Ashfia Binte Habib
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Ashfia Binte Habib has a total of 21 co-authors. Here are the top co-authors by number of shared publications:
- Rajesh Palit
North South University, Dhaka, Bangladesh
3 shared publications Fahimul Haque
2 shared publications- Ibraheem Muhammad Moosa
Dept. of Electrical and Computer Engineering Pennsylvania State University North South University Dhaka Bangladesh
2 shared publications Mahmud Elahi Akhter
2 shared publications- Mahmud Elahi Akhter
Dept. of Electrical and Computer Engineering Pennsylvania State University North South University Dhaka Bangladesh
2 shared publications Rishad Arfin
2 shared publications
Latest Publications
- 20 December 2023
- Published by Springer Nature
- p. 377-387
COVID-19 Diagnosis and Classification from CXR Images Using Vision Transformer
- 20 December 2023
- Published by Springer Nature
- p. 377-387
COVID-19 Diagnosis and Classification from CXR Images Using Vision Transformer
- Open Access
Does Transliteration Help Multilingual Language Modeling?
Script diversity presents a challenge to Multilingual Language Models (MLLM) by reducing lexical overlap among closely related languages. Therefore, transliterating closely related languages that use different writing scripts to a common script may improve the downstream task performance of MLLMs. We empirically measure the effect of transliteration on MLLMs in this context. We specifically focus on the Indic languages, which have the highest script diversity in the world, and we evaluate our models on the IndicGLUE benchmark. We perform the Mann-Whitney U test to rigorously verify whether the effect of transliteration is significant or not. We find that transliteration benefits the low-resource languages without negatively affecting the comparatively high-resource languages. We also measure the cross-lingual representation similarity of the models using centered kernel alignment on parallel sentences from the FLORES-101 dataset. We find that for parallel sentences across different languages, the transliteration-based model learns sentence representations that are more similar.
Open AccessDoes Transliteration Help Multilingual Language Modeling?
Script diversity presents a challenge to Multilingual Language Models (MLLM) by reducing lexical overlap among closely related languages. Therefore, transliterating closely related languages that use different writing scripts to a common script may improve the downstream task performance of MLLMs. We empirically measure the effect of transliteration on MLLMs in this context. We specifically focus on the Indic languages, which have the highest script diversity in the world, and we evaluate our models on the IndicGLUE benchmark. We perform the Mann-Whitney U test to rigorously verify whether the effect of transliteration is significant or not. We find that transliteration benefits the low-resource languages without negatively affecting the comparatively high-resource languages. We also measure the cross-lingual representation similarity of the models using centered kernel alignment on parallel sentences from the FLORES-101 dataset. We find that for parallel sentences across different languages, the transliteration-based model learns sentence representations that are more similar.
- Top Cited PapersOpen Access
- 1 January 2023
- Published by Association for Computational Linguistics (ACL)
- p. 670-685
Does Transliteration Help Multilingual Language Modeling?
Ibraheem Muhammad Moosa, Mahmud Elahi Akhter, Ashfia Binte Habib. Findings of the Association for Computational Linguistics: EACL 2023. 2023.
- 1 January 2023
- Published by Association for Computational Linguistics (ACL)
- p. 670-685
Top Cited PapersOpen AccessDoes Transliteration Help Multilingual Language Modeling?
Ibraheem Muhammad Moosa, Mahmud Elahi Akhter, Ashfia Binte Habib. Findings of the Association for Computational Linguistics: EACL 2023. 2023.
- 3 July 2020
- Published by Springer Nature
- p. 166-174
Performance Analysis of Different 2D and 3D CNN Model for Liver Semantic Segmentation: A Review
- 3 July 2020
- Published by Springer Nature
- p. 166-174
Performance Analysis of Different 2D and 3D CNN Model for Liver Semantic Segmentation: A Review
- 3 July 2020
- Published by Springer Nature
- p. 229-241
An Analysis of Multi-organ Segmentation Performance of CNNs on Abdominal Organs with an Emphasis on Kidney
- 3 July 2020
- Published by Springer Nature
- p. 229-241
An Analysis of Multi-organ Segmentation Performance of CNNs on Abdominal Organs with an Emphasis on Kidney
- Open Access
A Modified Design of a Hexagonal Circular Photonic Crystal Fiber with Large Negative Dispersion Properties and Ultrahigh Birefringence for Optical Broadband Communication
In this paper, we propose a modified design of a hexagonal circular photonic crystal fiber (HC-PCF) which obtains a large negative dispersion and ultrahigh birefringence simultaneously. The optical properties of the proposed HC-PCF were investigated using the finite element method (FEM) incorporated with a circular perfectly matched layer at the boundary. The simulation results showed large negative dispersion of −1044 ps/nm.km and ultrahigh birefringence of 4.321 × 10−2 at the operating wavelength of 1550 nm for the optimum geometrical parameters. Our proposed HC-PCF exhibited the desirable optical properties without non-circular air holes in the core and cladding region which facilitates the fabrication process. The large negative dispersion of the proposed microstructure over the wide spectral range, i.e., 1350 nm to 1600 nm, and high birefringence make it a suitable candidate for high-speed optical broadband communication and different sensing applications.
Open AccessA Modified Design of a Hexagonal Circular Photonic Crystal Fiber with Large Negative Dispersion Properties and Ultrahigh Birefringence for Optical Broadband Communication
In this paper, we propose a modified design of a hexagonal circular photonic crystal fiber (HC-PCF) which obtains a large negative dispersion and ultrahigh birefringence simultaneously. The optical properties of the proposed HC-PCF were investigated using the finite element method (FEM) incorporated with a circular perfectly matched layer at the boundary. The simulation results showed large negative dispersion of −1044 ps/nm.km and ultrahigh birefringence of 4.321 × 10−2 at the operating wavelength of 1550 nm for the optimum geometrical parameters. Our proposed HC-PCF exhibited the desirable optical properties without non-circular air holes in the core and cladding region which facilitates the fabrication process. The large negative dispersion of the proposed microstructure over the wide spectral range, i.e., 1350 nm to 1600 nm, and high birefringence make it a suitable candidate for high-speed optical broadband communication and different sensing applications.