Dongxia Wang
7 papers · 2015–2026 · 4 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 ↓
๐
Conference Polyglot ยท Bronze
๐
Academic Marathon ยท Half
๐
Interdisciplinary Bridge ยท Bronze
๐บ๏ธ
Taxonomy Completionist
(11)
Top co-authors
Research topics
Keywords
large language model
(2)
anomaly detection
(1)
preference alignment
(1)
flow matching
(1)
safety alignment
(1)
sentence similarity
(1)
backdoor attack
(1)
support vector machine
(1)
text embedding
(1)
large vision-language model
(1)
recommender system
(1)
attack success rate
(1)
attention analysis
(1)
data distillation
(1)
tokenization robustness
(1)
text embedding model
(1)
transformer model
(1)
user simulation
(1)
sticky token detection
(1)
vector steering
(1)
Papers
FlowHijack: A Dynamics-Aware Backdoor Attack on Flow-Matching Vision-Language-Action Models
CVPR 2026
LLM-VA: Resolving the Jailbreak-Overrefusal Trade-off via Vector Alignment
ACL 2026
Mirroring Users: Towards Building Preference-aligned User Simulator with User Feedback in Recommendation
ACL 2026
Defending LVLMs Against Vision Attacks Through Partial-Perception Supervision
ICML 2025
Sticking to the Mean: Detecting Sticky Tokens in Text Embedding Models
ACL 2025
Quantifying and Improving the Robustness of Trust Systems
IJCAI 2015
Quantifying Robustness of Trust Systems against Collusive Unfair Rating Attacks Using Information Theory
IJCAI 2015