Search papers, labs, and topics across Lattice.
8
0
13
7
A single forecasting backbone unifies multivariate and multimodal time series, outperforming 16 SOTA models by disentangling media identity and mastering temporal scales.
Pretrained visual encoders are a temporary crutch: encoder-free multimodal architectures close the performance gap with encoder-based systems at roughly $10^{22}$ FLOPs, well within current frontier pretraining budgets.
Achieving 94.52% of state-of-the-art performance with just 25% of the training data reveals a groundbreaking efficiency in visual instruction tuning.
DiffPDE achieves faster and more accurate PDE code repair by focusing only on localized errors, challenging the inefficiencies of existing autoregressive methods.
Depth integration in audio-visual segmentation leads to over 10% performance gains, revealing a critical yet overlooked modality in multimodal perception.
Visual-Seeker outperforms proprietary models by actively engaging with visual details, redefining multimodal search capabilities.
Jointly training MTP and RL doesn't have to hurt: a simple coefficient calibration scheme unlocks performance gains on mathematical reasoning tasks.
Stop using VLMs just for answer generation: WikiSeeker shows how to unlock their potential in KB-VQA by using them as specialized agents for query refinement and retrieval inspection.
