SymTrack is the first systematic detection-free framework for scene text tracking that constructs benchmarks from video text spotting datasets and reports up to 11.97% AUC gains over prior trackers.
Title resolution pending
11 Pith papers cite this work. Polarity classification is still indexing.
11
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
verdicts
UNVERDICTED 11roles
background 1polarities
unclear 1representative citing papers
Bayesian softmax-gated mixture-of-experts models achieve posterior contraction for density estimation and parameter recovery using Voronoi losses, plus two strategies for choosing the number of experts.
MoE routing states at boundary and delimiter anchors form basins that align with final answers, enabling RAD, a string-free multi-rollout selector that matches majority voting on math and code tasks.
Task-aware expert grouping derived from family-specific co-activation traces cuts average communication cost 31.39% versus task-agnostic baselines in multi-task MoE inference while maintaining Jain fairness near 1.0.
RouteScan identifies malicious prompts in MoE LLMs using GPU expert routing telemetry as a privacy-preserving fingerprint, achieving AUROC above 0.93 on unseen harmful domains.
A new Ms-MoE-IFactFormer neural operator uses time-step routing and scale-specific experts to achieve stable fine-time-step long-horizon predictions of homogeneous isotropic turbulence and channel flow.
EMoE trains MoE models so they maintain performance when the number of activated experts changes at inference, expanding the usable range to 2-3 times the training k with higher peak results.
SafeMoE isolates unsafe knowledge in domain-specific LoRA experts and routes them via a lightweight gate trained on safe responses to produce safer and more informative LLM outputs with zero-shot generalization.
Token-Superposition Training combines multiple tokens into bags for multi-hot cross-entropy pre-training followed by a recovery phase, yielding up to 2.5x reduction in training time at 10B scale under equal-loss conditions.
A survey of LLMs for graph computation introduces a role-based taxonomy of executors versus planners and concludes that current models suit simple small-scale tasks but remain unreliable for large-scale exact computation.
The paper introduces a four-layer technical architecture for token-operations-oriented inference optimization in large models and reviews key technologies and industry status at each layer.
citing papers explorer
-
Elastic MoE: Unlocking the Inference-Time Scalability of Mixture-of-Experts
EMoE trains MoE models so they maintain performance when the number of activated experts changes at inference, expanding the usable range to 2-3 times the training k with higher peak results.