Generation Models Know Space: Unleashing Implicit 3D Priors for Scene Understanding

▲ 63 🎓 H-EmbodVis

• A video diffusion model is repurposed as a latent world simulator to enhance multimodal large language models with implicit 3D structural priors and p

SAMA: Factorized Semantic Anchoring and Motion Alignment for Instruction-Guided Video Edit

▲ 57 🏢 Baidu

• SAMA presents a factorized approach to video editing that separates semantic anchoring from motion modeling, enabling instruction-guided edits with pr

FASTER: Rethinking Real-Time Flow VLAs

▲ 41 🎓 The University of Hong Kong

• Fast Action Sampling for ImmediaTE Reaction (FASTER) reduces real-time reaction latency in Vision-Language-Action models by adapting sampling schedule

3DreamBooth: High-Fidelity 3D Subject-Driven Video Generation Model

▲ 41 🎓 Yonsei University

• A novel 3D-aware video customization framework is presented that decouples spatial geometry from temporal motion using a 1-frame optimization approach

Bridging Semantic and Kinematic Conditions with Diffusion-based Discrete Motion Tokenizer

▲ 34 🎓 MMLab@NTU

• A three-stage motion generation framework combines discrete token-based planning with diffusion-based synthesis to improve controllability and fidelit

MonoArt: Progressive Structural Reasoning for Monocular Articulated 3D Reconstruction

▲ 28 🎓 MMLab@NTU

• MonoArt presents a unified framework for reconstructing articulated 3D objects from single images through progressive structural reasoning that enable

Nemotron-Cascade 2: Post-Training LLMs with Cascade RL and Multi-Domain On-Policy Distilla

▲ 28 🏢 NVIDIA

• Nemotron-Cascade 2 is a 30B parameter Mixture-of-Experts model with 3B activated parameters that achieves exceptional reasoning and agentic capabiliti

Cubic Discrete Diffusion: Discrete Visual Generation on High-Dimensional Representation To

▲ 26 🎓 University of Hong Kong

• CubiD is a discrete generation model for high-dimensional representations that enables fine-grained masking and learns rich correlations across spatia

Evaluating Counterfactual Strategic Reasoning in Large Language Models

🏛️ National Technical University of Athens AI & Machine Learning

• Current large language models exhibit significant social biases that are not adequately addressed by existing debiasing
• The paper proposes UGID, a unified graph isomorphism approach for effectively debiasing large language models.

NavTrust: Benchmarking Trustworthiness for Embodied Navigation

🏛️ University of California, Riverside AI & Machine Learning

• Automated Alzheimer's Disease detection methods often lack alignment with clinical constructs, leading to potential misd
• The paper presents Agentic Cognitive Profiling, which realigns automated detection methods with clinical validity.

HierarchicalKV: A GPU Hash Table with Cache Semantics for Continuous Online Embedding Stor

🏛️ NVIDIA Systems & Infrastructure

• Traditional GPU hash tables waste memory by preserving every inserted key, leading to inefficiencies in high-bandwidth m
• HierarchicalKV introduces a GPU hash table with cache semantics that optimizes memory usage for continuous online embedd

Towards Exponential Quantum Improvements in Solving Cardinality-Constrained Binary Optimiz

🏛️ University of Cambridge Theory & Algorithms

• Cardinality-constrained binary optimization is a fundamental computational challenge with wide-ranging applications.
• Introduction of a Grover-based quantum algorithm that significantly improves the efficiency of solving these optimizatio

Post-Quantum Cryptography from Quantum Stabilizer Decoding

🏛️ Massachusetts Institute of Technology Applications

• Current post-quantum cryptography relies on a limited number of hardness assumptions, posing significant security risks.
• A new framework for post-quantum cryptography based on quantum stabilizer decoding techniques.

Exposing Cross-Modal Consistency for Fake News Detection in Short-Form Videos

🏛️ Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) Other CS

• Short-form video platforms are significant sources of misinformation, where cross-modal relationships can mislead viewer
• A framework that exposes cross-modal consistency for detecting fake news in short-form videos.

Risk-Based Auto-Deleveraging

🏛️ Columbia University Quantitative Finance

• The mechanisms of auto-deleveraging in cryptocurrency futures exchanges are poorly understood and under-researched.
• This paper proposes a risk-based framework to analyze auto-deleveraging mechanisms in the context of cryptocurrency futu

Consistencies in Social Ranking

🏛️ The University of Tokyo Economics

• Ranking individuals based on performance across different contexts is often inconsistent and problematic.
• A systematic approach to establish consistencies in social ranking across various coalitions.