Search papers, labs, and topics across Lattice.
8
0
12
7
CheckRLM cuts error accumulation in reasoning chains by correcting factual inaccuracies in real-time, outperforming traditional approaches.
By harnessing implicit supervision from environment dynamics, EnvRL boosts RL success rates by over 4% on long-horizon tasks, revealing a new frontier in agentic learning.
UniSVQ achieves state-of-the-art performance in 2-bit quantization, outperforming traditional methods while enhancing inference speed.
MemoryCard transforms long videos into coherent, topic-focused segments, boosting long-video QA accuracy by over 21% while maintaining visual-token efficiency.
Crafter revolutionizes scientific figure generation by enabling multi-type outputs and local editability, outperforming existing systems across diverse benchmarks.
Multi-source visual reasoning can actually *hurt* performance when modalities conflict, but MARS solves this by adaptively emphasizing mutual promotion and suppressing noise, leading to significant gains.
Forget separate structure and fidelity models – Khala shows you can generate high-quality music with text-vocal alignment using a single acoustic-token hierarchy.
Forget full attention: a hybrid sparse-linear attention model, MiniCPM-SALA, achieves 3.5x faster inference and supports 1M context length on a single GPU, all while maintaining comparable performance.
