Tsinghua AI

Canvas360 not only redefines panoramic generation with geometry-aware techniques but also delivers a dataset of 1 million samples that transforms how we approach in-context tasks.

State-of-the-art performance in endoscopic referring segmentation is now achievable through a novel attribute retrieval approach, transforming how we interpret complex medical imagery.

The Temporal Ratio reveals how attention shifts between future and present frames can predict a model's ability to generalize compositional tasks in video-action contexts.

Many text-to-image models are safer than expected, but a subset poses significant risks that traditional evaluation methods fail to capture.

Multimodal unlearning could revolutionize how we handle sensitive data in AI, enabling targeted removal without sacrificing model performance.

Existing text-to-image models struggle to capture individual aesthetic preferences, but PIPBench reveals critical gaps in their performance that could redefine personalized image generation.

Cortex outperforms traditional models by enabling zero-shot execution of complex long-horizon tasks, bridging the gap between high-level planning and low-level execution.

Extracting interaction cues from a frozen video model enables robots to achieve up to 90.6% success in manipulation tasks without costly rollout processes.

SEAM reduces boundary jerk by 28% and transition discontinuity by 27%, all while preserving task success and computational efficiency.

CuRe transforms video captioning reward design by shifting from holistic evaluations to precise claim-level verification, significantly boosting factual accuracy and diversity in generated captions.

Surpassing human performance in gaze estimation, PaGE closes the human-AI gap by over 60% while remaining lightweight for real-world applications.

HUGS achieves a remarkable balance between grasp success and diversity, synthesizing 3.2 million grasps that can adaptively handle objects from screws to large boxes.

UI-MOPD achieves a remarkable balance between retaining existing capabilities and adapting to new platforms, with task success rates that challenge conventional approaches in GUI agent learning.

Current avatar systems are more diverse than ever, yet foundational prior learning is often overlooked in discussions of photorealistic digital humans.

Automatically constructed data can dramatically enhance the temporal localization abilities of audio models, overcoming the limitations of manual annotation.

ResearchStudio-Reel not only automates research dissemination but does so with unprecedented quality, outperforming both traditional methods and leading LLMs in aesthetic appeal and information accuracy.

Reducing sampling steps from 50 to just 8 without sacrificing quality could revolutionize how we approach generative modeling.

Bridge-WA achieves superior task performance by predicting where and how the world will change, enabling robots to focus on relevant scene dynamics rather than irrelevant visual details.

Attention mechanisms can drastically improve pose sensing accuracy in the face of challenging visual conditions like occlusion and weak textures.

Unconstrained egocentric video generation now achieves unprecedented fidelity and control by disentangling hand and camera motion with a novel 3D-aware representation.

AutoMIA can generate complex 3D illusions in under 80 seconds, revolutionizing the intersection of art and computational design.

Achieving 100% task success in closed-loop execution, Embodied.cpp revolutionizes how embodied AI models are deployed across diverse hardware platforms.

MiShield outperforms leading moderation tools by effectively identifying harmful semantics in multi-image content that appear benign in isolation.

TTP enables robots to learn dexterous manipulation from human tactile experiences, achieving unprecedented performance in complex tasks.

MG-RWKV achieves state-of-the-art TFL performance with a groundbreaking O(T) complexity, redefining efficiency in audio-visual content authenticity verification.

Late visual-token updates can be safely ignored, leading to a 33.7% reduction in computational load without sacrificing performance.

Positional leakage in 3D masked autoencoders can be mitigated, leading to significantly improved semantic representation quality.

CoLT slashes inference time by over 10x while enabling multi-modal models to reason more efficiently with fewer steps.

Transforming image quality assessment from a single score to a nuanced diagnosis of multiple quality issues could revolutionize smartphone ISP tuning.

OVOW achieves unprecedented accuracy and speed in reconstructing 4D scenes from a single video, making it a game-changer for physics simulation in AI.

DynFly achieves a remarkable 4.69 improvement in navigation performance, showcasing how dynamic-aware trajectory generation can transform UAV navigation in complex urban environments.

Achieving a 22.3% word error rate with just 240 ms latency, LipsFlow redefines the capabilities of Visual Speech Recognition in challenging multi-speaker environments.

TACO reveals that agentic models can learn to optimize tool usage without external judges, achieving higher accuracy and efficiency in multimodal tasks.

GUICrafter achieves superior GUI agent performance with just a fraction of the data, revolutionizing the way we think about training in data-scarce environments.

Current T2I models fall short in scientific reasoning, but fine-tuning on the new SciIR dataset boosts performance by over 20%.

GeoEdit achieves unprecedented geometric accuracy and identity fidelity in object editing, overcoming the limitations of existing diffusion-based methods.

Fine-grained contact states can be distinguished through the dynamic correlation of tactile motion, transforming how we approach contact-rich manipulation in robotics.

OpenSPM achieves 85.6% task success with a control frequency of 1033.3 Hz, revolutionizing high-frequency robotic manipulation with minimal computational demands.

Effective keyframe extraction can boost MLLM performance by over 2% on complex video-guided tasks, revealing a critical link between video understanding and procedural learning.

Evaluating LLM agents in microservice failure diagnosis reveals that traditional outcome-based benchmarks miss critical reasoning processes, which these new datasets effectively capture.

Achieving 94.2% precision in item knowledge production at an unprecedented scale, Oxygen AIIC transforms e-commerce item management.

ProMSA achieves superior accuracy in KB-VQA by dynamically selecting retrieval strategies, outperforming traditional fixed pipelines.

A transparent probe-success rule boosts robot policy selection success rates by over 14 percentage points, revealing the hidden power of pre-deployment evaluations.

Pressure integration in humanoid motion imitation significantly enhances accuracy and stability, revealing the limitations of traditional vision-based methods.

HarmVideoBench reveals that existing benchmarks miss critical layers of harmful video understanding, while a new method boosts model accuracy by over 20%.

ViQ achieves a groundbreaking balance between semantic richness and detail in visual representations, enabling efficient multimodal training without sacrificing quality.

ReMMD-Agent achieves a remarkable 41.80% accuracy in detecting misinformation across complex multilingual and multi-image scenarios while slashing verification costs by up to 80%.

Counterfactual controllability in video generation could be the key to creating self-evolving world models that understand and adapt to their actions.

Strong proprietary models falter in grounding their predictions, revealing a critical flaw in current VideoQA systems that could reshape evaluation standards.

Achieving a 14-point boost in grounding accuracy, VistaRef redefines how we approach spatial orientation in AR and human-robot interaction.