Neural networks efficiently encode learned information within their parameters. Consequently, many tasks can be unified by treating neural networks themselves as input data. When doing so, recent studies demonstrated the importance of accounting for the symmetries and geometry of parameter spaces. However, those works developed architectures tailored to specific networks such as MLPs and CNNs without normalization layers, and generalizing such architectures to other types of networks can be challenging.

Text-to-3D modeling has seen exciting progress by combining generative text-to-image models with image-to-3D methods like Neural Radiance Fields. DreamFusion recently achieved high-quality results but requires a lengthy, per-prompt optimization to create 3D objects. To address this, we amortize optimization over text prompts by training on many prompts simultaneously with a unified model, instead of separately. With this, we share computation across a prompt set, training in less time than per-prompt optimization.

Studies have shown that action video game players have enhanced visual abilities in various domains, such as multiple object tracking, size of the useful field of view, and visual search speed and accuracy. These improvements have been attributed to either a general advantage in “learning to learn” abilities, or domain-specific enhancement(s) in the “common demands” between specific games and experimental tasks. To investigate these two theories, we conducted six experiments examining whether and how players and non-players differ in various aspects of visual search.

Computer graphics research has long prioritized image quality over frame rate. Yet demand for an alternative is growing, with many esports players turning off visual effects to improve frame rates. Is it time for graphics researchers to reconsider their goals? A workshop at the 2023 SIGGRAPH Conference explored this question. Three researchers made provocative presentations, each of which were then discussed by dozens of research and industry attendees.

We present a Multimodal Interlaced Transformer (MIT) that jointly considers 2D and 3D data for weakly supervised point cloud segmentation. Research studies have shown that 2D and 3D features are complementary for point cloud segmentation. However, existing methods require extra 2D annotations to achieve 2D-3D information fusion. Considering the high annotation cost of point clouds, effective 2D and 3D feature fusion based on weakly supervised learning is in great demand.

Recent text-to-3D generation approaches produce impressive 3D results but require time-consuming optimization that can take up to an hour per prompt. Amortized methods like ATT3D optimize multiple prompts simultaneously to improve efficiency, enabling fast text-to-3D synthesis. However, ATT3D cannot capture high-frequency geometry and texture details and struggles to scale to large prompt sets, so it generalizes poorly. We introduce Latte3D, addressing these limitations to achieve fast, high-quality generation on a significantly larger prompt set.