Search papers, labs, and topics across Lattice.
8
0
10
Key contribution not extracted.
Memory management in video generation just got a makeover鈥攑rioritizing novelty over recency leads to better visual coherence and fidelity.
Native computer use can be achieved at scale, enabling agents to outperform leading systems while significantly enhancing security against adversarial attacks.
Qwen-AgentWorld achieves unprecedented simulation fidelity, outperforming existing models and enabling scalable agentic reinforcement learning across diverse real-world environments.
Adapter design can make or break coding performance in OpenClaw-style agents, with a full adapter boosting success rates by over 50 percentage points.
Don't let valuable steps in failed trajectories go unnoticed: GraphGPO leverages state-transition graphs for fine-grained credit assignment in agentic RL, boosting performance and efficiency.
Cycle-consistent learning unlocks self-improvement in vision-language models, enabling them to reason about their own generations and boosting performance across understanding and generation tasks.
A new family of GUI agents, GUI-Owl-1.5, leapfrogs existing open-source models on 20+ GUI benchmarks, proving that multi-platform, real-time GUI automation is now within reach.
