A5033564621- Natural Language Processing Techniques
- Topic Modeling
- Multimodal Machine Learning Applications
- Artificial Intelligence in Healthcare
- Problem and Project Based Learning
- Machine Learning in Healthcare
- Interactive and Immersive Displays
- Mechatronics Education and Applications
- Online Learning and Analytics
- Experimental Learning in Engineering
- Hand Gesture Recognition Systems
- COVID-19 diagnosis using AI
- Mathematics, Computing, and Information Processing
- Augmented Reality Applications
- Advanced Text Analysis Techniques
Superior University
2022-2023
Google (United States)
2019-2020
University of Lahore
2020
Human Computer Interaction (Switzerland)
2015
Helwan University
2015
Yinfei Yang, Daniel Cer, Amin Ahmad, Mandy Guo, Jax Law, Noah Constant, Gustavo Hernandez Abrego, Steve Yuan, Chris Tar, Yun-hsuan Sung, Brian Strope, Ray Kurzweil. Proceedings of the 58th Annual Meeting Association for Computational Linguistics: System Demonstrations. 2020.
We introduce two pre-trained retrieval focused multilingual sentence encoding models, respectively based on the Transformer and CNN model architectures. The models embed text from 16 languages into a single semantic space using multi-task trained dual-encoder that learns tied representations translation bridge tasks (Chidambaram al., 2018). provide performance is competitive with state-of-the-art on: (SR), pair bitext (BR) question answering (ReQA). On English transfer learning tasks, our...
Popular QA benchmarks like SQuAD have driven progress on the task of identifying answer spans within a specific passage, with models now surpassing human performance. However, retrieving relevant answers from huge corpus documents is still challenging problem, and places different requirements model architecture. There growing interest in developing scalable retrieval trained end-to-end, bypassing typical document step. In this paper, we introduce Retrieval Question-Answering (ReQA),...
In Natural Language Processing (NLP), topic modeling is the technique to extract abstract information from documents with huge amount of text. This leads towards identification topics in document. One way retrieve keyphrase extraction. Keyphrases are a set terms which represent high level description Different techniques extraction for prediction have been proposed multiple languages i.e. English, Arabic, etc. However, this area needs be explored other e.g. Urdu. Therefore, paper, novel...
RemoAct is a wearable depth sensing and projection system that enables interaction on many surfaces. It makes with the environment more intuitive through sharing sending data surrounding devices by applying certain gestures. This offers mobile solution for interacting using projected surface habitual flat Every user has their public private areas, where can create tiles fly share it others these are shown to other users augmented reality. Interaction made hand gestures, finger tracking...
"The rising trend of students dropping out universities without completing their degrees is becoming a concerning issue for institutions. To address this problem, the reasons behind phenomenon need to be explored. However, most educational data sets have small sample sizes and varying patterns. Currently, there are few machine learning approaches Pakistani higher education student performance. This study presents learning-based approach predict withdrawals identify them. The proposed...
Summarization is one of the most common tasks performed by large language models (LLMs), especially in applications like Retrieval-Augmented Generation (RAG). However, existing evaluations hallucinations LLM-generated summaries, and hallucination detection both suffer from a lack diversity recency LLM families considered. This paper introduces FaithBench, summarization benchmark comprising challenging made 10 modern LLMs 8 different families, with ground truth annotations human experts....
Multilingual information retrieval (MLIR) is a crucial yet challenging task due to the need for human annotations in multiple languages, making training data creation labor-intensive. In this paper, we introduce mAggretriever, which effectively leverages semantic and lexical features from pre-trained multilingual transformers (e.g., mBERT XLM-R) dense retrieval. To enhance inference efficiency, employ approximate masked-language modeling prediction computing features, reducing 70–85% GPU...