Manipulation on skulldchttps://skulldc.net/tags/manipulation/Recent content in Manipulation on skulldcHugoen-usThu, 05 Nov 2026 10:25:39 +0000NEO-VLA drops the vision encoder so action policies can read visual features at any depthhttps://skulldc.net/2026/11/neo-vla-drops-the-vision-encoder-so-action-policies-can-read-visual-fe/Thu, 05 Nov 2026 10:25:39 +0000https://skulldc.net/2026/11/neo-vla-drops-the-vision-encoder-so-action-policies-can-read-visual-fe/A paper from MMLab@NTU argues that vision-language-action models lose action-critical detail in their visual encoders and proposes NEO-VLA, an encoder-free design.