Meng Zhou
I'm a Quantitative Researcher at Point72. I was a Master's student at Carnegie Mellon University, where I studied machine learning and large-scale systems.
I received my Bachelor's Degree at Shanghai Jiao Tong University (SJTU), where I worked on the interpretability of deep learning with Professor Quanshi Zhang. I also worked on self-supervised learning in NLP and machine learning for healthcare at UC San Diego with Professor Pengtao Xie.
I spent a wonderful time doing NLP research and building production tools at Alibaba DAMO Academy. I also worked as a quantitative research intern at Higgs Asset, building high-frequency predictive models in the financial market, which is a lot of fun.
I'm interested in building high-performance reliable machine learning solutions to real-world problems. I'm exicited about the intersection of machine learning and systems, which is a promising direction to deal with large-scale datasets and models.
Email  / 
Google Scholar  / 
Linkedin  / 
Github  / 
More
|
|
Achieved a 4.0/4.0 GPA during my graduate studies.
Research Publications/Preprints
|
Enhancing Cross-lingual Prompting with Dual Prompt Augmentation
Meng Zhou, Xin Li, Yue Jiang, Lidong Bing
Findings of the Association for Computational Linguistics: ACL 2023
Self-supervised Regularization for Text Classification
Meng Zhou,
Zechen Li,
Pengtao Xie
Transactions of the Association for Computational Linguistics, Volume 9 (TACL), 2021
On the Generation of Medical Dialogs for COVID-19
Meng Zhou,
Zechen Li, Bowen Tan, Guangtao Zeng, Wenmian Yang, Xuehai He, Zeqian Ju, Subrato Chakravorty, Shu Chen, Xingyi Yang, Yichen Zhang, Qingyang Wu, Zhou Yu, Kun Xu, Eric Xing, Pengtao Xie
Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (ACL-IJCNLP), 2021
Towards Theoretical Analysis of Transformation Complexity of ReLU DNNs
Jie Ren, Mingjie Li, Meng Zhou, Shih-Han Chan, Quanshi Zhang
Proceedings of the 39th International Conference on Machine Learning (ICML) 2022
Towards a Unified Game-Theoretic View of Adversarial Perturbations and Robustness
Jie Ren, Die Zhang, Yisen Wang, Lu Chen, Zhanpeng Zhou, Yiting Chen, Xu Cheng, Xin Wang, Meng Zhou, Jie Shi, Quanshi Zhang
Advances in Neural Information Processing Systems 34 (NeurIPS), 2021
Meddialog: A large-scale medical dialogue dataset
Guangtao Zeng, Wenmian Yang, Zeqian Ju, Yue Yang, Sicheng Wang, Ruisi Zhang, Meng Zhou, Jiaqi Zeng, Xiangyu Dong, Ruoyu Zhang, Hongchao Fang, Penghui Zhu, Shu Chen, Pengtao Xie
Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), 2020
CERT: Contrastive self-supervised learning for language understanding
Hongchao Fang, Sicheng Wang, Meng Zhou, Jiayuan Ding, Pengtao Xie
arXiv:2005.12766
From Clozing to Comprehending: Retrofitting Pre-trained Language Model to Pre-trained Machine Reader
Weiwen Xu, Xin Li, Wenxuan Zhang, Meng Zhou, Lidong Bing, Wai Lam, Luo Si
arXiv:2212.04755
|
Projects
Here are a few projects that I enjoyed much:
|
CMU 11785 Deep Learning Kaggle Competition
Achieved 1st place in the ASR and attention-based speech recognition task and 2nd
place in the face verification task among ∼200 students. Additional details regarding these competitions can be found here and here.
CMU 15719 Advanced Cloud Computing
(1) Utilized Terraform for the automation of AWS resource provisioning and fine-tuned AWS Auto-Scaling Policies and Controller to suit various workloads, leading to cost reductions and a 30% improvement in RPS.
(2) Constructed a large-scale ETL Processing pipeline using Spark to process the CommonCrawl dataset, which contains 2.85 billion web pages.
Needle: NEcessary Elements of Deep Learning
Implemented forward and backward functionality of various operators to support a minimal functional PyTorch-like deep learning library. By leveraging these operators, built common neural network modules, e.g. Linear layer, Residual Block, Convolution layer, and LSTM. Further, some hardware acceleration techniques like tiling in the array backend for matrix multiplication were used to speed up the Needle framework.
CloudFS: A Hybrid Linux File System Backed by Cloud Storage
Built a fully-functional hybrid file system using a local SSD and Amazon S3, known as CloudFS. The system supports block-level deduplication to reduce cloud storage costs and snapshots to back up the state of the file system at a certain timestamp. Caching further reduces the data transfer cost by about 50%.
BusTub: A Relational Database Management System
Implemented essential components for BusTub including: (1) A buffer pool manager moving physical pages back and forth from main memory to disk; (2) A disk-backed extendible hash table for indexing; (3) A query executor with the iterator query processing model; (4) A lock manager used to support concurrent query execution.
Xv6: A Unix-like Operating System with RISC-V
Developed various features in the xv6 operating system, including: (1) Copy-on-write fork to defer allocating copying physical memory pages
(2) System calls like symlink, sysinfo, sigalarm and mmap, etc.; (3) Concurrent memory allocator and buffer cache.
|
11-785 Introduction to Deep Learning (PhD Level), Fall 2023, CMU
|
Reviewer for:  NeurIPS 2022-2023,  ICML 2022,  ACL 2022,  ACL Rolling Review,  AAAI 2021,  NAACL 2021,  ICLR 2021 workshop “Machine Learning for Preventing and Combating Pandemics”
|
- Cat Lover (as shown in the photo...)
- Texas Hold'em Calling Machine
- A Fan of Coffee/Bubble Tea
- To see the world, things dangerous to come to, to see behind walls, to draw closer, to find each other and to feel. That is the purpose of life.
- Still having possibilities is a luxury.
- Earn your own respect is not easy, but it deserves.
|