in_context | Awesome RoboticsRecent interesting robotics papershttps://broadcastchannel-6vg.pages.devInstant Policy: In-Context Imitation Learning via Graph Diffusion- formulate in-context imitation learning as a graph generation problem- demonstrate impressive one-shot learning performance#ICLR2025 #manipulation #in_contextX Overview, Project Websitehttps://broadcastchannel-6vg.pages.dev/posts/21https://broadcastchannel-6vg.pages.dev/posts/21Thu, 01 May 2025 22:44:25 GMTInstant Policy: <mark>In</mark>-<mark>Context</mark> Imitation Learning via Graph Diffusion<br /><br />- formulate <mark>in</mark>-<mark>context</mark> imitation learning as a graph generation problem<br />- demonstrate impressive one-shot learning performance<br /><br /><a href="/search/%23ICLR2025">#ICLR2025</a> <a href="/search/%23manipulation">#manipulation</a> <a href="/search/%23in_context">#in_context</a><br /><br /><a href="https://fixupx.com/Ed__Johns/status/1884234163458564585" target="_blank">X Overview</a>, <a href="https://www.robot-learning.uk/instant-policy" target="_blank">Project Website</a><a href="https://fixupx.com/Ed__Johns/status/1884234163458564585" target="_blank"> <div><i><b>🧵</b></i> Thread • FixupX</div> <div> <i></i> <div> <video src="https://cdn4.telesco.pe/file/7943225c12.mp4?token=GhV6hMWDgB03X7dK2V_yKP5FTZ8izZOcx1VbO1ekvCBCwKsUe4Z1YTU0TmTc1a-c4d7PCdesGEynKZDIPibZZZQJK26cW5dHyLzvibFkEh_muwJp8K_23O_lv0kKDFjMVNtv5BwE79kwVj7Azd5Uc5YNdmELNS16zWvnUytuPIuc1rYEAvB84ONzSGFKprqOKnIUTSSWgNf56T5VFevW2GzWZYACqDMAooGX27HeH3pZjG8NXSgVSdMSM3I0G1nn8OtXxcfos99LTaqvrxsygZfIsZ4pHRmOk2AFSp6SYymIX0Jd4TEtfoqkp-ouJh1ogqf9LA89kj7VITakm4Xpow" width="100%" height="100%"></video> </div> </div> <div>Edward Johns (@Ed__Johns)</div> <div>Happy to announce that "Instant Policy" was accepted at ICLR 2025!<br /><br />Below is a single uncut video: the robot is learning several novel tasks instantly, after just one demonstration each ...<br /><br />We've achieved <mark>in</mark>-<mark>context</mark> learning <mark>in</mark> robotics!<br /><br />http://www.robot…</div> </a>Vision Language Models are In-Context Value Learners- formulating value learning as an autoregressive prediction task over shuffled sequence of the input video- suitable across different tasks and in-context learning#ICLR2025 #manipulation #in_contextX Overview, Project Websitehttps://broadcastchannel-6vg.pages.dev/posts/20https://broadcastchannel-6vg.pages.dev/posts/20Thu, 01 May 2025 22:30:13 GMTVision Language Models are <mark>In</mark>-<mark>Context</mark> Value Learners<br /><br />- formulating value learning as <b>an autoregressive prediction task over shuffled sequence of the input video</b><br />- suitable across different tasks and <mark>in</mark>-<mark>context</mark> learning<br /><br /><a href="/search/%23ICLR2025">#ICLR2025</a> <a href="/search/%23manipulation">#manipulation</a> <a href="/search/%23in_context">#in_context</a><br /><br /><a href="https://fixupx.com/JasonMa2020/status/1854616567603986437" target="_blank">X Overview</a>, <a href="https://generative-value-learning.github.io/" target="_blank">Project Website</a><a href="https://fixupx.com/JasonMa2020/status/1854616567603986437" target="_blank"> <div><i><b>🧵</b></i> Thread • FixupX</div> <div> <i></i> <div> <video src="https://cdn4.telesco.pe/file/d8865ccff3.mp4?token=JypneC4Iud_Rnk0HZ3H9o4yJ0aEpTeLhqERO32BzeWcAURI-nzFZ8uDWI3tMXlEnp8Is_yB1X3TVIMYTcxPL4GKbikTmyQkD7ne9gfFBpt6zHZ5B13X3GPHFnFYac31ZOhxeExGNGF8L1j5URVxrKnOI4UtrT_Agm99sSsUb7ivRaK89E6pkRF0-7Br0rE0ztcyGDtVy7kmLSHaDrinDE0U42lZHUGHM8r0R9zOFQ6I8G_hBQ5Md11Ty0hFNTRe1M4iH6BPnKChYzg7cNqtKFT1oufD0y4SPAoW14H5gxwMN1PEsl0YYmM02VOyttSfvV90G186Vyeen9LxSPRN4vA" width="100%" height="100%"></video> </div> </div> <div>Jason Ma (@JasonMa2020)</div> <div>Excited to finally share Generative Value Learning (GVL), my @GoogleDeepMind project on extracting universal value functions from long-<mark>context</mark> VLMs via <mark>in</mark>-<mark>context</mark> learning!<br /><br />We discovered a simple method to generate zero-shot and few-shot values for 300+…</div> </a>