302.3415679"},{"key":"e_1_3_1_16_2","unstructured":"A. G. Gonzalez M. Mahoney Joseph N. Golmant and Z. Yao. 2018. pytorchhessian-eigenthings: Efficient pytorch hessian eigendecomposition October 2018. [Online]. Available: Retrieved from https:\/\/github.com\/noahgolmant\/pytorch-hessian-eigenthings."},{"key":"e_1_3_1_17_2","doi-asserted-by":"publisher","DOI":"10.1145\/3665279"},{"key":"e_1_3_1_18_2","volume-title":"Proceedings of the International Conference on Learning Representations (ICLR)","author":"Veli\u010dkovi\u0107 P.","year":"2018","unstructured":"P. Veli\u010dkovi\u0107, G. Cucurull, A. Casanova, A. Romero, P. Li\u00f2, and Y. Bengio. 2018. Graph attention networks. In Proceedings of the International Conference on Learning Representations (ICLR), 2018."},{"key":"e_1_3_1_19_2","volume-title":"Proceedings of the Advances in Neural Information Processing Systems","volume":"30","author":"Hamilton W.","year":"2017","unstructured":"W. Hamilton, Z. Ying, and J. Leskovec. 2017. Inductive representation learning on large graphs. In Proceedings of the Advances in Neural Information Processing Systems 30 (2017)."},{"issue":"1","key":"e_1_3_1_20_2","first-page":"87\u2013110","article-title":"A survey on vision transformer","volume":"45","author":"Han K.","year":"2022","unstructured":"K. Han, Y. Wang, H. Chen, X. Chen, J. Guo, Z. Liu, Y. Tang, A. Xiao, C. Xu, Y. Xu, Z. Yang, Y. Zhang, and D. Tao. 2022. A survey on vision transformer. IEEE Transactions on Pattern Analysis and Machine Intelligence 45, 1 (2022), 87\u2013110.","journal-title":"IEEE Transactions on Pattern Analysis and Machine Intelligence"},{"key":"e_1_3_1_21_2","doi-asserted-by":"publisher","DOI":"10.23919\/DATE56975.2023.10137240"},{"key":"e_1_3_1_22_2","doi-asserted-by":"publisher","DOI":"10.1145\/3609097"},{"key":"e_1_3_1_23_2","first-page":"15\u201329","article-title":"Hygcn: A gcn accelerator with hybrid architecture","author":"Yan M.","year":"2020","unstructured":"M. Yan, L. Deng, X. Hu, L. Liang, Y. Feng, X. Ye, Z. Zhang, D. Fan, and Y. Xie. 2020. Hygcn: A gcn accelerator with hybrid architecture. In Proceedings of the 2020 IEEE International Symposium on High Performance Computer Architecture (HPCA) (2020), 15\u201329.","journal-title":"Proceedings of the 2020 IEEE International Symposium on High Performance Computer Architecture (HPCA)"},{"key":"e_1_3_1_24_2","first-page":"1099\u20131112","article-title":"FlowGNN: A dataflow architecture for real-time workload-agnostic graph neural network inference","author":"Sarkar R.","year":"2023","unstructured":"R. Sarkar, S. Abi-Karam, Y. He, L. Sathidevi, and C. Hao. 2023. FlowGNN: A dataflow architecture for real-time workload-agnostic graph neural network inference. In Proceedings of the 2023 IEEE International Symposium on High-Performance Computer Architecture (HPCA) (2023), 1099\u20131112.","journal-title":"Proceedings of the 2023 IEEE International Symposium on High-Performance Computer Architecture (HPCA)"},{"key":"e_1_3_1_25_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCSI.2024.3384748"},{"issue":"2","key":"e_1_3_1_26_2","first-page":"1\u201326","article-title":"NEM-GNN: DAC\/ADC-less, scalable, reconfigurable, graph and sparsity-aware near-memory accelerator for graph neural networks","volume":"21","author":"Raman S.","year":"2024","unstructured":"S. Raman, S. Raman, L. John, and J. P. Kulkarni. 2024. NEM-GNN: DAC\/ADC-less, scalable, reconfigurable, graph and sparsity-aware near-memory accelerator for graph neural networks. ACM Transactions on Architecture and Code Optimization 21, 2 (2024), 1\u201326.","journal-title":"ACM Transactions on Architecture and Code Optimization"},{"key":"e_1_3_1_27_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISVLSI59464.2023.10238622"},{"key":"e_1_3_1_28_2","doi-asserted-by":"publisher","DOI":"10.1109\/TVLSI.2021.3110721"},{"key":"e_1_3_1_29_2","first-page":"1\u20136","article-title":"Geniex: A generalized approach to emulating non-ideality in memristive xbars using neural networks","author":"Chakraborty I.","year":"2020","unstructured":"I. Chakraborty, M.-F. Ali, D.-E. Kim, A. Ankit, and K. Roy. 2020. Geniex: A generalized approach to emulating non-ideality in memristive xbars using neural networks. In Proceedings of the 2020 57th ACM\/IEEE Design Automation Conference (DAC) (2020), 1\u20136.","journal-title":"Proceedings of the 2020 57th ACM\/IEEE Design Automation Conference (DAC)"},{"key":"e_1_3_1_30_2","first-page":"1","article-title":"Noise injection adaption: End-to-end ReRAM crossbar non-ideal effect adaption for neural network mapping","author":"He Z.","year":"2019","unstructured":"Z. He, J. Lin, R. Ewetz, J. Yuan, and D. Fan. 2019. Noise injection adaption: End-to-end ReRAM crossbar non-ideal effect adaption for neural network mapping. In Proceedings of the 56th Annual Design Automation Conference (DAC) (2019), 1-6.","journal-title":"Proceedings of the 56th Annual Design Automation Conference (DAC)"},{"key":"e_1_3_1_31_2","doi-asserted-by":"publisher","DOI":"10.1145\/3400302.3415664"},{"key":"e_1_3_1_32_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-023-42110-y"},{"key":"e_1_3_1_33_2","doi-asserted-by":"publisher","DOI":"10.1038\/s44172-023-00074-3"},{"key":"e_1_3_1_34_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.amc.2014.03.110"},{"key":"e_1_3_1_35_2","doi-asserted-by":"publisher","DOI":"10.1137\/S1064827595287997"},{"key":"e_1_3_1_36_2","doi-asserted-by":"publisher","DOI":"10.1145\/3292500.3330925"},{"key":"e_1_3_1_37_2","doi-asserted-by":"publisher","DOI":"10.1109\/TEVC.2007.900837"},{"key":"e_1_3_1_38_2","first-page":"1\u20136","article-title":"Energy-efficient ReRAM-based ML training via mixed pruning and reconfigurable ADC","author":"Ogbogu C.","year":"2023","unstructured":"C. Ogbogu, M. Soumen, B.-K. Joardar, J.-R. Doppa, D. Heo, K. Chakrabarty, and P.-P. Pande. 2023. Energy-efficient ReRAM-based ML training via mixed pruning and reconfigurable ADC. In Proceedings of the 2023 IEEE\/ACM International Symposium on Low Power Electronics and Design (ISLPED) (2023), 1\u20136.","journal-title":"Proceedings of the 2023 IEEE\/ACM International Symposium on Low Power Electronics and Design (ISLPED)"},{"key":"e_1_3_1_39_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCSI.2011.2107214"},{"key":"e_1_3_1_40_2","doi-asserted-by":"publisher","DOI":"10.1109\/JETCAS.2023.3327748"},{"key":"e_1_3_1_41_2","doi-asserted-by":"publisher","DOI":"10.1080\/07468342.1997.11973825"},{"key":"e_1_3_1_42_2","first-page":"20104\u201320117","article-title":"Hessian eigenspectra of more realistic nonlinear models","volume":"34","author":"Liao Z.","year":"2021","unstructured":"Z. Liao and M. W. Mahoney. 2021. Hessian eigenspectra of more realistic nonlinear models. Advances in Neural Information Processing Systems 34 (2021), 20104\u201320117.","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_3_1_43_2","doi-asserted-by":"crossref","unstructured":"Y. C. Hsiao R. Yue and A. Dutta. 2025. Derivation of back-propagation for graph convolutional networks using matrix calculus and its application to explainable artificial intelligence. IEEE Transactions on Artificial Intelligence.","DOI":"10.1109\/TAI.2025.3569519"},{"key":"e_1_3_1_44_2","doi-asserted-by":"publisher","DOI":"10.23919\/DATE64628.2025.10993275"},{"key":"e_1_3_1_45_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISLPED58423.2023.10244486"},{"key":"e_1_3_1_46_2","doi-asserted-by":"crossref","unstructured":"Y. Ouyang F. Tang C. Hu W. Zhou and Q. Wang. 2021. Mmnnn: A tree-based multicast mechanism for noc-based deep neural network accelerators. Microprocessors and Microsystems 85 (2021) 104242.","DOI":"10.1016\/j.micpro.2021.104242"},{"issue":"2","key":"e_1_3_1_47_2","first-page":"1\u201315","article-title":"CACTI 7: New tools for interconnect exploration in innovative off-chip memories","volume":"14","author":"Balasubramonian R.","year":"2017","unstructured":"R. Balasubramonian, A.-B. Kahng, N. Muralimanohar, A. Shafiee, and V. Srinivas. 2017. CACTI 7: New tools for interconnect exploration in innovative off-chip memories. ACM Transactions on Architecture and Code Optimization (TACO) 14, 2 (2017), 1\u201315.","journal-title":"ACM Transactions on Architecture and Code Optimization (TACO)"},{"key":"e_1_3_1_48_2","doi-asserted-by":"publisher","DOI":"10.1145\/3007787.3001139"},{"issue":"11","key":"e_1_3_1_49_2","first-page":"2872\u20132884","article-title":"ReHy: A ReRAM-based digital\/analog hybrid PIM architecture for accelerating CNN training","volume":"33","author":"Jin H.","year":"2021","unstructured":"H. Jin, C. Liu, H. Liu, R. Luo, J. Xu, F. Mao, and X. Liao. 2021. ReHy: A ReRAM-based digital\/analog hybrid PIM architecture for accelerating CNN training. IEEE Transactions on Parallel and Distributed Systems 33, 11 (2021), 2872\u20132884.","journal-title":"IEEE Transactions on Parallel and Distributed Systems"},{"key":"e_1_3_1_50_2","doi-asserted-by":"publisher","DOI":"10.1109\/IEDM13553.2020.9372091"},{"key":"e_1_3_1_51_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISPASS.2013.6557149"},{"key":"e_1_3_1_52_2","article-title":"Hotspot 6.0: Validation, acceleration and extension","author":"Zhang R.","year":"2015","unstructured":"R. Zhang, M. Stan, and K. Skadron. 2015. Hotspot 6.0: Validation, acceleration and extension. In Proceedings of the University of Virginia, Tech. Rep (2015).","journal-title":"Proceedings of the University of Virginia, Tech. Rep"},{"key":"e_1_3_1_53_2","first-page":"7793\u20137804","article-title":"Beyond homophily in graph neural networks: Current limitations and effective designs","volume":"33","author":"Zhu J.","year":"2020","unstructured":"J. Zhu, Y. Yan, L. Zhao, M. Heimann, L. Akoglu, and D. Koutra. 2020. Beyond homophily in graph neural networks: Current limitations and effective designs. Advances in Neural Information Processing Systems 33 (2020), 7793\u20137804.","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_3_1_54_2","doi-asserted-by":"publisher","DOI":"10.1145\/3543507.3583268"},{"key":"e_1_3_1_55_2","unstructured":"J. Bazinska A. Ivanov T. Ben-Nun N. Dryden M. Besta S. Shen and T. Hoefler. 2023. Cached operator reordering: A unified view for fast gnn training. arXiv preprint arXiv:2308.12093."},{"key":"e_1_3_1_56_2","unstructured":"J. Hou H. Liu and S. Zhu. 2024. RCM++:Reverse cuthill-mckee ordering with bi-criteria node finder. arXiv preprint arXiv:2409.04171."},{"key":"e_1_3_1_57_2","unstructured":"L. Zhao and L. Akoglu. 2019. Pairnorm: Tackling oversmoothing in gnns. arXiv preprint arXiv:1909.12223."},{"key":"e_1_3_1_58_2","doi-asserted-by":"publisher","DOI":"10.1145\/3394486.3403076"},{"key":"e_1_3_1_59_2","doi-asserted-by":"publisher","DOI":"10.1145\/3489517.3530564"}],"container-title":["ACM Transactions on Embedded Computing Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3759918","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3759918","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,1]],"date-time":"2025-10-01T17:13:54Z","timestamp":1759338834000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3759918"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,9,26]]},"references-count":58,"journal-issue":{"issue":"5s","published-print":{"date-parts":[[2025,11,30]]}},"alternative-id":["10.1145\/3759918"],"URL":"https:\/\/doi.org\/10.1145\/3759918","relation":{},"ISSN":["1539-9087","1558-3465"],"issn-type":[{"value":"1539-9087","type":"print"},{"value":"1558-3465","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,9,26]]},"assertion":[{"value":"2025-08-03","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-08-04","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-09-26","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}