Search

In this paper we share findings from our effort to build practical machine translation (MT) systems capable of translating across over one thousand languages. We describe results in three research domains: (i) Building clean, web-mined datasets for 1500+ languages by leveraging semi-supervised pre-training for language identification and developing data-driven filtering techniques; (ii) Developing practical MT models for under-served languages by leveraging massively multilingual models trained with supervised parallel data for over 100 high-resource languages and monolingual datasets for an additional 1000+ languages; and (iii) Studying the limitations of evaluation metrics for these languages and conducting qualitative analysis of the outputs from our MT models, highlighting several frequent error modes of these types of models. We hope that our work provides useful insights to practitioners working towards building MT systems for currently understudied languages, and highlights research directions that can complement the weaknesses of massively multilingual models in data-sparse settings., Comment: V2: updated with some details from 24-language Google Translate launch in May 2022 V3: spelling corrections, additional acknowledgements

The COVID-19 pandemic is the worst pandemic to strike the world in over a century. Crucial to stemming the tide of the SARS-CoV-2 virus is communicating to vulnerable populations the means by which they can protect themselves. To this end, the collaborators forming the Translation Initiative for COvid-19 (TICO-19) have made test and development data available to AI and MT researchers in 35 different languages in order to foster the development of tools and resources for improving access to information about COVID-19 in these languages. In addition to 9 high-resourced, "pivot" languages, the team is targeting 26 lesser resourced languages, in particular languages of Africa, South Asia and South-East Asia, whose populations may be the most vulnerable to the spread of the virus. The same data is translated into all of the languages represented, meaning that testing or development can be done for any pairing of languages in the set. Further, the team is converting the test and development data into translation memories (TMXs) that can be used by localizers from and to any of the languages.

The paper investigates the feasibility of confidence estimation for neural machine translation models operating at the high end of the performance spectrum. As a side product of the data annotation process necessary for building such models we propose sentence level accuracy $SACC$ as a simple, self-explanatory evaluation metric for quality of translation. Experiments on two different annotator pools, one comprised of non-expert (crowd-sourced) and one of expert (professional) translators show that $SACC$ can vary greatly depending on the translation proficiency of the annotators, despite the fact that both pools are about equally reliable according to Krippendorff's alpha metric; the relatively low values of inter-annotator agreement confirm the expectation that sentence-level binary labeling $good$ / $needs\ work$ for translation out of context is very hard. For an English-Spanish translation model operating at $SACC = 0.89$ according to a non-expert annotator pool we can derive a confidence estimate that labels 0.5-0.6 of the $good$ translations in an "in-domain" test set with 0.95 Precision. Switching to an expert annotator pool decreases $SACC$ dramatically: $0.61$ for English-Spanish, measured on the exact same data as above. This forces us to lower the CE model operating point to 0.9 Precision while labeling correctly about 0.20-0.25 of the $good$ translations in the data. We find surprising the extent to which CE depends on the level of proficiency of the annotator pool used for labeling the data. This leads to an important recommendation we wish to make when tackling CE modeling in practice: it is critical to match the end-user expectation for translation quality in the desired domain with the demands of annotators assigning binary quality labels to CE training data.

Ying Liu,1,* Man He,1,* Mengmeng Niu,1 Yiqing Zhao,1 Yuanzhang Zhu,1 Zhenhua Li,2 Nianping Feng1 1Department of Pharmaceutical Sciences, School of Pharmacy, Shanghai University of Traditional Chinese Medicine, Shanghai, People’s Republic of China; 2Cedars-Sinai Medical Center, Los Angeles, CA, USA *These authors contributed equally to this work Abstract: Rapid drug release at the specific site of action is still a challenge for antitumor therapy. Development of stimuli-responsive hybrid nanocarriers provides a promising strategy to enhance therapeutic effects by combining the unique features of each component. The present study explored the use of drug–gold nanoparticle conjugates incorporated into liposomes to enhance antitumor efficiency. A model drug, vincristine sulfate, was physically conjugated with gold nanoparticles and verified by UV-visible and fourier transform infrared spectroscopy, and differential scanning calorimetry. The conjugates were incorporated into liposomes by film dispersion to yield nanoparticles (113.4 nm) with light-responsive release properties, as shown by in vitro release studies. Intracellular uptake and distribution was studied in HeLa cells using transmission electron microscopy and confocal laser scanning microscopy. This demonstrated liposome internalization and localization in endosomal–lysosomal vesicles. Fluorescence intensity increased in cells exposed to UV light, indicating that this stimulated intracellular drug release; this finding was confirmed by quantitative analyses using flow cytometry. Antitumor efficacy was evaluated in HeLa cells, both in culture and in implants in vivo in nude mice. HeLa cell viability assays showed that light exposure enhanced liposome cytotoxicity and induction of apoptosis. Furthermore, treatment with the prepared liposomes coupled with UV light exposure produced greater antitumor effects in nude mice and reduced side effects, as compared with free vincris

Ying Liu,1,* Man He,1,* Mengmeng Niu,1 Yiqing Zhao,1 Yuanzhang Zhu,1 Zhenhua Li,2 Nianping Feng1 1Department of Pharmaceutical Sciences, School of Pharmacy, Shanghai University of Traditional Chinese Medicine, Shanghai, People’s Republic of China; 2Cedars-Sinai Medical Center, Los Angeles, CA, USA *These authors contributed equally to this work Abstract: Rapid drug release at the specific site of action is still a challenge for antitumor therapy. Development of stimuli-responsive hybrid nanocarriers provides a promising strategy to enhance therapeutic effects by combining the unique features of each component. The present study explored the use of drug–gold nanoparticle conjugates incorporated into liposomes to enhance antitumor efficiency. A model drug, vincristine sulfate, was physically conjugated with gold nanoparticles and verified by UV-visible and fourier transform infrared spectroscopy, and differential scanning calorimetry. The conjugates were incorporated into liposomes by film dispersion to yield nanoparticles (113.4 nm) with light-responsive release properties, as shown by in vitro release studies. Intracellular uptake and distribution was studied in HeLa cells using transmission electron microscopy and confocal laser scanning microscopy. This demonstrated liposome internalization and localization in endosomal–lysosomal vesicles. Fluorescence intensity increased in cells exposed to UV light, indicating that this stimulated intracellular drug release; this finding was confirmed by quantitative analyses using flow cytometry. Antitumor efficacy was evaluated in HeLa cells, both in culture and in implants in vivo in nude mice. HeLa cell viability assays showed that light exposure enhanced liposome cytotoxicity and induction of apoptosis. Furthermore, treatment with the prepared liposomes coupled with UV light exposure produced greater antitumor effects in nude mice and reduced side effects, as compared with free vincris

Mengmeng Niu1, Yi Lu1, Lars Hovgaard2, Wei Wu11School of Pharmacy, Fudan University, Shanghai, People's Republic of China; 2Oral Formulation Development, Novo Nordisk A/S, Maalov, DenmarkBackground: Oral delivery of insulin is challenging and must overcome the barriers of gastric and enzymatic degradation as well as low permeation across the intestinal epithelium. The present study aimed to develop a liposomal delivery system containing glycocholate as an enzyme inhibitor and permeation enhancer for oral insulin delivery.Methods: Liposomes containing sodium glycocholate were prepared by a reversed-phase evaporation method followed by homogenization. The particle size and entrapment efficiency of recombinant human insulin (rhINS)-loaded sodium glycocholate liposomes can be easily adjusted by tuning the homogenization parameters, phospholipid:sodium glycocholate ratio, insulin:phospholipid ratio, water:ether volume ratio, interior water phase pH, and the hydration buffer pH.Results: The optimal formulation showed an insulin entrapment efficiency of 30% ± 2% and a particle size of 154 ± 18 nm. A conformational study by circular dichroism spectroscopy and a bioactivity study confirmed the preserved integrity of rhINS against preparative stress. Transmission electron micrographs revealed a nearly spherical and deformed structure with discernable lamella for sodium glycocholate liposomes. Sodium glycocholate liposomes showed better protection of insulin against enzymatic degradation by pepsin, trypsin, and a-chymotrypsin than liposomes containing the bile salt counterparts of sodium taurocholate and sodium deoxycholate.Conclusion: Sodium glycocholate liposomes showed promising in vitro characteristics and have the potential to be able to deliver insulin orally.Keywords: liposomes, glycocholate, insulin, enzymatic degradation, oral

Peipei Guan1, Yi Lu1, Jianping Qi1, Mengmeng Niu1, Ruyue Lian1, Fuqiang Hu2, Wei Wu11School of Pharmacy, Fudan University, Shanghai, People's Republic of China; 2School of Pharmacy, Zhejiang University, Hangzhou, People's Republic of ChinaAbstract: The main purpose of this study was to evaluate liposomes containing a bile salt, sodium deoxycholate (SDC), as oral drug delivery systems to enhance the oral bioavailability of the poorly water-soluble and poorly permeable drug, cyclosporine A (CyA). Liposomes composed of soybean phosphatidylcholine (SPC) and SDC were prepared by a thin-film dispersion method followed by homogenization. Several properties of the liposomes including particle size, polydispersity index, and entrapment efficiency were characterized. The in vitro release of CyA from these liposomes was less than 5% at 12 hours as measured by a dynamic dialysis method. The pharmacokinetic results in rats showed improved absorption of CyA in SPC/SDC liposomes, compared with CyA-loaded conventional SPC/cholesterol (Chol) liposomes and microemulsion-based Sandimmune Neoral®. The relative oral bioavailability of CyA-loaded SPC/SDC and SPC/Chol liposomes was 120.3% and 98.6%, respectively, with Sandimmun Neoral as the reference. The enhanced bioavailability of CyA was probably due to facilitated absorption by the liposomes containing SDC rather than improved release rate.Keywords: liposomes, bile salt, sodium deoxycholate, cyclosporine A, oral bioavailability

Peipei Guan1, Yi Lu1, Jianping Qi1, Mengmeng Niu1, Ruyue Lian1, Fuqiang Hu2, Wei Wu11School of Pharmacy, Fudan University, Shanghai, People's Republic of China; 2School of Pharmacy, Zhejiang University, Hangzhou, People's Republic of ChinaAbstract: The main purpose of this study was to evaluate liposomes containing a bile salt, sodium deoxycholate (SDC), as oral drug delivery systems to enhance the oral bioavailability of the poorly water-soluble and poorly permeable drug, cyclosporine A (CyA). Liposomes composed of soybean phosphatidylcholine (SPC) and SDC were prepared by a thin-film dispersion method followed by homogenization. Several properties of the liposomes including particle size, polydispersity index, and entrapment efficiency were characterized. The in vitro release of CyA from these liposomes was less than 5% at 12 hours as measured by a dynamic dialysis method. The pharmacokinetic results in rats showed improved absorption of CyA in SPC/SDC liposomes, compared with CyA-loaded conventional SPC/cholesterol (Chol) liposomes and microemulsion-based Sandimmune Neoral®. The relative oral bioavailability of CyA-loaded SPC/SDC and SPC/Chol liposomes was 120.3% and 98.6%, respectively, with Sandimmun Neoral as the reference. The enhanced bioavailability of CyA was probably due to facilitated absorption by the liposomes containing SDC rather than improved release rate.Keywords: liposomes, bile salt, sodium deoxycholate, cyclosporine A, oral bioavailability

Mengmeng Niu1, Yi Lu1, Lars Hovgaard2, Wei Wu11School of Pharmacy, Fudan University, Shanghai, People's Republic of China; 2Oral Formulation Development, Novo Nordisk A/S, Maalov, DenmarkBackground: Oral delivery of insulin is challenging and must overcome the barriers of gastric and enzymatic degradation as well as low permeation across the intestinal epithelium. The present study aimed to develop a liposomal delivery system containing glycocholate as an enzyme inhibitor and permeation enhancer for oral insulin delivery.Methods: Liposomes containing sodium glycocholate were prepared by a reversed-phase evaporation method followed by homogenization. The particle size and entrapment efficiency of recombinant human insulin (rhINS)-loaded sodium glycocholate liposomes can be easily adjusted by tuning the homogenization parameters, phospholipid:sodium glycocholate ratio, insulin:phospholipid ratio, water:ether volume ratio, interior water phase pH, and the hydration buffer pH.Results: The optimal formulation showed an insulin entrapment efficiency of 30% ± 2% and a particle size of 154 ± 18 nm. A conformational study by circular dichroism spectroscopy and a bioactivity study confirmed the preserved integrity of rhINS against preparative stress. Transmission electron micrographs revealed a nearly spherical and deformed structure with discernable lamella for sodium glycocholate liposomes. Sodium glycocholate liposomes showed better protection of insulin against enzymatic degradation by pepsin, trypsin, and a-chymotrypsin than liposomes containing the bile salt counterparts of sodium taurocholate and sodium deoxycholate.Conclusion: Sodium glycocholate liposomes showed promising in vitro characteristics and have the potential to be able to deliver insulin orally.Keywords: liposomes, glycocholate, insulin, enzymatic degradation, oral