Run-Time Efficient RNN Compression for Inference on Edge Devices
Open accessThe DOI matches its Crossref record; 9 of the 9 references it cites were found in a registry or an index.
Abstract
Recurrent neural networks can be large and compute-intensive, yet many applications that benefit from RNNs run on small devices with very limited compute and storage capabilities while still having run-time constraints. As a result, there is a need for compression techniques that can achieve significant compression without negatively impacting inference run-time and task accuracy. This paper explores a new compressed RNN cell implementation called Hybrid Matrix Decomposition (HMD) that achieves this dual objective. HMD creates dense matrices that results in output features where the upper sub-vector has "richer" features while the lower sub vector has "constrained" features". On the benchmarks evaluated in this paper, this results in faster inference runtime than pruning and better accuracy than matrix factorization for compression factors of 2-4x.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition
Ordóñez, F., & Roggen, D. (2016) · Sensors · doi:10.3390/s16010115 · Cited by 2,761 · cited 1×
“The second HAR network is based on the work in [12].”
BackgroundSpeech Commands: A Dataset for Limited-Vocabulary Speech Recognition
Warden, P. (2018) · arXiv · doi:10.48550/arxiv.1804.03209 · Cited by 1,033
To prune, or not to prune: exploring the efficacy of pruning for model compression
Zhu, M., & Gupta, S. (2017) · arXiv · doi:10.48550/arxiv.1710.01878 · Cited by 805
Collecting complex activity datasets in highly rich networked sensor environments
Roggen, D., Calatroni, A., Rossi, M., Holleczek, T., Forster, K., Troster, G., Lukowicz, P., Bannach, D., Pirkl, G., Ferscha, A., et al. (2010) · 2010 Seventh International Conference on Networked Sensing Systems (INSS) · doi:10.1109/inss.2010.5573462 · Cited by 788
Deep, Convolutional, and Recurrent Models for Human Activity Recognition Using Wearables
Hammerla, N. Y., Halloran, S., & Ploetz, T. (2016) · arXiv · doi:10.48550/arxiv.1604.08880 · Cited by 606
Low-rank matrix factorization for Deep Neural Network training with high-dimensional output targets
Sainath, T. N., Kingsbury, B., Sindhwani, V., Arisoy, E., & Ramabhadran, B. (2013) · 2013 IEEE International Conference on Acoustics, Speech and Signal Processing · doi:10.1109/icassp.2013.6638949 · Cited by 593 · cited 1×
“However, LMF has had mixed results in maintaining model accuracy [1, 5, 12].”
ResultFactorization tricks for LSTM networks
Kuchaiev, O., & Ginsburg, B. (2017) · arXiv · doi:10.48550/arxiv.1703.10722 · Cited by 99 · cited 2×
“Low Rank Matrix Factorization (LMF) [8] expresses a larger matrix A of dimension m × n as a product of two smaller matrices U and V of dimension m × d and d × n , respectively.”
BackgroundCompressing RNNs for IoT devices by 15-38x using Kronecker Products
Thakker, U., Beu, J., Gope, D., Zhou, C., Fedorov, I., Dasika, G., & Mattina, M. (2019) · arXiv · doi:10.48550/arxiv.1906.02876 · Cited by 30 · cited 1×
“ompression trade-offs of HMD with network pruning and matrix factorization 2 Related Work The research in NN compression can be categorized under 4 topics-Pruning[7,21],structuredmatrixbasedtechniques[3, 13, 14, 16], quantization [4] and tensor decomposition [8, 19]. HMD belongs in the structured matrix category. We c…”
MethodSelf-citationCompressing Language Models using Doped Kronecker Products
Thakker, U., Whatmough, P. N., Liu, Z.-G., Mattina, M., & Beu, J. (2020) · arXiv · doi:10.48550/arxiv.2001.08896 · Cited by 7
Self-citation
Research works printed but not linked: 14
- narang, 2017, Block-sparse recurrent neural networks, CoRR, abs 1711 2782
- sindhwani, 2015, Structured transforms for small-footprint deep learning, Advances in neural information processing systems, 28, 3088
- 10.1109/IJCNN.2017.7966420 · DOI · Paper page
- kuchaiev, 2017, CoRR, abs 1703 10722
- 10.1145/3123939.3124552 · DOI · Paper page
- hammerla, 2016, IJCA-2016
- zaremba, 2014, CoRR, abs 1409 2329
- 10.1007/978-3-319-69900-4_44 · DOI
- ding, 2018, Structured weight matrices-based hardware accelerators in deep neural networks: Fpgas and asics, Proceedings of the 2018 on Great Lakes Symposium on VLSI ser GLSVLSI '18, 353, 10.1145/3194554.3194625 · DOI
- chen, 2018, Adaptive mixture of low-rank factorizations for compact neural modeling, Advances in neural information processing systems (CDNNRIA workshop), 10.1007/978-3-030-04167-0 · DOI
- li, 2018, Efficient recurrent neural networks using structured matrices in fpgas, CoRR, abs 1803 7661
- thomas, 2018, Learning compressed transforms with low displacement rank, Advances in neural information processing systems, 31, 9066
- han, 2016, Deep compression: Compressing deep neural networks with pruning, trained quantization and huffman coding, International Conference on Learning Representations (ICLR)
- thakker, 2019, Measuring scheduling efficiency of rnns for NLP applications, CoRR, abs 1904 3302
Cited by
Edge Intelligence: The Confluence of Edge Computing and Artificial Intelligence
Deng, S., Zhao, H., Fang, W., Yin, J., Dustdar, S., & Zomaya, A. Y. (2020) · IEEE Internet of Things Journal · doi:10.1109/jiot.2020.2984887 · Cited by 1,147
A Survey on Federated Learning for Resource-Constrained IoT Devices
Imteaj, A., Thakker, U., Wang, S., Li, J., & Amini, M. H. (2022) · IEEE Internet of Things Journal · doi:10.1109/jiot.2021.3095077 · Cited by 774
“A potential direction can be found in [179].”
Self-citationBanbury, C., Zhou, C., Fedorov, I., Navarro, R. M., Thakker, U., Gope, D., Reddi, V. J., Mattina, M., & Whatmough, P. N. (2020) · arXiv · doi:10.48550/arxiv.2010.11267 · Cited by 183
Self-citationCompressing RNNs for IoT devices by 15-38x using Kronecker Products
Thakker, U., Beu, J., Gope, D., Zhou, C., Fedorov, I., Dasika, G., & Mattina, M. (2019) · arXiv · doi:10.48550/arxiv.1906.02876 · Cited by 30
“The size of typical RNN layers can pro-hibit their deployment on IoT devices or reduce execution efficiency (Thakker et al., 2019b).”
BackgroundMethodSelf-citationCompressing RNNs to Kilobyte Budget for IoT Devices Using Kronecker Products
Thakker, U., Fedorov, I., Zhou, C., Gope, D., Mattina, M., Dasika, G., & Beu, J. (2021) · ACM Journal on Emerging Technologies in Computing Systems · doi:10.1145/3440016 · Cited by 8
Self-citation
Metadata
- Authors
- Thakker, U., Beu, J., Gope, D., Dasika, G., & Mattina, M.
- Venue
- 2019 2nd Workshop on Energy Efficient Machine Learning and Cognitive Computing for Embedded Applications (EMC2)
- Publisher
- IEEE
- Published
- 2019-02
- Pages
- 26-30
- OpenAlex
- W2949896761
- Cited by
- 24
- References
- 9
- FWCI
- 1.43
- Type
- conference paper
- Language
- en
- Open access
- green
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Run-Time Efficient RNN Compression for Inference on Edge Devices. https://app.cochat.ai/papers/10.1109/emc249363.2019.00013
- The paper
- Thakker, U., Beu, J., Gope, D., Dasika, G., & Mattina, M. (2019). Run-Time Efficient RNN Compression for Inference on Edge Devices. 2019 2nd Workshop on Energy Efficient Machine Learning and Cognitive Computing for Embedded Applications (EMC2). https://doi.org/10.1109/emc249363.2019.00013
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.