I'm a third-year PhD from the University of Amsterdam in the Netherlands, supervised by Dr. Christof Monz and Dr. Vlad Niculae.
My main research interest is efficient pretraining, finetuning and inference of LLM, including memory/parameter-efficient finetuning, quantization, sample-efficient training and so on.
I obtained my master degree from RWTH Aachen University in Germany, and my bachelor degree from Shanghai Jiao Tong University in China.
11/07/2024: I will give a talk about my NeurIPS 2024 accepted paper at eBay.
10/21/2024: I will attend EMNLP 2024 in person and present three papers, including one EMNLP main, one EMNLP Findings and one WMT system paper.
11/07/2024: Give a talk about my NeurIPS 2024 acceptance at eBay (3-in-1: 2D Rotary Adaptation for Efficient Finetuning, Efficient Batching and Composability).
12/05/2023: Give a talk about my NeurIPS 2023 acceptance at eBay (Make Pre-trained Model Reversible: From Parameter to Memory Efficient Fine-Tuning).
11/30/2023: Give a talk about my NeurIPS 2023 acceptance at aiXplain (Make Pre-trained Model Reversible: From Parameter to Memory Efficient Fine-Tuning).
🎉: co-first author 💯: supervised student
3-in-1: 2D Rotary Adaptation for Efficient Finetuning, Efficient Batching and Composability
Baohao Liao, Christof Monz
NeurIPS 2024 / pdf
IKUN for WMT24 General MT Task: LLMs Are here for Multilingual Machine Translation
Baohao Liao, Christian Herold, Shahram Khadivi, Christof Monz
WMT 2024 / pdf
7 first places and 5 second places in General Task
Apiq: Finetuning of 2-bit quantized large language model
Baohao Liao, Christian Herold, Shahram Khadivi, Christof Monz
The SIFo Benchmark: Investigating the Sequential Instruction Following Ability of Large Language Models
Xinyi Chen, Baohao Liao, Jirui Qi, Panagiotis Eustratiadis, Christof Monz, Arianna Bisazza, Maarten de Rijke
EMNLP Findings 2024 / pdf / code and dataset
Make Pre-trained Model Reversible: From Parameter to Memory Efficient Fine-Tuning
Baohao Liao, Shaomu Tan, Christof Monz
Ask Language Model to Clean Your Noisy Translation Data
Quinten Bolding 💯, Baohao Liao, Brandon James Denis, Jun Luo, Christof Monz
EMNLP Findings 2023 / pdf
Parameter-Efficient Fine-Tuning without Introducing New Latency
Baohao Liao, Yan Meng, Christof Monz
ACL 2023 / pdf
Mask more and mask later: Efficient pre-training of masked language models by disentangling the [MASK] token
Baohao Liao, David Thulke, Sanjika Hewavitharana, Hermann Ney, Christof Monz
EMNLP Findings 2022 / pdf
Back-translation for large-scale multilingual machine translation
Baohao Liao, Shahram Khadivi, Sanjika Hewavitharana
WMT 2021 / pdf
2 second-places in Large-Scale Multilingual Machine Translation Competition
Unifying input and output smoothing in neural machine translation
Yingbo Gao🎉, Baohao Liao🎉, Hermann Ney
COLING 2020 / pdf
Multi-agent mutual learning at sentence-level and token-level for neural machine translation
Baohao Liao, Yingbo Gao, Hermann Ney
EMNLP Findings 2020 / pdf
WMT shared task
7 first-places and 5 second-places at WMT 2024 general task.
2 second-places at WMT 2021 Large-Scale Multilingual Machine Translation Competition.
11/1315 (🥇) for Google American Sign Language Fingerspelling Recognition in 2023.
29/744 (🥈) for Bengali.AI Speech Recognition in 2023.
Robotics Competition
4th place and Most Creative Award in the 26th International Design Contest Robocon in 2015.
2nd prize in DJI’s RoboMaster National Competition in 2015.
Scholarship
German Scholarship (top 20%) in 2018.
Chinese National Scholarship (top 2%) in 2015.
Global embedding learning from different modalities (link)
Baohao Liao, Sanjika Hewavitharana, Michael Damian Kozielski, Friedrich Leonard Dahlmann, Shahram Khadivi
eBay
Systems and methods for creating listing for items for sale in an electronic marketplace (link)
Ramesh Periyathambi, Tomer Lancewicki, Sanjika Hewavitharana, Ryan Reeves, Senthil Kumar Padmanabhan, Daniel Stein, Luther S Boorn, Baohao Liao, Simon Alexander Becker, Sivan Elkis, Saral Bharathi Sukumar Jeyaseelan
eBay
Teaching assistants:
Supervised interns
2024: Matteo Nulli, intern at eBay for multimodal learning.
2024: Stefan Vasilev, intern at eBay for machine unlearning.
Supervised thesis
2022-2023: Quiten Bolding's master thesis jointly with Huawei Amsterdam, with a publication at EMNLP Findings 2023.
Reviewer
2024: ARR Feb/June, NeurIPS, ICLR, TMLR, WMT
2023: EMNLP, ACL
2022: EMNLP, ACL
