Amir Zamir
Tenure Track Assistant Professor
Leads the Visual Intelligence and Learning Lab, working on multi-task and multimodal vision models and perception for embodied agents.
Amir Zamir's lab at EPFL works on vision-language & multimodal, embodied & active vision and multimodal & grounded language. 15 papers since 2023, 15 at top venues, mostly ICLR and ECCV. 10 PhD students. 3 alumni, most often to Anthropic and Apple. Applications close .
At a glance
- papers, 2023–2026
- 4/ year
- at T1 venues
- 100%
- PhD students
- 10
- MS students
- 0
- alumni listed
- 3
Papers by yearfrom the lab homepage, 2023–2026
- 4 papers2026 · T1 4 · collecting
- 3 papers2025 · T1 3
- 5 papers2024 · T1 5
- 3 papers2023 · T1 3
T1 venuesOtherVenues: ICLR, ECCV, ICML, NeurIPS, CVPR
Research focuspaper titles that use each topic's words
- Vision-language & multimodal6 papers
- Embodied & active vision6 papers
- Multimodal & grounded language4 papers
- vision models0 papers
- visual intelligence0 papers
Paper list
From the lab homepage.
20264 papers
- ICML (T1)(1D) Ordered Tokens Enable Efficient Test-Time Search
- ICLR (T1)How Well Does GPT-4o Understand Vision? Evaluating Multimodal Foundation Models on Standard Computer Vision Tasks
- ICLR (T1)Multimodality as Supervision: Self-Supervised Specialization to the Test Environment via Multimodality
- ICML (T1)VideoFlexTok: Flexible-Length Coarse-to-Fine Video Tokenization
20253 papers
- TMLR (T1)Controlled Training Data Generation with Diffusion Models
- ICML (T1)FlexTok: Resampling Images into 1D Token Sequences of Flexible Length
- ICLR (T1)Large (Vision) Language Models are Unsupervised In-Context Learners
20245 papers
- NeurIPS (T1)4M-21: An Any-to-Any Vision Model for Tens of Tasks and Modalities
- ECCV (T1)BRAVE: Broadening the visual encoding of vision-language models
- ECCV (T1)Solving Vision Tasks using Simple Photoreceptors instead of Cameras
- ICLR (T1)Unraveling the Key Components of OOD Generalization via Diversification
- ECCV (T1)ViPer: Visual Personalization of Generative Models via Individual Preference Learning
20233 papers
- NeurIPS (T1)4M: Massively Multimodal Masked Modeling
- CVPR (T1)Modality-invariant Visual Odometry for Embodied Navigation
- ICCV (T1)Rapid Network Adaptation: Learning to Adapt Neural Networks Using Test-Time Feedback
Where alumni went3 listed, 100% to industry
Industry 3Academia 0
Most often to
- Anthropic
- Apple
- Google DeepMind
Apply to EPFLFall 2027
EDIC Computer and Communication Sciences PhD
- Round
- Round 1
- Opens
- —
- GRE
- optional
- English
- TOEFL 100 / IELTS 7.0
- Fee
- none
Confirmed on the program page.
Similar labsnearest by research summary
- CNPeking
Chuanmin Jia
- Video understanding
- CNTsinghua
Lu Fang
- Computational imaging & photography
- CNSJTU
Yiling Xu
- 3D vision & neural rendering
- USUCLA
Stefano Soatto
- 3D vision & neural rendering
- SGNTU
Deepu Rajan
- Video understanding
- USUCSD
Zhuowen Tu
- 3D vision & neural rendering
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
EPFLcloses