Timothy Dai | Researcher Profile | Sotabase

Career

· ML Foundation Engineer, NVIDIA2024–

· Teaching Assistant, Stanford University2024–2024

· ML Foundation Intern, NVIDIA2023–2023

· Master of Science - MS, Computer Science, Stanford University2023–2024

· Intern, Meta2022–2022

· Deep Learning Intern, NVIDIA2022–2022

· Participant, Stanford Machine Learning Group2021–

· Research Assistant, Stanford University2021–2023

· Schneider Fellow, U.S. Green Building Council2021–2021

· Research Assistant, Stanford Doerr School of Sustainability2020–2020

· Bachelor of Science - BS, Computer Science, Stanford University2019–2024

Research on the Application of Deep Learning-based BERT Model with Additional Pretraining and Multitask Fine-Tuning

cited

BERT: Battling Overfitting with Multitask Learning and Ensembling

Combining Contrastive Learning with Layer Utility Analysis and Experimental Multi-Task Finetuning to Improve mini-BERT Performance

Fine-tuning minBERT for multi-task prediction

How to make your BERT model an xBERT in multitask learning?

Improving minBERT on Downstream Tasks Through Combinatorial Extensions

MinBERT and Downstream Tasks

Multi-task BERT Fine-Tuning with Gradient Tricks

QAN-et al.: Exploring Extensions on QANet

2022

Rhapsody on a Theme of Gradient Surgery: Variations to Improve minBERT for Multi-Task Learning

Sotabase