|
|
| CARVIEW |
Bio
Mengye Ren is an assistant professor of computer science and data science at New York University (NYU). He runs the Agentic Learning AI Lab. Before joining NYU, he was a visiting faculty researcher at Google Brain Toronto working with Prof. Geoffrey Hinton. From 2017 to 2021, he was a senior research scientist at Uber Advanced Technologies Group (ATG) and Waabi, working on self-driving vehicles. He received Ph.D. in Computer Science from the University of Toronto, advised by Prof. Richard Zemel and Prof. Raquel Urtasun. His research focuses on making machine learning more natural and human-like, in order for AIs to continually learn, adapt, and reason in naturalistic environments.
Research
General areas: machine learning, computer vision, representation learning, continual learning, meta-learning, few-shot learning, artificial intelligence.
My key research question is: how do we enable human-like, agent-based machine intelligence to continually learn, adapt, and reason in naturalistic environments? I am interested in the emergence of intelligence by learning from a point-of-view experience. Current research topics in my lab are:
Visual representation learning and planning in the wild
Adaptive agents and foundation models
Few-shot concept learning, reasoning, and abstraction
I am also looking for motivated students at all levels in the following areas:
Continual learning for LLMs, VLMs, and agents
Machine creativity for vision, language, and planning
Teaching
NYU DS-GA 3001: Advanced Topics in Embodied Learning and Vision [2025 spring]
NYU DS-GA 1008 / CSCI-GA 2572: Deep Learning [2024 spring]
NYU CSCI-GA 2565: Machine Learning [2023 fall] [2024 fall]
NYU DS-GA 1003: Machine Learning [2023 spring]
Vector Institute: Deep Learning II [2020 fall]
UofT CSC 411: Machine Learning and Data Mining [2019 winter]
News
2025/12: I will serve as an area chair for ICML 2026.
2025/12: One paper is accepted at World Modeling Workshop 2026.
2025/10: One paper is accepted at NeurIPS 2025 Workshop on Embodied World Models for Decision Making.
2025/08: I will serve as an area chair for ICLR 2026.
2025/06: Two papers [1, 2] are accepted at ICML 2025 workshops.
2025/03: I will serve as a communications chair for NeurIPS 2025.
2025/01: I will serve as an area chair for COLM 2025.
2025/01: One paper is accepted at ICLR 2025.
2024/12: I will teach a new course on Embodied Learning and Vision in Spring 2025 at NYU.
2024/10: I will serve as an associate program chair for CoLLAs 2025.
Group
- Chris Hoang (2023-)
- Jack Lu (2023-)
- Ryan Teehan (2022-)
- Ying Wang (2024-)
- Yanlai Yang (2022-)
Recent Papers
[Full List] [Google Scholar] [dblp]
SkillFactory: Self-distillation for learning cognitive behaviors. Zayne Sprague, Jack Lu, Manya Wadhwa, Sedrick Keh, Mengye Ren, Greg Durrett. arXiv preprint arXiv:2512.04072, 2025. [arxiv]
When does verification pay off? A closer look at LLMs as solution verifiers. Jack Lu, Ryan Teehan, Jinran Jin, Mengye Ren. arXiv preprint arXiv:2512.02304, 2025. [arxiv]
Opinion: Learning intuitive physics may require more than visual data. Ellen Su, Solim LeGris, Todd M. Gureckis, Mengye Ren. NeurIPS 2025 Workshop on Embodied World Models for Decision Making, 2025. [arxiv]
In-context clustering with large language models. Ying Wang, Mengye Ren, Andrew Gordon Wilson. arXiv preprint arXiv:2510.08466, 2025. [webpage] [arxiv]
Local reinforcement learning with action-conditioned root mean squared Q-functions. Frank Wu, Mengye Ren. arXiv preprint arXiv:2510.06649, 2025. [webpage] [arxiv]
Midway Network: Learning representations for recognition and motion from latent dynamics. Christopher Hoang, Mengye Ren. arXiv preprint arXiv:2510.05558, 2025. [webpage] [arxiv]
StreamMem: Query-agnostic KV cache memory for streaming video understanding. Yanlai Yang, Zhuokai Zhao, Satya Narayan Shukla, Aashu Singh, Shlok Kumar Mishra, Lizhu Zhang, Mengye Ren. arXiv preprint arXiv:2508.15717, 2025. [webpage] [arxiv]
Context tuning for in-context optimization. Jack Lu, Ryan Teehan, Zhenbang Yang, Mengye Ren. Workshop on Test-Time Adaptation at ICML, 2025. [webpage] [arxiv] [code]
Discrete JEPA: Learning discrete token representations without reconstruction. Junyeob Baek, Hosung Lee, Christopher Hoang, Mengye Ren, Sungjin Ahn. WoTok: Workshop on Tokenization at ICML, 2025. [arxiv]
Memory Storyboard: Leveraging temporal segmentation for streaming self-supervised learning from egocentric videos. Yanlai Yang, Mengye Ren. CoLLAs, 2025. [webpage] [arxiv] [code]
Replay can provably increase forgetting. Yasaman Mahdaviyeh, James Lucas, Mengye Ren, Andreas S. Tolias, Richard Zemel, Toniann Pitassi. CoLLAs, 2025. [arxiv]
Are LLMs prescient? A continuous evaluation using daily news as the oracle. Hui Dai, Ryan Teehan, Mengye Ren. ICML, 2025. [webpage] [arxiv] [code] [dataset]
PooDLe: Pooled and dense self-supervised learning from naturalistic videos. Alex N. Wang*, Christopher Hoang*, Yuwen Xiong, Yann LeCun, Mengye Ren. ICLR, 2025. [webpage] [arxiv] [code] [dataset] [pdf] [html]
Reawakening knowledge: Anticipatory recovery from catastrophic interference via structured training. Yanlai Yang, Matt Jones, Michael C. Mozer, Mengye Ren. NeurIPS, 2024. [webpage] [arxiv] [code] [pdf] [html]
CoLLEGe: Concept embedding generation for large language models. Ryan Teehan, Brenden M. Lake, Mengye Ren. COLM, 2024. [webpage] [arxiv] [code] [pdf] [html]
ProCreate, don’t reproduce! Propulsive energy diffusion for creative generation. Jack Lu, Ryan Teehan, Mengye Ren. ECCV, 2024. [webpage] [arxiv] [code] [pdf] [html]
Integrating present and past in unsupervised continual learning. Yipeng Zhang, Laurent Charlin, Richard Zemel, Mengye Ren. CoLLAs, 2024. [arxiv] [pdf] [html]
Self-supervised learning of video representations from a child’s perspective. Emin Orhan, Wentao Wang, Alex N. Wang, Mengye Ren, Brenden M. Lake. CogSci, 2024. [arxiv] [pdf] [html]
Learning and forgetting unsafe examples in large language models. Jiachen Zhao, Zhun Deng, David Madras, James Zou, Mengye Ren. ICML, 2024. [arxiv] [pdf] [html]
LifelongMemory: Leveraging LLMs for answering queries in long-form egocentric videos. Ying Wang, Yanlai Yang, Mengye Ren. arXiv preprint 2312.05269, 2023. [arxiv] [pdf] [html]
Recent Talks
- Lifelong concept learning.
- Lifelong and human-like learning in foundation models.
- Visual learning in the open world.