ETF Dashboard
View how BlackRock’s ETFs Change Daily
Posts by Collection
portfolio
projects
Probing LeWM: What Does a JEPA World Model Actually Learn?
Published:
LeWM is a latent world model for robot manipulation trained with a JEPA objective. It encodes visual observations with a ViT, then learns a transformer predictor that forecasts future latent states conditioned on actions — without decoding back to pixels.
Skin Cancer Detection (Kaggle Competition)
2D Medical Image Classification for Kaggle ISIC 2024 Competition
publications
A Transformer-Based Classification System for Volcanic Seismic Signals
Published in Unpublished, 2022
This paper is about applying the idea of multi-headed self-attention to the volcanic event classification problem.
Download here
Leveraging Auxiliary Domain Parallel Data In Intermediate Task Fine-Tuning For Low-Resource Translation
Published in The Eleventh International Conference on Learning Representations, 2023
This paper is about fine-tuning strategies for neural machine translation of low-resource languages.
Download here
A deep learning pipeline for three-dimensional brain-wide mapping of local neuronal ensembles in teravoxel light-sheet microscopy
Published in Nature Methods, 2025
This paper introduces ACE, an end-to-end pipeline using 3D deep learning segmentation and cluster-wise statistical algorithms for unbiased mapping of local neuronal activity and connectivity in teravoxel light-sheet microscopy data.
Download here
Mapping cellular targets of covalent cancer drugs in the entire mammalian body
Published in Cell, 2026
This paper introduces vCATCH (volumetric clearing-assisted tissue click chemistry), enabling cellular-resolution imaging of covalent drug targets across the whole adult mouse body using 3D click labeling, HYBRiD imaging, and virtual reality visualization.
Download here
talks
The Tenth Annual Canadian Statistics Student Conference
Published:
Selected as one of few students to present an academic poster at the Annual Canadian Statistics Student Conference.
Neural Machine Translation for Low-Resource Languages
Published:
This talk presented research that I contributed to, headed by Professor Annie En-Shiun Lee from the University of Toronto. The team working on this research includes individuals from the University of Toronto, University of Moratuwa, University of New York Tirana, and Microsoft.
teaching
Teaching experience 1
Undergraduate course, University 1, Department, 2014
This is a description of a teaching experience. You can use markdown like any other post.
Teaching experience 2
Workshop, University 1, Department, 2015
This is a description of a teaching experience. You can use markdown like any other post.