Xin Geng

24

Papers

16

Total Citations

Papers (24)

Mimic In-Context Learning for Multimodal Tasks

Building Variable-Sized Models via Learngene Pool

Selective Label Enhancement Learning for Test-Time Adaptation

Improved Noise Schedule for Diffusion Training

One-Shot Knowledge Transfer for Scalable Person Re-Identification

Inheriting Generalized Learngene for Efficient Knowledge Transfer across Multiple Tasks

Transformer as Linear Expansion of Learngene

Correlation-Induced Label Prior for Semi-Supervised Multi-Label Learning

ULAREF: A Unified Label Refinement Framework for Learning with Inaccurate Supervision

Aligned Objective for Soft-Pseudo-Label Generation in Supervised Learning

Vision Transformers as Probabilistic Expansion from Learngene

Logistic Boosting Regression for Label Distribution Learning

Multi-Person Pose Estimation With Enhanced Channel-Wise and Spatial Information

Label Distribution Learning on Auxiliary Label Space Graphs for Facial Expression Recognition

Towards Effective Visual Representations for Partial-Label Learning

Efficient Diffusion Training via Min-SNR Weighting Strategy

Learning with Partial-Label and Unlabeled Data: A Uniform Treatment for Supervision Redundancy and Insufficiency

WAVE: Weight Templates for Adaptive Initialization of Variable-sized Models

Redefining <Creative> in Dictionary: Towards an Enhanced Semantic Understanding of Creative Generation

Provably Consistent Partial-Label Learning

Instance-Dependent Partial Label Learning

One Positive Label is Sufficient: Single-Positive Multi-Label Learning with Label Enhancement

Exploring Diverse In-Context Configurations for Image Captioning

Learning From Biased Soft Labels