CIDEr-R: Robust Consensus-based Image Description Evaluation

This paper shows that CIDEr-D, a traditional evaluation metric for image description, does not work properly on datasets where the number of words in the sentence is significantly greater than those in the MS COCO Captions dataset. We also show that CIDEr-D has performance hampered by the lack of multiple reference sentences and high variance of sentence length. To bypass this problem, we introduce CIDEr-R, which improves CIDEr-D, making it more flexible in dealing with datasets with high sentence length variance. We demonstrate that CIDEr-R is more accurate and closer to human judgment than CIDEr-D; CIDEr-R is more robust regarding the number of available references. Our results reveal that using Self-Critical Sequence Training to optimize CIDEr-R generates descriptive captions. In contrast, when CIDEr-D is optimized, the generated captions' length tends to be similar to the reference length. However, the models also repeat several times the same word to increase the sentence length.

Wine is not v i n. On the Compatibility of Tokenizations across Languages

The size of the vocabulary is a central design choice in large pretrained language models, with respect to both performance and memory requirements. Typically, subword tokenization algorithms such as byte pair encoding and WordPiece are used. In this work, we investigate the compatibility of tokenizations for multilingual static and contextualized embedding spaces and propose a measure that reflects the compatibility of tokenizations across languages. Our goal is to prevent incompatible tokenizations, e.g., wine'' (word-level) in English vs. v i n'' (character-level) in French, which make it hard to learn good multilingual semantic representations. We show that our compatibility measure allows the system designer to create vocabularies across languages that are compatible -- a desideratum that so far has been neglected in multilingual models.

Constructing Emotional Consensus and Utilizing Unpaired Data for Empathetic Dialogue Generation

Researches on dialogue empathy aim to endow an agent with the capacity of accurate understanding and proper responding for emotions. Existing models for empathetic dialogue generation focus on the emotion flow in one direction, that is, from the context to response. We argue that conducting an empathetic conversation is a bidirectional process, where empathy occurs when the emotions of two interlocutors could converge on the same point, i.e., reaching an emotional consensus. Besides, we also find that the empathetic dialogue corpus is extremely limited, which further restricts the model performance. To address the above issues, we propose a dual-generative model, Dual-Emp, to simultaneously construct the emotional consensus and utilize some external unpaired data. Specifically, our model integrates a forward dialogue model, a backward dialogue model, and a discrete latent variable representing the emotional consensus into a unified architecture. Then, to alleviate the constraint of paired data, we extract unpaired emotional data from open-domain conversations and employ Dual-Emp to produce pseudo paired empathetic samples, which is more efficient and low-cost than the human annotation. Automatic and human evaluations demonstrate that our method outperforms competitive baselines in producing coherent and empathetic responses.

Study of the Upper Ordovician - Current Time tectonic evolution of the Elward Area in the Syrian Euphrates Graben Using restoration techniques

The inclined shear restoration technique was used in this research as the primary method to remove the effects of fault displacements. These displacements were resulted from the impact of the NE-SW trending extensional forces. The inclined shear restoration technique was applied to the NE-SW trending seismic section (Inline A2157) along the Elward Area, using 2D move software. The vertical shear restoration technique was used as a complementary method to remove the effects of folding associated with faulting, especially to formations under the Base Upper Cretaceous Unconformity (BKU). The inclined shear and the vertical shear restoration techniques contrib uted to build many geological sections according to depth seismic section (Inline A 2157). These sections showed the Tectonic setting of the study area from Upper Ordovician till current time.

Study of the success grafting of pear varieties Cocia and Williams and Quince variety Saidawi on the Syrian rootstock Pyrus syriaca Boiss. And development of produced seedlings

This research was conducted in seedling area belonging to Tishreen University and in the Coast seedling belonging to Agricultural Directorate in Lattakia. Results has shown that, the pear variety Cocia grafted on pear Syrian rootstock was significantly better than pear variety Williams and Quince variety Saidawi grafted on the same rootstock, where the percentage of grafting were 100%, 90% and 66.67% consequently, and in the next year 93.33%, 76.67% and 54.00% consequently. Results has shown a good agreement between a variety Cocia and a used rootstock in comparison with the tow other varieties. The results of analytical calculation has shown that, the variety Cocia was better than the tow other varieties Williams and Saidawi in graft-height, also, Williams variety was better than Saidawi variety, and this confirm that, the Cocia variety in medium growth when it grafted on pear Syrian rootstock, whereas, Williams variety was in poor growth and Saidawi variety was very poor in growth when they were grafted on the same rootstalk especially in the first years of grafting.