影响力指数
论文质量、代表作、近期表现、广度与样本量置信度综合计算
71.52/100
前 2.4%
全站排名 #1,526
发表论文18 篇
平均评分
年均产出6.0 篇/年
Dongyoon Han
研究方向
Representation learning · Language modeling · Multimodal learning · Unlearning · Efficient Deep Learning · Deep neural network architectures
28
RL makes MLLMs see better than SFT
ICLR 2026Poster
三作20
Beyond Synthetic Replays: Turning Diffusion Features into Few-Shot Class-Incremental Learning Knowledge
ICLR 2026Rejected
三作17
Aligned Novel View Image and Geometry Synthesis via Cross-modal Attention Instillation
ICLR 2026Poster
30
SyMerge: From Non-Interference to Synergistic Merging via Single-Layer Adaptation
ICLR 2026Rejected
三作5
When Test-Time Adaptation Meets Self-Supervised Model
ICLR 2026Withdrawn
三作20
Less is Not Worse: Effective Reasoning Without Complete Reasoning Traces
ICLR 2026Rejected
通讯4
Exploring Conditions for Diffusion models in Robotic Control
ICLR 2026Withdrawn
三作23
StoryCoder: Bridging Narratives and Formality for Code Generation in LLMs
ICLR 2026Withdrawn
二作4
Vision Transformers Secretly Crave Noise
ICLR 2026Withdrawn
通讯19
Morphing Tokens Draw Strong Masked Image Models
ICLR 2025Poster
三作25
DaWin: Training-free Dynamic Weight Interpolation for Robust Adaptation
ICLR 2025Poster
通讯27
Token Bottleneck: One Token to Remember Dynamics
NeurIPS 2025Poster
二作31
NegMerge: Consensual Weight Negation for Strong Machine Unlearning
ICLR 2025Rejected
二作22
Token-Supervised Value Models for Enhancing Mathematical Problem-Solving Capabilities of Large Language Models
ICLR 2025Poster
15
Peri-LN: Revisiting Normalization Layer in the Transformer Architecture
ICML 2025Poster
13
NegMerge: Sign-Consensual Weight Merging for Machine Unlearning
ICML 2025Poster
二作