More Like this

1. 3U-EdgeAI: Ultra-Low Memory Training, Ultra-Low Bitwidth Quantization, and Ultra-Low Latency Acceleration

   https://doi.org/10.1145/3453688.3461738  

   Chen, Yao ; Hawkins, Cole ; Zhang, Kaiqi ; Zhang, Zheng ; Hao, Cong ( June 2021 , Great Lakes Symposium on VLSI)
2. Towards Fast-Convergence, Low-Delay and Low-Complexity Network Optimization

   https://doi.org/10.1145/3154492  

   Wang, Sinong ; Shroff, Ness ( December 2017 , Proceedings of the ACM on Measurement and Analysis of Computing Systems)
3. Formation of Low Volatility Organic Compounds and Secondary Organic Aerosol from Isoprene Hydroxyhydroperoxide Low-NO Oxidation

   https://doi.org/10.1021/acs.est.5b02031  

   Krechmer, Jordan E. ; Coggon, Matthew M. ; Massoli, Paola ; Nguyen, Tran B. ; Crounse, John D. ; Hu, Weiwei ; Day, Douglas A. ; Tyndall, Geoffrey S. ; Henze, Daven K. ; Rivera-Rios, Jean C. ; et al ( September 2015 , Environmental Science & Technology)
4. Strengthening Low-resource Neural Machine Translation through Joint Learning: The Case of Farsi-Spanish [Strengthening Low-resource Neural Machine Translation through Joint Learning: The Case of Farsi-Spanish]

   https://doi.org/10.5220/0010362604750481  

   Ahmadnia, Benyamin ; Aranovich, Raul ; Dorr, Bonnie ( January 2021 , Proceedings of the 13th International Conference on Agents and Artificial Intelligence - Volume 1: NLPinAI)

   This paper describes a systematic study of an approach to Farsi-Spanish low-resource Neural Machine Translation (NMT) that leverages monolingual data for joint learning of forward and backward translation models. As is standard for NMT systems, the training process begins using two pre-trained translation models that are iteratively updated by decreasing translation costs. In each iteration, either translation model is used to translate monolingual texts from one language to another, to generate synthetic datasets for the other translation model. Two new translation models are then learned from bilingual data along with the synthetic texts. The key distinguishing feature between our approach and standard NMT is an iterative learning process that improves the performance of both translation models, simultaneously producing a higher-quality synthetic training dataset upon each iteration. Our empirical results demonstrate that this approach outperforms baselines.
5. Conditions leading to the unprecedented low Antarctic sea ice extent during the 2016 austral spring season: RECORD LOW 2016 ANTARCTIC SEA ICE EXTENT

   https://doi.org/10.1002/2017GL074691  

   Stuecker, Malte F. ; Bitz, Cecilia M. ; Armour, Kyle C. ( September 2017 , Geophysical Research Letters)