{"arxiv_id":"2608.18077","incoming":{"papers":0,"sentences":0,"groups":[],"unlabelled_sentences":0},"outgoing":{"papers":34,"sentences":61,"groups":[{"kind":"extends","label":"Builds on","papers":2,"sentences":2,"quotes":[{"arxiv_id":"2505.22944","title":"ATI: Any Trajectory Instruction for Controllable Video Generation","url":"/paper/2505.22944","section":"Introduction","sentence":"Building on trajectory-based visual conditioning in ATI and Wan-Move [ref] , our world model uses this condition to predict the command's consequences, a mode we call kinematically grounded video prediction.","intent_p":0.85,"contests_p":null},{"arxiv_id":"2512.08765","title":null,"url":"/paper/2512.08765","section":"Introduction","sentence":"Building on trajectory-based visual conditioning in ATI and Wan-Move [ref] , our world model uses this condition to predict the command's consequences, a mode we call kinematically grounded video prediction.","intent_p":0.85,"contests_p":null}]},{"kind":"method_or_tool","label":"Uses the method or tool of","papers":5,"sentences":7,"quotes":[{"arxiv_id":"2605.18739","title":"LongLive-2.0: An NVFP4 Parallel Infrastructure for Long Video Generation","url":"/paper/2605.18739","section":"Efficient Causal Rollout","sentence":"Following LongLive-2.0 [ref] , we initialize the distilled student from the causal autoregressive checkpoint and generate each chunk with four denoising steps through its causal cache while preserving the same absolute-offset motion slices.","intent_p":0.94,"contests_p":null},{"arxiv_id":"2311.18828","title":"One-step Diffusion with Distribution Matching Distillation","url":"/paper/2311.18828","section":"Efficient Causal Rollout","sentence":"For efficient robotics deployment, we combine autoregressive conversion for causal long-horizon generation with improved distribution-matching distillation (DMD2) [ref] for few-step sampling.","intent_p":1.0,"contests_p":null},{"arxiv_id":"2405.14867","title":"Improved Distribution Matching Distillation for Fast Image Synthesis","url":"/paper/2405.14867","section":"Efficient Causal Rollout","sentence":"For efficient robotics deployment, we combine autoregressive conversion for causal long-horizon generation with improved distribution-matching distillation (DMD2) [ref] for few-step sampling.","intent_p":1.0,"contests_p":null},{"arxiv_id":"2506.07310","title":"AllTracker: Efficient Dense Point Tracking at High Resolution","url":"/paper/2506.07310","section":"Constructing and Sampling Action Flow","sentence":"For these data, we first recover dense image-plane trajectories and visibility labels with a flow tracker [ref] , then segment the tracks using grounded masks for the visible embodiment and manipulated objects [ref] .","intent_p":1.0,"contests_p":null},{"arxiv_id":"2106.09685","title":"LoRA: Low-Rank Adaptation of Large Language Models","url":"/paper/2106.09685","section":"Action-Flow-Conditioned Video Prediction","sentence":"We freeze the pretrained video-backbone weights except for the DiT patch embedding and train rank-64 low-rank adaptation (LoRA) modules [ref] on the $q$, $k$, $v$, and $o$ attention projections and the two feed-forward projections.","intent_p":1.0,"contests_p":null}]},{"kind":"data_used","label":"Uses data or values from","papers":6,"sentences":7,"quotes":[{"arxiv_id":"2603.08546","title":null,"url":"/paper/2603.08546","section":"Data Efficiency from Multi-Embodiment Mid-Training","sentence":"We use six tasks from the Interactive World Simulator (IWS) dataset [ref] : Bimanual box, Bimanual rope, Bimanual sweep, Push-T, Single chain-in-box, and Single grasp.","intent_p":0.99,"contests_p":null},{"arxiv_id":"2402.10329","title":"Universal Manipulation Interface: In-The-Wild Robot Teaching Without In-The-Wild Robots","url":"/paper/2402.10329","section":"Datasets","sentence":"EgoDex [ref] provides egocentric human-hand manipulation captured with Apple Vision Pro, while Deform360 [ref] provides handheld-gripper demonstrations of deformable-object manipulation using UMI grippers [ref] .","intent_p":0.96,"contests_p":null},{"arxiv_id":"2505.11709","title":"EgoDex: Learning Dexterous Manipulation from Large-Scale Egocentric Video","url":"/paper/2505.11709","section":"Datasets","sentence":"EgoDex [ref] provides egocentric human-hand manipulation captured with Apple Vision Pro, while Deform360 [ref] provides handheld-gripper demonstrations of deformable-object manipulation using UMI grippers [ref] .","intent_p":0.96,"contests_p":null},{"arxiv_id":"2607.05390","title":"Deform360: A Massive Multi-view Visuotactile Dataset for Deformable World Models","url":"/paper/2607.05390","section":"Datasets","sentence":"EgoDex [ref] provides egocentric human-hand manipulation captured with Apple Vision Pro, while Deform360 [ref] provides handheld-gripper demonstrations of deformable-object manipulation using UMI grippers [ref] .","intent_p":0.96,"contests_p":null},{"arxiv_id":"2606.27375","title":"Scalable Behavior Cloning with Open Data, Training, and Evaluation","url":"/paper/2606.27375","section":"Datasets","sentence":"ABC-130k [ref] and MolmoAct2 [ref] provide bimanual teleoperation.","intent_p":0.98,"contests_p":null}]},{"kind":"background","label":"Cites as background","papers":13,"sentences":13,"quotes":[{"arxiv_id":"2511.11520","title":null,"url":"/paper/2511.11520","section":"Neural Simulators and Robotic World Models","sentence":"Action-conditioned video simulators have been used for policy evaluation and ranking, synthetic policy-improvement data, and interactive rollouts: [ref] and [ref] study policy ranking; [ref] and [ref] also use imagined trajectories for policy improvement or training; and Veo [ref] studies nominal and out-of-distribution policy behavior.","intent_p":0.84,"contests_p":null},{"arxiv_id":"2512.10675","title":null,"url":"/paper/2512.10675","section":"Neural Simulators and Robotic World Models","sentence":"Action-conditioned video simulators have been used for policy evaluation and ranking, synthetic policy-improvement data, and interactive rollouts: [ref] and [ref] study policy ranking; [ref] and [ref] also use imagined trajectories for policy improvement or training; and Veo [ref] studies nominal and out-of-distribution policy behavior.","intent_p":0.84,"contests_p":null},{"arxiv_id":"2603.08546","title":null,"url":"/paper/2603.08546","section":"Neural Simulators and Robotic World Models","sentence":"Action-conditioned video simulators have been used for policy evaluation and ranking, synthetic policy-improvement data, and interactive rollouts: [ref] and [ref] study policy ranking; [ref] and [ref] also use imagined trajectories for policy improvement or training; and Veo [ref] studies nominal and out-of-distribution policy behavior.","intent_p":0.84,"contests_p":null},{"arxiv_id":"2510.10125","title":"Ctrl-World: A Controllable Generative World Model for Robot Manipulation","url":"/paper/2510.10125","section":"Neural Simulators and Robotic World Models","sentence":"Action-conditioned video simulators have been used for policy evaluation and ranking, synthetic policy-improvement data, and interactive rollouts: [ref] and [ref] study policy ranking; [ref] and [ref] also use imagined trajectories for policy improvement or training; and Veo [ref] studies nominal and out-of-distribution policy behavior.","intent_p":0.84,"contests_p":null},{"arxiv_id":"2412.08261","title":"FLIP: Flow-Centric Generative Planning as General-Purpose Manipulation World Model","url":"/paper/2412.08261","section":"Motion-Conditioned Video Generation","sentence":"Neighboring geometric and flow interfaces include kinematic pointmap conditioning, optical-flow world-action models, robot rendering, flow-centric generative planning, trajectory-conditioned cross-embodiment skill transfer, and contact-flow conditioning [ref] .","intent_p":0.95,"contests_p":null}]}],"unlabelled_sentences":32},"claims":[],"thresholds":{"intent_p":0.8,"supports_p":0.95,"contests_p":0.7},"method":"Citing sentences come from each paper's LaTeX source; every \\cite marks the sentence and the works it cites. Pith's classifier typed each sentence. A label is shown at intent probability 0.8 or more (0.95 for support), a contest at 0.7 or more. Measured accuracy against two blind model families is at https://pith.science/accuracy. The quoted sentences are the citing authors' own.","held_up":null,"_links":{"html":"https://pith.science/paper/2608.18077"}}