Hi, I am Howard (Jingbiao) Mei, currently a third year PhD student at Machine Inteligence Lab, University of Cambridge. My interests are Deep Learning and its various applications.
My current research is on multimodal large language models and retrieval systems.
Here is some of my latest work:
- BayesFT: Bayesian Optimization for Fault Tolerant Neural Network Architecture. 2021 58th ACM/IEEE Design Automation Conference link
- Fine-grained Late-interaction Multi-modal Retrieval for Retrieval Augmented Visual Question Answering. NeurIPS2023 link
- Improving Hateful Meme Detection through Retrieval-Guided Contrastive Learning. ACL2024 Main link
- PreFLMR: Scaling Up Fine-Grained Late-Interaction Multi-modal Retrievers. ACL2024 Main Link
- Control-DAG: Constrained Decoding for Non-Autoregressive Directed Acyclic T5 using Weighted Finite State Automata. NAACL link 2024 Main
- On Extending Direct Preference Optimization to Accommodate Ties link
Here is a copy of my CV to download.
嗨，我是Howard，目前是剑桥大学机器智能实验室的三年级博士生。我感兴趣的领域是深度学习及其多种应用。
我当前的研究方向是多模态大语言模型和检索系统。
以下是我的一些最新工作：
- BayesFT: 面向容错神经网络架构的贝叶斯优化，2021年第58届ACM/IEEE设计自动化大会 链接
- Fine-grained Late-interaction Multi-modal Retrieval for Retrieval Augmented Visual Question Answering，NeurIPS 2023 链接
- Improving Hateful Meme Detection through Retrieval-Guided Contrastive Learning，ACL 2024 主会 链接
- PreFLMR: 扩展细粒度后期交互多模态检索模型，ACL 2024 主会 链接
- Control-DAG: 使用加权有限状态自动机进行非自回归有向无环T5的受约束解码，NAACL 2024 主会 链接
- On Extending Direct Preference Optimization to Accommodate Ties 链接
我的最新简历可下载。
