Continuous audio encoders compress audio along feature and time axes through latent width and frame rate, but their joint effect on downstream performance remains unclear. We train sixteen encoders spanning four widths and four frame rates, with downstream adapters and probes, using matched training protocols. Despite...
Kyudan Jung, Sehyun Lee, Song-ha Jo et al.· 0 citations
3D Gaussian Splatting (3DGS) represents scenes with anisotropic Gaussian primitives optimized from multi-view images and rendered through differentiable rasterization and alpha compositing. Conventional 3DGS training primarily constrains camera-visible surfaces, whereas occluded interior regions remain insufficiently s...
Vision-language-action models let robots follow language instructions, but their language backbones of several billion parameters are the main obstacle to running them on robot hardware, and removing 63% of it from OpenVLA-OFT drops LIBERO-Long success from 93.2% to 0.8%.
LogitLens analyses and a targeted LM head intervention support the conclusion that acoustic underuse is not explained solely by encoder-side information loss and that readout alignment can be a dominant bottleneck.
Song-ha Jo, Sehyun Lee, Soyoon Kim et al.· 0 citations
Transfer learning with ImageNet-pretrained backbones is ubiquitous in computer vision, yet the question of which intermediate layer best serves a downstream task is typically resolved by convention rather than controlled measurement. Despite the prevalence of this practice, no systematic comparison of layer-wise frozen...
Sanghyuk Roy Choi, Chanhoe Gu, S. Kim et al.· International Conference on...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
