🤝 Open to Collab

4 64 11

Juanxi Tian

Juanxi

https://tianshijing.github.io

AI & ML interests

Efficient AI & Gen AI

Recent Activity

upvoted a paper 7 days ago

DreamX-World 1.0: A General-Purpose Interactive World Model

upvoted a paper 7 days ago

JoyAI-VL-Interaction: Real-Time Vision-Language Interaction Intelligence

upvoted a paper 9 days ago

MiniMax Sparse Attention

View all activity

Organizations

Posts 5

Post

187

🧐 BlogXiv provides a scholarly discovery interface for researchers to trace emerging ideas, compare technical perspectives, and engage with high-quality research communication.

Welcome any contributions such as star and submissions, thanks

Github: https://github.com/OpenEnvision/BlogXiv
Website: https://openenvision.github.io/BlogXiv/

View all Posts

Articles 1

Article

models 0

None public yet

datasets 2

Juanxi/PhyVideo

Viewer • Updated Oct 18, 2025 • 6 • 23

Juanxi/Phys101

Viewer • Updated Oct 18, 2025 • 103 • 69

Juanxi Tian

AI & ML interests

Recent Activity

Organizations

Posts 5

Articles 1

Backbone-Optimizer Coupling Bias: The Hidden Co-Design Principle

Collections 3

Taming LLMs by Scaling Learning Rates with Gradient Grouping

Switch EMA: A Free Lunch for Better Flatness and Sharpness

Unveiling the Backbone-Optimizer Coupling Bias in Visual Representation Learning

OpenMixup: Open Mixup Toolbox and Benchmark for Visual Representation Learning

Envision: Benchmarking Unified Understanding & Generation for Causal World Process Insights

MergeVQ: A Unified Framework for Visual Generation and Representation with Disentangled Token Merging and Quantization

WideRange4D: Enabling High-Quality 4D Reconstruction with Wide-Range Movements and Scenes

Trans4D: Realistic Geometry-Aware Transition for Compositional Text-to-4D Synthesis

Taming LLMs by Scaling Learning Rates with Gradient Grouping

Switch EMA: A Free Lunch for Better Flatness and Sharpness

Unveiling the Backbone-Optimizer Coupling Bias in Visual Representation Learning

OpenMixup: Open Mixup Toolbox and Benchmark for Visual Representation Learning

Envision: Benchmarking Unified Understanding & Generation for Causal World Process Insights

MergeVQ: A Unified Framework for Visual Generation and Representation with Disentangled Token Merging and Quantization

WideRange4D: Enabling High-Quality 4D Reconstruction with Wide-Range Movements and Scenes

Trans4D: Realistic Geometry-Aware Transition for Compositional Text-to-4D Synthesis

Papers 12

models 0

datasets 2

Juanxi/PhyVideo

Juanxi/Phys101

Juanxi Tian

AI & ML interests

Recent Activity

Organizations

Posts 5

Articles 1

Backbone-Optimizer Coupling Bias: The Hidden Co-Design Principle

Collections 3

Papers 12

models 0

datasets 2 Sort: Recently updated

datasets 2