Adapting word order from one language to another is a key problem in cross-lingual structured prediction. Current sentence encoders (e.g., RNN, Transformer with position embeddings) are usually word order sensitive. Even with uniform word form repres
entations (MUSE, mBERT), word order discrepancies may hurt the adaptation of models. In this paper, we build structured prediction models with bag-of-words inputs, and introduce a new reordering module to organizing words following the source language order, which learns task-specific reordering strategies from a general-purpose order predictor model. Experiments on zero-shot cross-lingual dependency parsing, POS tagging, and morphological tagging show that our model can significantly improve target language performances, especially for languages that are distant from the source language.
Ad hoc abbreviations are commonly found in informal communication channels that favor shorter messages. We consider the task of reversing these abbreviations in context to recover normalized, expanded versions of abbreviated messages. The problem is
related to, but distinct from, spelling correction, as ad hoc abbreviations are intentional and can involve more substantial differences from the original words. Ad hoc abbreviations are also productively generated on-the-fly, so they cannot be resolved solely by dictionary lookup. We generate a large, open-source data set of ad hoc abbreviations. This data is used to study abbreviation strategies and to develop two strong baselines for abbreviation expansion.
Due to the development of deep learning, the natural language processing tasks have made great progresses by leveraging the bidirectional encoder representations from Transformers (BERT). The goal of information retrieval is to search the most releva
nt results for the user's query from a large set of documents. Although BERT-based retrieval models have shown excellent results in many studies, these models usually suffer from the need for large amounts of computations and/or additional storage spaces. In view of the flaws, a BERT-based Siamese-structured retrieval model (BESS) is proposed in this paper. BESS not only inherits the merits of pre-trained language models, but also can generate extra information to compensate the original query automatically. Besides, the reinforcement learning strategy is introduced to make the model more robust. Accordingly, we evaluate BESS on three public-available corpora, and the experimental results demonstrate the efficiency of the proposed retrieval model.
Vast amounts of data in healthcare are available in unstructured text format, usually in the local language of the countries. These documents contain valuable information. Secondary use of clinical narratives and information extraction of key facts a
nd relations from them about the patient disease history can foster preventive medicine and improve healthcare. In this paper, we propose a hybrid method for the automatic transformation of clinical text into a structured format. The documents are automatically sectioned into the following parts: diagnosis, patient history, patient status, lab results. For the Diagnosis'' section a deep learning text-based encoding into ICD-10 codes is applied using MBG-ClinicalBERT - a fine-tuned ClinicalBERT model for Bulgarian medical text. From the Patient History'' section, we identify patient symptoms using a rule-based approach enhanced with similarity search based on MBG-ClinicalBERT word embeddings. We also identify symptom relations like negation. For the Patient Status'' description, binary classification is used to determine the status of each anatomic organ. In this paper, we demonstrate different methods for adapting NLP tools for English and other languages to a low resource language like Bulgarian.
Complex question answering often requires finding a reasoning chain that consists of multiple evidence pieces. Current approaches incorporate the strengths of structured knowledge and unstructured text, assuming text corpora is semi-structured. Build
ing on dense retrieval methods, we propose a new multi-step retrieval approach (BeamDR) that iteratively forms an evidence chain through beam search in dense representations. When evaluated on multi-hop question answering, BeamDR is competitive to state-of-the-art systems, without using any semi-structured information. Through query composition in dense space, BeamDR captures the implicit relationships between evidence in the reasoning chain. The code is available at https://github.com/ henryzhao5852/BeamDR.
We present a set of assignments for a graduate-level NLP course. Assignments are designed to be interactive, easily gradable, and to give students hands-on experience with several key types of structure (sequences, tags, parse trees, and logical form
s), modern neural architectures (LSTMs and Transformers), inference algorithms (dynamic programs and approximate search) and training methods (full and weak supervision). We designed assignments to build incrementally both within each assignment and across assignments, with the goal of enabling students to undertake graduate-level research in NLP by the end of the course.
Abstract Although current CCG supertaggers achieve high accuracy on the standard WSJ test set, few systems make use of the categories' internal structure that will drive the syntactic derivation during parsing. The tagset is traditionally truncated,
discarding the many rare and complex category types in the long tail. However, supertags are themselves trees. Rather than give up on rare tags, we investigate constructive models that account for their internal structure, including novel methods for tree-structured prediction. Our best tagger is capable of recovering a sizeable fraction of the long-tail supertags and even generates CCG categories that have never been seen in training, while approximating the prior state of the art in overall tag accuracy with fewer parameters. We further investigate how well different approaches generalize to out-of-domain evaluation sets.
Due to the absence of operational requirements and its unidentified solutions ,this
study presented definition alternative analytical and technical conditions for the
continuation of nitrate-selective electrode working. Many electrolytes were teste
d
(ammonium sulfate, ammonium acetate and ammonium chloride). The technical and
analytical conditions shows that the best response which give the Nernst slope of the
electrode had been obtained by adding 0.5ml of ammonium sulfate (3mol / l), and the
response time was 20sec, by the time that values stabilized for 15sec. The study of pH
effect on the electrode response also showed that electrode works within the range (2-8) of
pH and that the best response was at pH = 5.5 by using acetic acid-sodium acetate buffer.
The effect of temperature on the potential electrode was studied ,it showed that the Nernest
electrode response realized within the range (18-37) C0of temperature.While Studying the
effect of some interfering ions such as (OH-, HCO3
-, CO3
2-,Cl-,NO2
-) on the work of the
electrode showed low values of the coefficient of selectivity, which did not exceed 3×10-2
that the electrode was not affected by the presence of these ions. The chosen conditions
were experimented on standard solutions, which showed that the electrode is working
within the range (10-4-10-1) mol / l of concentrations. When the concentrations was about
10-5mol/l or less than that, it was necessary to soak the electrode in solution which has low
concentration for at least one hour, and diluted the taken volumes ten times . These
technical and analytical conditions was applied on standard solutions and environmental
samples(water), which showed the accuracy ,correctly and authenticity of the
measurements.
The effect of different culture medias (water, MS solid, MS 1/2 solid) on germination
and growth of heliotropiun hirsutissimum Grauer. in vitro was studied with different
concentrations of gibberellic acid (0.001-0.01-0.1 and 1 mg/l).
The results
showed that best germination rate (80.5%) and root (20mm) and shoot
(35mm) growth were obtained in MS solid medium with 0.1mg/l GA3.
Addition of 0.1mg/l GA3 to MS1/2 solid medium improved germination rate
(75.44%), root (19mm), and shoot (24mm)growth.
When GA3 (0.1 mg/l) was added to water medium, germination rate reached
(65.33%), as well as root and shoot growth (12.66mm and20mm) alternatively after 2
weeks of planting.
Seedling of MS solid medium were transferred into pots contained torp medium to
adapting them with outside environment, and then surviving their growth until maturity
after 4 weeks.
يهدف بحثنا إلى دراسة إمكانية التحكم بعمل دارة الاشتعال الترانزستورية المستخدمة في سيارة حديثة بشكل آلي لإلغاء حادثة الطرق التي تحدث في المحرك عند التزود بوقود ذو عدد أوكتاني منخفض، و عند سرعات الدوران المنخفضة للجذع المعقوف و ذلك باقتراح استخدام منظم
أوكتاني الكتروني ثم تصميم المنظم الالكتروني المقترح للاستخدام، حيث تم دراسة إمكانية استخدام المتحكمات الصغرية من أجل تعيير زاوية تسبيق الاشتعال بشكل آلي و منع حادثة الطرق في اسطوانات المحرك لأن هذه الظاهرة تعمل على إنقاص عمر المحرك.
