We propose a deep generative model that performs typography analysis and font reconstruction by learning disentangled manifolds of both font style and character shape. Our approach enables us to massively scale up the number of character types we can
effectively model compared to previous methods. Specifically, we infer separate latent variables representing character and font via a pair of inference networks which take as input sets of glyphs that either all share a character type, or belong to the same font. This design allows our model to generalize to characters that were not observed during training time, an important task in light of the relative sparsity of most fonts. We also put forward a new loss, adapted from prior work that measures likelihood using an adaptive distribution in a projected space, resulting in more natural images without requiring a discriminator. We evaluate on the task of font reconstruction over various datasets representing character types of many languages, and compare favorably to modern style transfer systems according to both automatic and manually-evaluated metrics.
Code summarization aims to generate concise natural language descriptions of source code, which can help improve program comprehension and maintenance. Recent studies show that syntactic and structural information extracted from abstract syntax trees
(ASTs) is conducive to summary generation. However, existing approaches fail to fully capture the rich information in ASTs because of the large size/depth of ASTs. In this paper, we propose a novel model CAST that hierarchically splits and reconstructs ASTs. First, we hierarchically split a large AST into a set of subtrees and utilize a recursive neural network to encode the subtrees. Then, we aggregate the embeddings of subtrees by reconstructing the split ASTs to get the representation of the complete AST. Finally, AST representation, together with source code embedding obtained by a vanilla code token encoder, is used for code summarization. Extensive experiments, including the ablation study and the human evaluation, on benchmarks have demonstrated the power of CAST. To facilitate reproducibility, our code and data are available at https://github.com/DeepSoftwareAnalytics/CAST.
Neural sequence-to-sequence (Seq2Seq) models and BERT have achieved substantial improvements in abstractive document summarization (ADS) without and with pre-training, respectively. However, they sometimes repeatedly attend to unimportant source phra
ses while mistakenly ignore important ones. We present reconstruction mechanisms on two levels to alleviate this issue. The sequence-level reconstructor reconstructs the whole document from the hidden layer of the target summary, while the word embedding-level one rebuilds the average of word embeddings of the source at the target side to guarantee that as much critical information is included in the summary as possible. Based on the assumption that inverse document frequency (IDF) measures how important a word is, we further leverage the IDF weights in our embedding-level reconstructor. The proposed frameworks lead to promising improvements for ROUGE metrics and human rating on both the CNN/Daily Mail and Newsroom summarization datasets.
Historical linguists have identified regularities in the process of historic sound change. The comparative method utilizes those regularities to reconstruct proto-words based on observed forms in daughter languages. Can this process be efficiently au
tomated? We address the task of proto-word reconstruction, in which the model is exposed to cognates in contemporary daughter languages, and has to predict the proto word in the ancestor language. We provide a novel dataset for this task, encompassing over 8,000 comparative entries, and show that neural sequence models outperform conventional methods applied to this task so far. Error analysis reveals a variability in the ability of neural model to capture different phonological changes, correlating with the complexity of the changes. Analysis of learned embeddings reveals the models learn phonologically meaningful generalizations, corresponding to well-attested phonological shifts documented by historical linguistics.
This study aimed at identifying Mechanisms of Human Capital Development in Tishreen University, and its role in reconstruction of Syria. In addition, the extent Tishreen University applied these mechanisms and the strategic objectives identified in t
he Higher Education Plan. The researcher gave a questionnaire to 335 people, they are members of the Teaching Staff in Tishreen University, the main result of the study was that the universities follow to these mechanisms will contribute in Reconstruction. We have provided a summary of the main result that have found out is that Tishreen University did not apply mechanisms of Human Capital Development and the strategic objectives as required, and there is no interest in scientific research, absence of research plans and weak budget, and a dire need to contract with scientific data and research engines.
The transport sector provides suitable and appropriate conditions for the development of
the various sectors of the national economy, productivity and other services. Its
relationship with these sectors is characterized by a mutual influence relati
onship, which
means that no progress or development can be achieved in any of these sectors without
preceded or accompanied by significant progress in the transport sector.
The government has given special importance to this division since the inception of
independence, when the sector was allocated large provision in its successive five-year
plans. After the year 2000, there was a big leap in the provision of the allocations assigned
to the transport sector, where reached (to the ninth five-year plan (2001-2005)) the
equivalent to the expenditure on this sector in the previous five-year plans combined.
The systematic destruction and devastation of the crisis in Syria since March 2011 aimed at
splitting the country and paralyzing the traffic of passengers and goods. As a result, the
transport sector suffered heavy economic losses in its infrastructure and transportation.
These losses were evaluated in this research for the years 2011-2012, 2013 and the state of
the transport sector was reviewed before the crisis.
In view of the importance of this sector and its key role in the reconstruction, there was a
package of proposals and visions that were put forward for the reconstruction of the
infrastructure of the sector itself as a priority and strengthening the role of this sector in the
reconstruction and development within the framework of comprehensive regional planning
and sustainable development.
Perhaps the most important impact of the crisis was the decision of the Syrian state oriental
direction in its policy, which directly affects the implementation of transport networks and
infrastructure and means of transport.
Based on the goal of linking the five seas, which have set a higher goal in the national
framework for regional planning, some visions have been reviewed that can achieve this
goal in order to attract international transport and transit through Syria and exploit its
geographical position as a transport hub leading to an increase in economic growth rate To
raise the contribution of the transport sector to GDP.
Background: The defects in the leg is one of the most important challenges facing
reconstructive surgeons because of the lack of soft tissue covering the front surface of the
leg and the lack of blood perfusion in the lower third of the leg. The mu
scle and
fasciocutaneous flaps are one of the most important reconstruction options as there have
been no local studies about this method so far.
Objective: Evaluation of methods of reconstruction the defects of the soft tissues of
the leg using the muscle flaps and fasciocutaneous flaps and comparing them with the
international studies.
Methods: The study included 25 patients undergoing transposition of the
gastrocnemius and / or soleus flaps, sural flaps for treating soft tissues in the leg, from
2015 to 2017 in both Al-Asaad and Tishreen University hospitals in Lattakia , gathering
information on the etiology the lesion, the time of reconstruction, the flap used to cover the
lesion, and the function of the gastrocnemius-soleus unit, and the following Sensory
disorders.
The management of lower extremity trauma has evolved over the last two decades to
the point that many extremities that would have required amputation are routinely
salvaged. The use of transposition muscle is regarded as the optimal method for the
treatment of extensive defects of soft tissues in limb traumatology. The aim of the study is
to investigate the therapeutic effect of repairing soft tissue defects of mid-tibia open
fractures with muscle flap pedicled with medial half of soleus. A retrospective study of our
outcome of 10 performed soft-tissue reconstruction of an open tibial wound cases. All
these patients were operated using the proposed surgical technique between 2013 and 2015
in AL-ASSAD University Hospital in Lattakia, Syria. The results of this technique of
reconstruction are excellent. Medial hemisoleus flap is a valuable option for soft tissue
coverage of middle third of lower leg. It does not sacrifice the whole function of the Soleus
muscle. Due to its longer arc of rotation, this flap can cover the defects of different size
and shape in middle third of leg. The operation is relatively easy. The muscle can fill the
narrow cavity and repair the soft tissue defect simultaneously. The donor-site injury is
minor.
This paper aims to identify key characteristics projects in
reconstruction stage in order to assist decision makers to produce
appropriate approach to manage those projects effectively. A list of
characteristics that may exist in reconstruction projects were
identified through intensive literature review and pilot study with
various stakeholders involved in in reconstruction stage.
The reconstruction of the nasal ala requires the achievement of esthetic and functional results. Composite auricular grafts are extremely valuable here. The purpose of this study is to evaluate the surgical outcome after using composite grafts for al
ar nasal reconstruction according to our experience. A retrospective study of our outcome of 10 performed alar nasal reconstruction cases. All these patients were operated using the proposed surgical technique between 2011 and 2015 in AL-ASSAD University Hospital in Lattakia, Syria. The cosmetic and functional results of this relatively simple technique of reconstruction are excellent. It provides an internal lining and support. The integration was total. We could fill the alar nasal defect using composite grafts. The defect has disappeared consequently. The form of the nose could be improved. The use of this technique efficiently enables a one-stage reconstruction of nasal defects. The morbidity rate within the donor areas was low. We believe that composite grafts have an important and under-utilized role as an adjunct to the flaps in nasal reconstruction.
