Results 11 to 20 of about 1,119,437 (280)
Low-Resource Neural Headline Generation [PDF]
Proceedings of the Workshop on New Frontiers in Summarization, 2017 Recent neural headline generation models have shown great results, but are generally trained on very large datasets. We focus our efforts on improving headline quality on smaller datasets by the means of pretraining. We propose new methods that enable pre-training all the parameters of the model and utilize all available text, resulting in improvements
Tilk, Ottokar, Alumäe, Tanel
openaire +3 more sources
Boosting the Transformer with the BERT Supervision in Low-Resource Machine Translation
Applied Sciences, 2022 Previous works trained the Transformer and its variants end-to-end and achieved remarkable translation performance when there are huge parallel sentences available.
Rong Yan +4 more
doaj +1 more source
Improving Hybrid CTC/Attention Architecture for Agglutinative Language Speech Recognition
Sensors, 2022 Unlike the traditional model, the end-to-end (E2E) ASR model does not require speech information such as a pronunciation dictionary, and its system is built through a single neural network and obtains performance comparable to that of traditional methods.
Zeyu Ren +4 more
doaj +1 more source
Experimental study on characteristics of grounawater fracture in coalmine overlying rock
Meitan kexue jishu, 2023 The seepage characteristics of groundwater in the overlying fissures are the key factors to be considered in the later maintenance of coal mining and goaf, especially in the underground reservoir construction area of coal mines.
Haixiang LI +4 more
doaj +1 more source
Autism services in low-resource areas
Neurosciences, 2023 To explore access to intervention services for children with autism spectrum disorder (ASD) in Jordan.We used prospective cross sectional design and survey methodology to collect information from the parents of a convenient sample of children with ASD aged 2.5-17 years and who attended pediatric neurology clinics in 3 different university affiliated ...
Masri, Amira T. +9 more
openaire +2 more sources
Environment International, 2020 Activated sludge processes with an ultra-short sludge retention time (ultra-short-SRT) are considered to have potential for energy and resource recovery from wastewater.
Yuting Shao +11 more
doaj +1 more source
Life cycle energy and carbon emissions of colorants extraction from Hibiscus sabdariffa
Energy Reports, 2022 The food industry is interested in new plant-based, non-toxic, food additives to substitute synthetic ones. Aiming to support the development of H. sabdariffa bio-colorant extracts at an early stage, a cradle-to-gate life cycle assessment (LCA) approach ...
Helena Monteiro, Bruna Moura
doaj +1 more source
Data Augmentation for Low-Resource Neural Machine Translation [PDF]
, 2017 The quality of a Neural Machine Translation system depends substantially on the availability of sizable parallel corpora. For low-resource language pairs this is not the case, resulting in poor translation quality. Inspired by work in computer vision, we
Bisazza, Arianna +2 more
core +2 more sources
Framework to assess eco-efficiency improvement: Case study of a meat production industry
Energy Reports, 2021 The industry sector accounts for nearly a quarter of the total global final energy and heat makes up two-thirds of that parcel. Sectors such as food & drinks, steel, cement, ceramic and glass, among others represent a considerable part of the energy ...
Muriel Iten +2 more
doaj +1 more source
Low-resource synchronous coincidence processor for positron emission tomography [PDF]
, 2011 We developed a new FPGA-based method for coincidence detection in positronemissiontomography. The method requires low device resources and no specific peripherals in order to resolve coincident digital pulses within a time window of a few nanoseconds ...
Andrés Santos +14 more
core +2 more sources