Seyed Arshan Dalili
Undergraduate Student of Computer Engineering
- arshandalili@gmail.com
- +98-912-070-9646
- arshandalili.github.io
- Sharif University of Technology, Tehran, Iran
My name is Seyed Arshan Dalili. I am a undergraduate student of Computer Engineering at Sharif University of Technology, and I will graduate with a bachelor’s degree next year. My fields of interest are Natural Language Processing (NLP), Information Retrieval, Artificial Intelligence, and Machine Learning. I am working on Visual Word Sense Disambiguation challenge in SemEval 2023 as my B.Sc. project.
I graduated and got my B.Sc. in Computer Engineering from Sharif University of Technology.
Curriculum Vitae
Education
- Bachelor in Computer EngineeringSharif University of Technology, Tehran, IranSep. 2019 - Jul. 2023
- Diploma of MathematicsShahid Beheshti School, National Organization for Development of Exceptional Talents (NODET), Sari, IranSep. 2016 - Jun. 2019
Awards
- Full ScholarshipReceived full scholarship (tuition waiver) from Sharif University of Technology for Bachelor’s degree.2019
- University Entrance ExamRanked 6th nationwide among more than 164,000 participant in Iran National University Entrance Exam (Konkour) in Mathematics Branch2019
Publications
Team SUT at SemEval-2023 Task 1: Prompt Generation for Visual Word Sense Disambiguation (B.Sc. Project)
A novel model for Visual Word Sense Disambiguation challenge.
Interests
Natural Language Processing
Information Retrieval
Social Networks
Artificial Intelligence and Machine Learning
Research Experience
- Visual Word Sense Disambiguation (B.Sc. Project, Ongoing)Working on a model to select the appropriate image for ambiguous words in a given context. Developing models with Transformers and BERT to 1) transform pictures and texts into a same space and 2) bring related pictures and texts that are semantically similar around each other. Also working on using pre-trained models like CLIP to select relevant images from ambiguous textual contexts.2022
- Computational SemanticsExploring various methods for understanding the meaning of Natural Language and analyzing methods for enabling multimodal models to comprehend Natural Language. Particularly, work is being done to create a dataset to enable models to learn the semantics of both textual and visual contexts. WordNet and Text-to-Image models will be used to create a dataset of images that correspond to a given meaning of a word.2022
- Implemented Credible Early Detection Point model in PytorchImplemented and published the CED model in Pytorch in order to make it easier to use compared to paper’s original old version of Tensorflow. Also, reviewed the mechanism of some fake news detection models such as Early Detection of Fake News on Social Media Through Propagation Path Classification with Recurrent and Convolutional Networks.2022