Search papers, labs, and topics across Lattice.
Graduate School, Tsinghua Shenzhen International, Tsinghua University, Peng Cheng Laboratory
Tsinghua AI7
0
10
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.
MemVenom reveals that web agents can be compromised with up to 99.15% success through sophisticated memory poisoning attacks that bypass traditional defenses.
Current image restoration models still fail to strike the right balance between noise reduction, detail fidelity, and accurate color in real-world, low-light portrait scenarios, highlighting a critical gap this challenge aims to close.
Current Chinese AI-generated text detection benchmarks are too homogeneous; C-ReD fixes this with real-world prompts and diverse LLMs, enabling better generalization.
By optimizing gradient inversions across hierarchical GAN feature spaces, GIFD achieves pixel-perfect reconstructions of private data in federated learning, even in challenging out-of-distribution scenarios.
Spatial awareness is the secret ingredient to unlocking better visual in-context learning, boosting performance across diverse vision tasks.
Achieve state-of-the-art long-horizon video understanding by compressing multimodal memories into high-level semantic schemas, enabling efficient reasoning without losing crucial details.
