Erfan Shayegani

3 papers · 2024–2025 · 3 venues · across top CS/AI venues

Authors are matched by name, so researchers who share a name ("Yang Liu", "Wei Wang") appear as a single profile and their paper counts, venues and achievements are combined. How identity is handled.

Achievements

Jump to papers ↓

🌍 Conference Polyglot · Bronze 🌉 Interdisciplinary Bridge · Bronze 🐣 Hot Topic Early Bird ❓ The Questioner 📈 Trend Setter

Venues

EMNLP (1) ICLR (1) ICML (1)

Top co-authors

Yue Dong (3) Nael B. Abu-Ghazaleh (2) Trishna Chakraborty (2) Chengyu Song (2) Amit Roy-Chowdhury (2) Zikui Cai (1) Rohit Lal (1) Arindam Dutta (1) Nael Abu-Ghazaleh (1) Saketh Bachu (1)

Research topics

Artificial Intelligence (3) Core AI (3) Deep Learning (2) Learning Types · Deep Learning (2) Machine Learning (1) Reinforcement Learning (1) Learning Types · Machine Learning (1) Methods (1)

Keywords

vision-language model (3) adversarial attack (2) early exit (1) jailbreak attack (1) multimodal language model (1) attack success rate (1) cross-modality alignment (1) harmful information (1) textual unlearning (1) cross-modality attack (1) clipped-proximal policy optimization (1) image encoder layers (1) machine unlearning (1) layer-wise ppo (1) safety alignment (1)

Papers

Layer-wise Alignment: Examining Safety Alignment Across Image Encoder Layers in Vision Language Models ICML 2025 Can Textual Unlearning Solve Cross-Modality Safety Alignment? EMNLP 2024 Jailbreak in pieces: Compositional Adversarial Attacks on Multi-Modal Language Models ICLR 2024