{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,27]],"date-time":"2026-05-27T06:05:44Z","timestamp":1779861944474,"version":"3.53.1"},"reference-count":79,"publisher":"IEEE","license":[{"start":{"date-parts":[[2026,4,26]],"date-time":"2026-04-26T00:00:00Z","timestamp":1777161600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,4,26]],"date-time":"2026-04-26T00:00:00Z","timestamp":1777161600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/100000028","name":"Semiconductor Research Corporation","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100000028","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,4,26]]},"DOI":"10.1109\/ispass69572.2026.00062","type":"proceedings-article","created":{"date-parts":[[2026,5,26]],"date-time":"2026-05-26T19:39:19Z","timestamp":1779824359000},"page":"589-600","source":"Crossref","is-referenced-by-count":0,"title":["Towards System-2 AI: Workloads and Characterizations of Energy-Based Models"],"prefix":"10.1109","author":[{"given":"Hanchen","family":"Yang","sequence":"first","affiliation":[{"name":"Georgia Institute of Technology,Atlanta,GA,USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jiayi","family":"Qian","sequence":"additional","affiliation":[{"name":"Georgia Institute of Technology,Atlanta,GA,USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zishen","family":"Wan","sequence":"additional","affiliation":[{"name":"Georgia Institute of Technology,Atlanta,GA,USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jingtian","family":"Dang","sequence":"additional","affiliation":[{"name":"Georgia Institute of Technology,Atlanta,GA,USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ziwei","family":"Li","sequence":"additional","affiliation":[{"name":"Georgia Institute of Technology,Atlanta,GA,USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yilun","family":"Du","sequence":"additional","affiliation":[{"name":"Harvard University,Cambridge,MA,USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tushar","family":"Krishna","sequence":"additional","affiliation":[{"name":"Georgia Institute of Technology,Atlanta,GA,USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","article-title":"Do phd-level llms truly grasp elementary addition? probing rule learning vs. memorization in large language models","author":"Yan","year":"2025","journal-title":"arXiv preprint arXiv:2504.05262"},{"key":"ref2","first-page":"arXiv\u20132412","article-title":"Escapebench: Pushing language models to think outside the box","author":"Qian","year":"2024","journal-title":"arXiv e-prints"},{"key":"ref3","volume-title":"Thinking, fast and slow","author":"Kahneman","year":"2011"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA61900.2025.00064"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1109\/ISPASS61541.2024.00033"},{"key":"ref6","article-title":"From system 1 to system 2: A survey of reasoning large language models","author":"Li","year":"2025","journal-title":"arXiv preprint arXiv:2502.17419"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1109\/TCASAI.2024.3462692"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA68181.2026.11408458"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1007\/s10462-023-10448-w"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v35i17.17765"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1093\/nsr\/nwac035"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1109\/JSSC.2026.3658288"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.3233\/SW-223228"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1145\/3760250.3762235"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1109\/JRPROC.1961.287775"},{"key":"ref16","article-title":"Efficient processing of neuro-symbolic ai: A tutorial and cross-layer co-design case study","volume-title":"Proceedings of the International Conference on Neuro-symbolic Systems","author":"Wan"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1109\/DAC63849.2025.11133088"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1109\/CODES-ISSS60120.2024.00012"},{"key":"ref19","article-title":"Cross-layer design of vectorsymbolic computing: Bridging cognition and brain-inspired hardware acceleration","author":"Du","year":"2025","journal-title":"arXiv preprint arXiv:2508.14245"},{"key":"ref20","article-title":"Does reinforcement learning really incentivize reasoning capacity in llms beyond the base model?","author":"Yue","year":"2025","journal-title":"arXiv preprint arXiv:2504.13837"},{"key":"ref21","article-title":"Reasoning with sampling: Your base model is smarter than you think","author":"Karan","year":"2025","journal-title":"arXiv preprint arXiv:2510.14901"},{"key":"ref22","article-title":"How to train your energy-based models","author":"Song","year":"2021","journal-title":"arXiv preprint arXiv:2101.03288"},{"key":"ref23","first-page":"8489","article-title":"Reduce, reuse, recycle: Compositional generation with energy-based diffusion models and mcmc","volume-title":"International conference on machine learning","author":"Du"},{"key":"ref24","article-title":"Learning iterative reasoning through energy diffusion","author":"Du","year":"2024","journal-title":"arXiv preprint arXiv:2406.11179"},{"key":"ref25","article-title":"Learning Generalizable Systems by Learning Composable Energy Landscapes","author":"Du","year":"2025"},{"key":"ref26","article-title":"Energy-based transformers are scalable learners and thinkers","author":"Gladstone","year":"2025","journal-title":"arXiv preprint arXiv:2507.02092"},{"key":"ref27","article-title":"Generalizable reasoning through compositional energy minimization","author":"Oarga","year":"2025","journal-title":"arXiv preprint arXiv:2510.20607"},{"key":"ref28","article-title":"Compositional scene understanding through inverse generative modeling","author":"Wang","year":"2025","journal-title":"arXiv preprint arXiv:2505.21780"},{"key":"ref29","article-title":"Compositional generative modeling: A single model is not all you need","author":"Du","year":"2024","journal-title":"arXiv preprint arXiv:2402.01103"},{"key":"ref30","article-title":"Compositional diffusion-based continuous constraint solvers","author":"Yang","year":"2023","journal-title":"arXiv preprint arXiv:2309.00966"},{"key":"ref31","first-page":"5570","article-title":"Learning iterative reasoning through energy minimization","volume-title":"International Conference on Machine Learning","author":"Du"},{"key":"ref32","first-page":"1","article-title":"Energy-based models for continual learning","volume-title":"Conference on lifelong learning agents","author":"Li"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-19790-1_26"},{"key":"ref34","article-title":"Model based planning with energy based models","author":"Du","year":"2021"},{"key":"ref35","first-page":"15608","article-title":"Unsupervised learning of compositional energy concepts","volume":"34","author":"Du","year":"2021","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref36","first-page":"23166","article-title":"Learning to compose visual relations","volume":"34","author":"Liu","year":"2021","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref37","article-title":"Compositional visual generation and inference with energy based models","author":"Du","year":"2020","journal-title":"arXiv preprint arXiv:2004.06030"},{"key":"ref38","article-title":"Improved contrastive divergence training of energy based models","author":"Du","year":"2020","journal-title":"arXiv preprint arXiv:2012.01316"},{"key":"ref39","article-title":"Exponential family estimation via adversarial dynamics embedding","volume":"32","author":"Dai","year":"2019","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref40","article-title":"Implicit generation and modeling with energy based models","volume":"32","author":"Du","year":"2019","journal-title":"Advances in neural information processing systems"},{"key":"ref41","article-title":"Model based planning with energy based models","volume-title":"Conference on Robot Learning","author":"Du"},{"key":"ref42","first-page":"681","article-title":"Bayesian learning via stochastic gradient langevin dynamics","volume-title":"Proceedings of the 28th International Conference on Machine Learning (ICML)","author":"Welling"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.2307\/3318418"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1201\/b10905-6"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1162\/089976602760128018"},{"key":"ref46","article-title":"Equilibrium matching: Generative modeling with implicit energy-based models","author":"Wang","year":"2025","journal-title":"arXiv preprint arXiv:2510.02300"},{"key":"ref47","first-page":"182","article-title":"Regularizing model-based planning with energy-based models","volume-title":"Conference on Robot Learning","author":"Boney"},{"key":"ref48","article-title":"Compositional generative inverse design","volume-title":"The Twelfth International Conference on Learning Representations","author":"Wu"},{"key":"ref49","first-page":"8748","article-title":"Learning transferable visual models from natural language supervision","volume-title":"International conference on machine learning","author":"Radford"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1109\/ICRA.2011.5980391"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1609\/icaps.v30i1.6739"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.52202\/068431-2679"},{"key":"ref53","article-title":"Predicting stable configurations for semantic placement of novel objects","volume-title":"Proceedings of the Conference on Robot Learning (CoRL)","author":"Paxton"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.15607\/RSS.2023.XIX.061"},{"key":"ref55","article-title":"Introducing PyTorch Profiler - the new and improved performance tool","author":"Lukiyanov","year":"2021"},{"issue":"5","key":"ref56","first-page":"2","article-title":"The cifar-10 dataset","volume":"55","author":"Krizhevsky","year":"2014"},{"key":"ref57","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.215"},{"key":"ref58","first-page":"6545","article-title":"Satnet: Bridging deep learning and logical reasoning using a differentiable satisfiability solver","volume-title":"International Conference on Machine Learning","author":"Wang"},{"key":"ref59","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2015.425"},{"key":"ref60","article-title":"Deep compression: Compressing deep neural networks with pruning, trained quantization and huffman coding","author":"Han","year":"2015","journal-title":"arXiv preprint arXiv:1510.00149"},{"key":"ref61","article-title":"Gptq: Accurate post-training quantization for generative pretrained transformers","author":"Frantar","year":"2023","journal-title":"arXiv preprint arXiv:2210.17323"},{"key":"ref62","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.l803.03635"},{"key":"ref63","article-title":"Hawq-v2: Hessian aware quantization of neural networks with mixed precision","volume-title":"European Conference on Computer Vision","author":"Liu"},{"key":"ref64","article-title":"Sparse quantized representation for large language models","author":"Dettmers","year":"2023","journal-title":"arXiv preprint arXiv:2306.03078"},{"key":"ref65","article-title":"H2o: Heavy-hitter oracle for efficient kv cache compression in llms","author":"Peng","year":"2023","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref66","article-title":"Flashattention-2: Faster attention with better parallelism and work partitioning","volume-title":"International Conference on Machine Learning","author":"Dao"},{"key":"ref67","article-title":"Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity","volume-title":"International Conference on Machine Learning","author":"Fedus"},{"key":"ref68","doi-asserted-by":"publisher","DOI":"10.52202\/075280-0441"},{"key":"ref69","article-title":"Awq: Activation-aware weight quantization for llms","author":"Lin","year":"2023","journal-title":"arXiv preprint arXiv:2306.00978"},{"key":"ref70","doi-asserted-by":"publisher","DOI":"10.1109\/jssc.2016.2616357"},{"key":"ref71","doi-asserted-by":"publisher","DOI":"10.1145\/3360307"},{"key":"ref72","article-title":"Efficient large-scale language model training on gpu clusters","volume-title":"International Conference on Machine Learning","author":"Narayanan"},{"key":"ref73","article-title":"Spatten: Efficient sparse attention architecture for large language models","volume-title":"International Symposium on Computer Architecture (ISCA)","author":"Kim"},{"key":"ref74","article-title":"Sparta: Efficient structured sparse transformer acceleration","author":"Liu","year":"2023","journal-title":"MICRO"},{"key":"ref75","first-page":"2098","article-title":"Decoupled parallel backpropagation with convergence guarantee","volume-title":"International Conference on Machine Learning","author":"Huo"},{"key":"ref76","article-title":"Hogwild!: A lock-free approach to parallelizing stochastic gradient descent","volume":"24","author":"Recht","year":"2011","journal-title":"Advances in neural information processing systems"},{"key":"ref77","doi-asserted-by":"publisher","DOI":"10.1145\/3297858.3304019"},{"key":"ref78","article-title":"Mc 2 a: Enabling algorithm-hardware co-design for efficient markov chain monte carlo acceleration","author":"Zhao","year":"2025","journal-title":"arXiv preprint arXiv:2507.12935"},{"key":"ref79","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA53966.2022.00012"}],"event":{"name":"2026 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS)","location":"Seoul, Korea, Republic of","start":{"date-parts":[[2026,4,26]]},"end":{"date-parts":[[2026,4,28]]}},"container-title":["2026 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11527204\/11527232\/11527268.pdf?arnumber=11527268","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,27]],"date-time":"2026-05-27T05:39:40Z","timestamp":1779860380000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11527268\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,4,26]]},"references-count":79,"URL":"https:\/\/doi.org\/10.1109\/ispass69572.2026.00062","relation":{},"subject":[],"published":{"date-parts":[[2026,4,26]]}}}