Zhen-Yu Wu

Native desktop automation requires seamless GUI and code integration, a feat previously unachievable with prompt-level tool composition, but OS-Helix now unifies these into a single action space.

The Milky Way's split-arm architecture is surprisingly typical: 90% of comparable multi-arm spiral galaxies share its exact transition from two inner arms to multiple outer branches, with bifurcations consistently clustering between 4 and 7 kpc.

RGB-only salient object detection can outperform traditional RGB-D methods by leveraging reliable geometry distillation without using depth data during inference.

Safety judgments for reasoning traces are far more complex than for final responses, revealing critical gaps in current guardrail models' capabilities.

Intern-S2-Preview-397B not only excels in multimodal scientific reasoning but also enhances biological instruction performance without altering its foundational architecture.

MDB-Link boosts Text-to-SQL accuracy by over 30 points while streamlining the schema selection process in multi-database scenarios.

SkillMemo transforms robotic manipulation by enabling models to leverage reusable skill structures, leading to unprecedented compositional generalization in complex tasks.

A unified evaluation framework that simplifies the assessment of LLM-based agents could drastically enhance reproducibility and accelerate research breakthroughs.

Action-only decoding in GigaWorld-Policy-0.5 slashes inference latency to 85 ms, revolutionizing real-time robot control efficiency.

Evaluator quality for robotic policies hinges more on long-horizon consistency than on short-term visual fidelity, reshaping our approach to world model design.

Structured allocentric representations can boost spatial reasoning in foundation models by up to 18%, even in the absence of visual inputs.

Decomposing GUI agent trajectories into verifiable milestones and auditing the evidence chain yields a 10% boost in RL training performance, outperforming single-judge reward systems.

Strategic recovery from failures is key to deploying robots for complex assembly tasks in the real world.

Humanoid robots can now reliably transport objects on a tray in the real world, thanks to a hierarchical RL approach that isolates and cancels gait-induced disturbances.