The project aims at automatic information retrieval of the interlingual medical knowledge. We adopted cross-lingual information retrieval and large health corpora, supported by neural methods, and complemented with low-resourced languages. The project aims to enrich and obtain an effective automatic search engine which utilizes health-related information such as global research on COVID-19. The core of our approach is to use state-of-the-art deep neural networks trained and fine-tuned on medical domain data.

In: The Covid-19 MultiLingual Information Access (MLIA) Eval 2020, ELRA-ELRC, European Commission

(url, bibtex)

LSCP is hierarchically organized in a semantic taxonomy that focuses on multi-task informal Persian language understanding as a comprehensive problem. This encompasses the recognition of multiple semantic aspects in the human-level sentences, which naturally captures from the real-world sentences. The proposed corpus consists of 120M Persian (Farsi) sentences resulted from 27M tweets annotated with parsing tree, part-of-speech tags, sentiment polarity and translations in English, German, Czech, Italian and Hindi spoken languages.

In: Proceedings of the 12th International Conference on Language Resources and Evaluation (LREC 2020), pp. 6325-6329, European Language Resources Association, Marseille, France, ISBN 979-10-95546-34-4

(url, bibtex, web)

We introduced an end-to-end deep multimodal convolutional-recurrent network for learning both vision and language representations simultaneously to infer image-text similarity. The model learns which pairs are a match (positive) and which ones are a mismatch (negative) using a hinge-based triplet ranking. To learn about the joint representations, we leverage our newly extracted collection of tweets from Twitter. The main characteristic of our dataset is that the images and tweets are not standardized the same as the benchmarks. Furthermore, there can be a higher semantic correlation between the pictures and tweets contrary to benchmarks in which the descriptions are well-organized.

Based on Automatic Image Description Generation Using Deep Multimodal Embeddings: RG.2.2.21497.01129, In: arXiv e-prints:2002.10016 [cs.IR]

(url, bibtex)

The main goal of this study is to prevent accidents caused by fatigue, drowsiness, and driver distraction. To avoid these incidents, this study proposes an integrated safety system that continuously monitors the driver's attention and vehicle surroundings, and finally decides whether the actual steering control status is safe or not. For this purpose, we equipped an ordinary car called FARAZ with a vision system consisting of four mounted cameras along with a universal car tool for communicating with surrounding factory-installed sensors and other car systems, and sending commands to actuators. The proposed system leverages a scene understanding pipeline using a deep convolutional encoder-decoder network and a driver state detection pipeline. We have been identifying and assessing domestic capabilities for the development of technologies specifically of the ordinary vehicles in order to manufacture smart cars and eke providing an intelligent system to increase safety and to assist the driver in various conditions/situations.

In: Data Science: From Research to Application. Lecture Notes on Data Engineering and Communications Technologies (LNDECT), vol 45., pp. 322-336, Springer, Cham., Nature Switzerland AG, ISBN 978-3-030-37308-5

(url, bibtex, web)