{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,5]],"date-time":"2026-06-05T06:02:06Z","timestamp":1780639326124,"version":"3.54.1"},"reference-count":25,"publisher":"IEEE","license":[{"start":{"date-parts":[[2026,4,20]],"date-time":"2026-04-20T00:00:00Z","timestamp":1776643200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,4,20]],"date-time":"2026-04-20T00:00:00Z","timestamp":1776643200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,4,20]]},"DOI":"10.23919\/date69613.2026.11539743","type":"proceedings-article","created":{"date-parts":[[2026,6,4]],"date-time":"2026-06-04T19:53:10Z","timestamp":1780602790000},"page":"1-7","source":"Crossref","is-referenced-by-count":0,"title":["GCPT: Gradient-aware Clustering Method for Efficient Post-Training Quantization in Large Neural Networks"],"prefix":"10.23919","author":[{"given":"Chuyi","family":"Dai","sequence":"first","affiliation":[{"name":"College of Integrated Circuits & Micro-Nano Electronics School of Integrated Circuits, Fudan University,State Key Laboratory of Integrated Chips and Systems,Shanghai,China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chen","family":"Ye","sequence":"additional","affiliation":[{"name":"College of Integrated Circuits & Micro-Nano Electronics School of Integrated Circuits, Fudan University,State Key Laboratory of Integrated Chips and Systems,Shanghai,China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zeyu","family":"Li","sequence":"additional","affiliation":[{"name":"The Hong Kong University of Science and Technology,Electronics and Computer Engineering Department,Hong Kong,China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jun","family":"Tao","sequence":"additional","affiliation":[{"name":"College of Integrated Circuits & Micro-Nano Electronics School of Integrated Circuits, Fudan University,State Key Laboratory of Integrated Chips and Systems,Shanghai,China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wei","family":"Zhang","sequence":"additional","affiliation":[{"name":"The Hong Kong University of Science and Technology,Electronics and Computer Engineering Department,Hong Kong,China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Grace Li","family":"Zhang","sequence":"additional","affiliation":[{"name":"TU Darmstadt,Hardware for AI Group,Darmstadt,Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xin","family":"Li","sequence":"additional","affiliation":[{"name":"Duke Kunshan University,Digital Innovation Research Center,Kunshan,China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","article-title":"Gpt-4 technical report","author":"Achiam","year":"2023"},{"key":"ref2","article-title":"Llama: Open and efficient foundation language models","author":"Touvron","year":"2023"},{"key":"ref3","first-page":"1135","article-title":"Learning both weights and connections for efficient neural network","author":"Han","year":"2015","journal-title":"Adv. Neural Inf. Process. Syst"},{"key":"ref4","article-title":"Pushing the limit of post-training quantization by block reconstruction","author":"Li","year":"2021","journal-title":"ICLR"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.15"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1201\/9781003162810-13"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-acl.26"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00286"},{"key":"ref9","article-title":"A sparsequantized representation for near-lossless LLM weight compression","author":"Dettmers","year":"2023"},{"key":"ref10","article-title":"Oltron: Software-hardware co-design for outlier-aware quantization of LLMs with inter-\/intra-layer adaptation","author":"Xue","year":"2024","journal-title":"DAC"},{"key":"ref11","article-title":"GPTQ: Accurate post-training compression for generative pretrained transformers","author":"Frantar","year":"2023","journal-title":"ICLR"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1145\/3649329.3658498"},{"key":"ref13","article-title":"Deep compression: Compressing deep neural networks with pruning, trained quantization and Huffman coding","author":"Han","year":"2016","journal-title":"ICLR"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO50266.2020.00071"},{"key":"ref15","article-title":"K-means++: The advantages of careful seeding","author":"Arthur","year":"2006","journal-title":"Technical Report"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2022.3206148"},{"issue":"1","key":"ref17","first-page":"5485","article-title":"Exploring the limits of transfer learning with a unified text-to-text transformer","volume":"21","author":"Raffel","year":"2020","journal-title":"J. Mach. Learn. Res."},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-015-0816-y"},{"key":"ref19","article-title":"HAWQ-V3: Dyadic neural network quantization","author":"Yao","year":"2021","journal-title":"ICML"},{"key":"ref20","article-title":"Pointer sentinel mixture models","author":"Merity","year":"2016"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.3390\/su13020717"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v34i05.6239"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P19-1472"},{"key":"ref24","article-title":"Think you have solved question answering? Try ARC, the AI2 reasoning challenge","author":"Clark","year":"2018"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v34i05.6399"}],"event":{"name":"2026 Design, Automation & Test in Europe Conference (DATE)","location":"Verona, Italy","start":{"date-parts":[[2026,4,20]]},"end":{"date-parts":[[2026,4,22]]}},"container-title":["2026 Design, Automation & Test in Europe Conference (DATE)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11539023\/11539024\/11539743.pdf?arnumber=11539743","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,5]],"date-time":"2026-06-05T05:15:40Z","timestamp":1780636540000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11539743\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,4,20]]},"references-count":25,"URL":"https:\/\/doi.org\/10.23919\/date69613.2026.11539743","relation":{},"subject":[],"published":{"date-parts":[[2026,4,20]]}}}