メニュー
Enable spatially adaptive Gaussian allocation in 3D Gaussian Splatting without iterative optimization. Reduces primitive count by 70-90% while maintaining reconstruction quality through learned densification score prediction.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Convert arXiv papers into ready-to-use agent skills using category-aware extraction. First classifies the paper into one or more of 11 research categories, then applies a specialized extraction pipeline for each category — because different types of papers produce different types of usable knowledge. A single paper can yield multiple skills if it spans categories. Use this skill whenever the user wants to turn a paper into a skill, extract practical techniques from research, build a skill library from papers, convert arXiv papers into reusable agent instructions, or batch-process multiple papers into skills. Also trigger when someone asks about extracting actionable knowledge from papers, making research practical for LLM agents, or systematically converting academic contributions into structured agent capabilities.
Establish regret bounds for behavior cloning with discretized actions combining statistical error and quantization error terms. Prove smoothness requirements for safe quantizer design, show that learning-based quantizers fail these requirements, and propose model-based augmentation to reduce error dependence from H² to H.
Dynamically allocate LoRA ranks per-layer during fine-tuning instead of using fixed uniform ranks. Learn optimal rank for each layer and subject via variational framework with discretized exponential distribution, reducing memory footprint while maintaining fidelity and text-alignment.
Adapt general LLMs to specialized manufacturing domains via domain-adaptive pretraining on open-access journals and visual instruction tuning. Extract 50M tokens and 24K images from peer-reviewed papers, achieve >90% accuracy on domain knowledge tasks, and enable real-time defect identification from manufacturing images.
Future intelligence explosions will be plural, social, and entangled with humanity through distributed collaborative systems rather than singular superintelligence. Intelligence is inherently social, demanding infrastructure matching agent development; integrate governance, institutional frameworks, and constitutional checks across hierarchies of autonomous agents and human-AI centaurs in shifting configurations.
Build taxonomy-aware audio-text pretraining systems for species recognition from animal vocalizations. Train contrastive models that augment text prompts with hierarchical taxonomic structure (scientific/common names, phylogenetic sequences), evaluate on unseen species via rare-species test sets, and predict ecological traits directly from audio.
| name | f4splat-feed-forward-densification |
| title | F4Splat: Feed-Forward Predictive Densification for 3D Gaussian Splatting |
| version | 0.0.3 |
| engine | skillxiv-v0.0.3-claude-opus-4.6 |
| license | MIT |
| url | https://arxiv.org/abs/2603.21304 |
| keywords | ["3DGS","Gaussian Splatting","Densification","Feed-Forward","Adaptive Allocation"] |
| description | Enable spatially adaptive Gaussian allocation in 3D Gaussian Splatting without iterative optimization. Reduces primitive count by 70-90% while maintaining reconstruction quality through learned densification score prediction. |
Predictive Densification Module for Feed-Forward 3DGS
Traditional feed-forward 3D Gaussian Splatting uniformly distributes Gaussians across pixels or voxels, creating redundancy in simple regions and insufficient representation in complex areas. Each scene requires many primitives to maintain quality because allocation is not adaptive.
F4Splat replaces fixed pixel-to-Gaussian allocation with a learned densification score predictor that enables spatially adaptive allocation at inference time:
Densification Score Prediction Network - During training, the model observes per-region homodirectional view-space positional gradients (which indicate where additional Gaussians would help during optimization). These supervise a learned densification score (computed as log-scaled L2 norm of gradients). At inference, the score is available without per-scene optimization.
Multi-Scale Gaussian Maps - Instead of single-resolution allocation, the architecture predicts multi-scale parameter maps across three pyramid levels, enabling flexible representation levels per spatial region.
Adaptive Thresholding Mechanism - A region-wise selector chooses representation levels for each spatial location based on predicted densification scores, ensuring non-overlapping allocations across pyramid levels.
The paper demonstrates:
