Publications

All Learning Algorithms & Systems Model Serving & Inference Edge & On-Device ML Agentic Systems AI for Systems Miscellaneous

2026

  • SenSys'26 WildFiT: Autonomous In-situ Model Adaptation for Resource-Constrained IoT Systems.
    Mohammad Mehdi Rastikerdar, Jin Huang, Hui Guan, Deepak Ganesan
    ACM/IEEE International Conference on Embedded Artificial Intelligence and Sensing Systems, 2026
  • MLSys'26 ProTrain: Efficient LLM Training via Automatic Memory Management. [PDF]
    Hanmei Yang, Jin Zhou, Yao Fu, Xiaoquan Wang, Ramine Roane, Hui Guan, Tongping Liu
    The 9th Annual Conference on Machine Learning and Systems (MLSys 2026), Bellevue, WA, May 18-22, 2026
  • PacificVis'26 A Four-Stage Framework of Visual Complexity and Trust as Mediated by Effort.
    Kylie Lin, Hui Guan, David N Rapp, Cindy Xiong Bearfield
    IEEE Pacific Visualization Conference, 2026
  • MMSys'26 Atom: Efficient On-Device Video-Language Pipelines Through Modular Reuse.
    Kunjal Panchal, Saayan Mitra, Somdeb Sarkhel, Haoliang Wang, Ishita Dasgupta, Gang Wu, Hui Guan
    ACM Multimedia Systems Conference, 2026

2025

  • arXiv'25 Automated Cloud Infrastructure-as-Code Reconciliation with AI Agents. [PDF]
    Zhenning Yang, Hui Guan, Victor Nicolet, Brandon Paulsen, Joey Dodds, Daniel Kroening, Ang Chen
    arXiv, 2025
  • VLDB'25 Graph neural network training systems: A performance comparison of full-graph and mini-batch. [PDF]
    Saurabh Bajaj, Hojae Son, Juelin Liu, Hui Guan, and Marco Serafini
    Proceedings of the VLDB Endowment, 2025
  • EXAIT@ICML'25 Reimagining Parameter Space Exploration with Diffusion Models. [PDF]
    Lijun Zhang, Xiao Liu, Hui Guan
    First Exploration in AI Today Workshop at ICML (EXAIT at ICML 2025)
  • TKDD'25 Recurrent Neural Networks Meet Context-Free Grammar: Two Birds with One Stone. [PDF]
    Hui Guan, Umang Chaudhary, Yuanchao Xu, Lin Ning, Lijun Zhang, Xipeng Shen
    ACM Transactions on Knowledge Discovery from Data, Volume 20, Issue 2, 2025
  • Neurocomputing@25 Attacking all tasks at once using adversarial examples in multi-task learning. [PDF]
    Lijun Zhang, Xiao Liu, Kaleel Mahmood, Caiwen Ding, and Hui Guan
    Neurocomputing, 2025
  • ARITH'25 An Empirical Study of Microscaling Formats for Low-Precision LLM Training. [PDF]
    Hanmei Yang, Summer Deng, Amit Nagpal, Maxim Naumov, Mohammad Janani, Tongping Liu, Hui Guan
    32nd IEEE Symposium on Computer Arithmetic, Jun 23-25, 2025
  • MLSys'25 SPA: Scaling Graph Neural Network Training on Large Graphs via Probablistic Splitting. [PDF]
    Sandeep Polisetty, Juelin Liu, Yi Fung, Seung-Hwan Lim, Hui Guan, Marco Serafini
    The Eighth Annual Conference on Machine Learning and Systems, Santa Clara, May 12-15, 2025 (Acceptance Rate = 22% (61/271))
  • MLSys'25 DiffServe: Efficiently Serving Text-to-Image Diffusion Models with Query-Aware Model Scaling. [PDF]
    Sohaib Ahmad (co-first author), Qizheng Yang (co-first author), Haoliang Wang, Ramesh K. Sitaraman, and Hui Guan
    The Eighth Annual Conference on Machine Learning and Systems, Santa Clara, May 12-15, 2025. (Acceptance Rate = 22% (61/271))
  • CHI EA'25 What Makes a Visualization Visually Complex?. [PDF]
    Kylie Lin, Sean Sheng-tse Ru, David N Rapp, Hui Guan, Cindy Xiong Bearfield
    Extended Abstracts of the CHI Conference on Human Factors in Computing Systems, 2025

2024

  • NeurIPS'24 Thinking Forward: Memory-Efficient Federated Finetuning of Language Models. [PDF] [Code]
    Kunjal Panchal, Nisarg Parikh, Sunav Choudhary, Lijun Zhang, Yuriy Brun, Hui Guan
    NeurIPS '24, Mon, Dec 9, 2024 – Sun, Dec 15, 2024, Vancouver
  • NeurIPS'24 Attack-Resilient Image Watermarking Using Stable Diffusion. [PDF] [Code]
    Lijun Zhang, Xiao Liu, Antoni Viros i Martin, Cindy Xiong Bearfield, Yuriy Brun, Hui Guan
    NeurIPS '24, Mon, Dec 9, 2024 – Sun, Dec 15, 2024, Vancouver
  • MLforSys@NeurIPS'24 Understanding and Alleviating Memory Issue in RLHF for LLMs. [PDF]
    Jin Zhou, Hanmei Yang, Steven Jiaxun Tang, Mingcan Xiang, Hui Guan, Tongping Liu
    NeurIPS'24 Workshop MLforSys, Dec 14, 2024, Vancouver
  • AI4Mat@NeurIPS'24 Integrating Graph Neural Networks and Many-Body Expansion Theory for Potential Energy Surfaces. [PDF]
    Siqi Chen, Zhiqiang Wang, Xianqi Deng, Yili Shen, Cheng-Wei Ju, Jun Yi, Lin Xiong, Guo Ling, Dieaa Alhmoud, Hui Guan, Zhou Lin
    NeurIPS'24 Workshop AI4Mat, Dec 15, 2024, Vancouver
  • ACM MM'24 AdapMTL: Adaptive Pruning Framework for Multitask Learning Model. [PDF]
    Mingcan Xiang, Jiaxun Tang, Qizheng Yang, Hui Guan, Tongping Liu
    ACM MM '24, October 28-November 1, 2024, Melbourne, VIC, Australia
    https://doi.org/10.1145/3664647.3681426
  • MIPR'24 Structured Pruning for Multi-Task Deep Neural Networks. [PDF]
    Siddhant Garg, Lijun Zhang, Hui Guan
    International Conference on Multimedia Information Processing and Retrieval, August 07, 2024
  • MobiSys'24 CACTUS: Dynamically Switchable Context-aware micro-Classifiers for Efficient IoT Inference. [PDF] [Code]
    Mohammad Mehdi Rastikerdar, Jin Huang, Shiwei Fang, Hui Guan, Deepak Ganesan
    The 22nd ACM International Conference on Mobile Systems, Applications, and Services (MobiSys), Tokyo, Japan, June 3-7, 2024
    https://doi.org/10.1145/3643832.3661888
  • IEEE Access'24 Information-Enhanced Graph Neural Network for Transcending Homophily Barriers. [PDF]
    Xiao Liu, Lijun Zhang, Hui Guan
    In IEEE Access, 2024
  • HPDC'24 Loki: A System for Serving ML Inference Pipelines with Hardware and Accuracy Scaling. [PDF]
    Sohaib Ahmad, Hui Guan, Ramesh K. Sitaraman
    The 33rd International Symposium on
  • EuroSys'24 GMorph: Accelerating Multi-DNN Inference via Model Fusion. [PDF] [Code]
    Qizheng Yang, Tianyi Yang, Mingcan Xiang, Lijun Zhang, Haoliang Wang, Marco Serafini, Hui Guan
    The 2024 European Conference on Computer Systems (EuroSys), April 22-25, 2024
    https://doi.org/10.1145/3627703.3650074
  • ASPLOS'24 Proteus: A High-Throughput Inference-Serving System with Accuracy Scaling. [PDF] [Code]
    Sohaib Ahmad, Hui Guan, Brain D. Friedman, Thomas Williams, Ramesh K. Sitaraman, Thomas Woo
    The 2024 ACM Conference on Architectural Support for Programming Languages and Operating Systems, April 27-May 1, 2024
    https://doi.org/10.1145/3617232.3624849

2023

  • NeurIPS'23 Flow: Per-instance Personalized Federated Learning. [PDF] [Code]
    Kunjal Panchal, Sunav Choudhary, Nisarg Parikh, Lijun Zhang, Hui Guan
    The 2023 Conference on Neural Information Processing Systems, Dec. 10-16, 2023
  • PACT'23 GraphMini: Accelerating Graph Pattern Matching Using Auxiliary Graphs. [PDF] [Code]
    Juelin Liu, Sandeep Polisetty, Hui Guan, Marco Serafini
    The 32nd International Conference on Parallel Architectures and Compilation Techniques, Oct. 21-25, 2023
  • MobiCom'23 Re-thinking computation offload for efficient inference on IoT devices with duty-cycled radios. [PDF]
    Jin Huang, Hui Guan, Deepak Ganesan
    The 29th International Conference on Mobile Computing and Networking, Madrid, Spain, Oct. 2-6, 2023
  • ICML'23 Flash: Concept Drift Adaptation in Federated Learning. [PDF]
    Kunjal Panchal, Sunav Choudhary, Subrata Mitra, Koyel Mukherjee, Somdeb Sarkhel, Saayan Mitra, Hui Guan
    40th International Conference on Machine Learning, Jul. 23-29, 2023
  • ICML'23 Automatically marginalized MCMC in probabilistic programming. [PDF]
    Jinlin Lai, Javier Burroni, Hui Guan, Daniel Sheldon
    40th International Conference on Machine Learning, Jul. 23-29, 2023
  • TNNLS'23 A Tree-Structured Multi-Task Model Architectures Recommendation System. [PDF] [Code]
    Lijun Zhang, Xiao Liu, Hui Guan
    IEEE Transactions on Neural Networks and Learning Systems, 2023
  • ISMM'23 NUMAlloc: A Faster NUMA Memory Allocator. [PDF]
    Hanmei Yang, Xin Zhao, Jin Zhou, Wei Wang, Sandip Kundu, Bo Wu, Hui Guan, and Tongping Liu
    ACM SIGPLAN International Symposium on Memory Management, 2023
  • IEEE Access'23 An Alternative Hard-Parameter Sharing Paradigm for Multi-Domain Learning. [PDF]
    Lijun Zhang, Qizheng Yang, Xiao Liu, Hui Guan
    In IEEE Access, 2023

2022

  • NeurIPS'22 AutoMTL: A Programming Framework for Automating Efficient Multi-Task Learning. [PDF] [Code]
    Lijun Zhang, Xiao Liu, Hui Guan
    36th Conference on Neural Information Processing Systems (NeurIPS 2022), November 28, 2022. (Acceptance rate: 25.6%)
  • VLDB'22 COMET: A Novel Memory-Efficient Deep Learning Training Framework by Using Error-Bounded Lossy Compression. [PDF]
    Sian Jin, Chengming Zhang, Xintong Jiang, Yunhe Feng, Hui Guan, Guanpeng Li, Shuaiwen Leon Song, and Dingwen Tao
    In International Conference on Very Large Data Bases, 2022
  • ICME'22 Rethinking Hard-Parameter Sharing in Multi-Domain Learning. [PDF]
    Lijun Zhang, Qizheng Yang, Xiao Liu, Hui Guan
    IEEE INTERNATIONAL CONFERENCE ON MULTIMEDIA AND EXPO (ICME), Taipei, Taiwan, July 18-22, 2022. (Acceptance rate: 29%)
  • AutoML'22 A Tree-Structured Multi-Task Model Recommender. [PDF] [Code] [Teaser] [Video]
    Lijun Zhang, Xiao Liu, Hui Guan
    1st International Conference on Automated Machine Learning, July 25-27, 2022. (Acceptance rate: 19.2%)
  • AI4Science@ICML'22 Improving Subgraph Representation Learning via Multi-View Augmentation. [PDF] [talk]
    Yili Shen, Jiaxu Yan, Cheng-Wei Ju, Jun Yi, Zhou Lin, Hui Guan
    [ICML 2022 AI4Science Workshop](http://ai4science.net/icml22/)
  • CrossFL'22 Flow: Fine-grained Personalized Federated Learning through Dynamic Routing. [PDF] [Poster]
    Kunjal Panchal, Hui Guan
    [CrossFL 2022 Workshop @ MLSys'22](https://crossfl2022.github.io/program/)
  • CGO'22 Enabling Near Real-Time NLU-Driven Natural Language Programming through Dynamic Grammar Graph-Based Translation. [PDF]
    Zifan Nan, Xipeng Shen, Hui Guan
    The 2022 International Symposium on Code Generation and Optimization (CGO), Seoul, South Korea, 2022

2021

  • ICDM'21 Recurrent Neural Networks Meet Context-Free Grammar: Two Birds with One Stone. [PDF]
    Hui Guan, Umang Chaudhary, Yuanchao Xu, Lin Ning, Lijun Zhang, and Xipeng Shen
    In IEEE International Conference on Data Mining, 2021 (short paper). (Acceptance rate: 20% (198/990))
  • OSR'21 Scalable Graph Neural Network Training: The Case for Sampling. [PDF]
    Marco Serafini, Hui Guan
    In ACM SIGOPS Operating Systems Review, 2021
  • MCHPC'21 FreeLunch: Compression-based GPU Memory Management for Convolutional Neural Networks. [PDF]
    Shaurya Patel, Tongping Liu, Hui Guan
    In MCHPC'21 Workshop
  • ICS'21 NumaPerf: Predictive and Comprehensive NUMA Profiling. [PDF]
    Xin Zhao, Jin Zhou, Hui Guan, Wei Wang, Xu Liu, Tongping Liu
    In Proceedings of International Conference on Supercomputing, 2021. (Acceptance rate: 25% (39/157))
  • CACM'21 CoCoPIE: Enabling Real-Time AI on Off-the-Shelf Mobile Devices via Compression-Compilation Co-Design. [PDF]
    Hui Guan, Shaoshan Liu, Xiaolong Ma, Wei Niu, Bin Ren, Xipeng Shen, Yanzhi Wang, Pu Zhao. (Authors in Alphabetical Order)
    In Communications of the ACM, 2021
  • InformationSystems'21 Reuse-Centric K-Means Configuration.
    Lijun Zhang, Hui Guan, Yufei Ding, Xipeng Shen, Hamid Krim
    Information Systems, 2021
  • CC'21 Deep NLP-Based Co-Evolvement for Synthesizing Code Analysis from Natural Language. [PDF]
    Zifan Nan, Hui Guan, Xipeng Shen, and Chunhua Liao
    In The ACM SIGPLAN 2021 International Conference on Compiler Construction, 2021

2020

  • TPDS'20 An Automatic Synthesizer of Advising Tools for High Performance Computing. [PDF]
    Hui Guan, Xipeng Shen, and Hamid Krim
    In IEEE Transactions on Parallel and Distributed Systems (TPDS), 2020
  • FSE'20 HISyn: Human Learning-Inspired Natural Language Programming. [PDF]
    Zifan Nan, Hui Guan, Xipeng Shen
    In The ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering, Sacramento, California, United States, November 2020. (Acceptance rate: 101/360=28%)
  • MLSys'20 FLEET: Flexible Efficient Ensemble Training for Heterogeneous Deep Neural Networks. [PDF]
    Hui Guan, Laxmikant Kishor Mokadam, Xipeng Shen, Robert Patton
    MLSys'20. (Acceptance rate: 20.0% (34/170))

2019

  • NeurIPS'19 In-Place Zero-Space Memory Protection for CNN. [PDF]
    Hui Guan, Lin Ning, Zhen Lin, Xipeng Shen, Huiyang Zhou, and Seung-Hwan Lim
    In Advances in Neural Information Processing Systems, pp. 5735-5744. 2019. (Acceptance rate: 21.2% (1428/6743))
  • MLSys@NeurIPS'19 Post-Training 4-bit Quantization on Embedding Tables. [PDF]
    Hui Guan, Andrey Malevich, Jiyan Yang, Jongsoo Park, and Hector Yuen
    MLSys Workshop on Systems for ML @ NeurIPS, 2019
  • PLDI'19 Wootz: a Compiler-based Framework for Fast CNN Pruning via Composability. [PDF]
    Hui Guan, Xipeng Shen, and Seung-Hwan Lim
    In Proceedings of the 40th ACM SIGPLAN Conference on Programming Language Design and Implementation, pp. 717-730. ACM, 2019. (Acceptance rate: 27.7% (76/274))
  • ICDE'19 Adaptive Deep Reuse: Accelerating CNN Training on the Fly. [PDF]
    Lin Ning, Hui Guan, and Xipeng Shen
    In 2019 IEEE 35th International Conference on Data Engineering (ICDE), pp. 1538-1549. IEEE, 2019. (Acceptance rate: 18%)

2018

  • SC'18 Exploring Flexible Communications for Streamlining DNN Ensemble Training Pipelines. [PDF]
    Randall Pittman, Hui Guan, Xipeng Shen, Seung-Hwan Lim, and Robert M. Patton
    In Proceedings of the International Conference for High Performance Computing, Networking, Storage, and Analysis, p. 64. IEEE, 2018. (Acceptance rate: 23%)
  • ICDE'18 Reuse-Centric K-Means Configuration. [PDF]
    Hui Guan, Yufei Ding, Xipeng Shen, and Hamid Krim
    In 2018 IEEE 34th International Conference on Data Engineering (ICDE), pp. 1224-1227. IEEE, 2018. (short paper) (Acceptance rate: 23%)
  • SysML'18 TOP: A Compiler-Based Framework for Optimizing Machine Learning Algorithms through Generalized Triangle Inequality.
    Yufei Ding, Lin Ning, Hui Guan, Xipeng Shen, Madanlal Musuvathi, Todd Mytkowicz
    SysML, Feb 16th, 2018, Stanford University, 2018

2017

  • SC'17 Egeria: a Framework for Automatic Synthesis of HPC Advising Tools through Multi-Layered Natural Language Processing. [PDF]
    Hui Guan, Xipeng Shen, and Hamid Krim
    In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, p. 10. ACM, 2017. (Acceptance rate: 18% (61/327))
  • PLDI'17 Generalizations of the Theory and Deployment of Triangular Inequality for Compiler-Based Strength Reduction. [PDF]
    Yufei Ding, Lin Ning, Hui Guan, and Xipeng Shen
    In Proceedings of the 38th ACM SIGPLAN Conference on Programming Language Design and Implementation, pp. 33-48. ACM, 2017. (Acceptance rate: 15% (47/322))

2016

  • SPAWC'16 A topological collapse for document summarization. [PDF]
    Hui Guan, Wen Tang, Hamid Krim, James Keiser, Andrew Rindos, and Radmila Sazdanovic
    In 2016 IEEE 17th International Workshop on Signal Processing Advances in Wireless Communications (SPAWC), pp. 1-5. IEEE, 2016