Transfers RL-induced policy shifts from a smaller 'weak' teacher to a stronger target by using the teacher's post-/pre-RL log-ratio as a dense implicit reward applied on the student's on-policy states. Enables reuse of RL supervision without running RL rollouts on the target, improving sample/time efficiency.
AI Video Papers2026
Provides a reflexive agentic framework for long-horizon video understanding that replaces costly iterative reasoning with dual contextual states: a consolidated global multimodal script and parametric latent states for fast retrieval and response, improving speed and memory efficiency.
Trains a single diffusion model that unifies 3D scene reconstruction and generative modeling by operating directly in pixel/rendered-image space. Supervises diffusion on rendered views and adds a geometry-perception loss from a pretrained 3D foundation model, reducing latent information loss and improving 3D fidelity.
AI Video Papers2026
Autoregressively synthesizes long-horizon, playable video worlds conditioned on current state and user actions for real-time interaction. Ships as an open-source, full-stack framework covering data preparation, model architectures, training, inference acceleration, and deployment for interactive generative worlds.
Expresses diverse computer-vision tasks as instruction-driven text, image, or mixed generation from a single unified multimodal model, producing outputs for detection, segmentation, depth, pose, OCR and more. Trained on a converted SenseNova‑Vision instruction–response corpus and requires no task-specific prediction heads.
Introduces KronQ, a post-training quantization framework that incorporates gradient covariance via a Kronecker‑factored Hessian to guide input/output weight rotations and sensitivity-driven mixed-precision allocation. Demonstrates stable 2-bit weight-only quantization on LLaMA-3-70B (7.93 PPL).
Performs native structural reasoning for proteins, small molecules and inorganic crystals by tokenizing coordinates, topologies and periodic connectivities into a unified structure-aware vocabulary. Treats structural tokens as addressable evidence to produce interpretable prediction traces and improves accuracy across biology, chemistry and materials benchmarks.
AI Video Papers2026
Pretrains a DiT-based Mixture-of-Experts video foundation model for embodied intelligence by augmenting internet videos with robot-centric footage and using a multi-dimensional reward system to prioritize physical realism and task completion while scaling MoE for better capacity vs. inference trade-offs.
Reconstructs historical experience into latent memory tokens and weaves short- and long-term latent memories directly into vision-language-action reasoning to improve long-horizon robotic manipulation. Uses a four-part pipeline (curator, seeker, condenser, weaver) so memory participates natively in multimodal action formation.
Creates an open-ended interactive world simulator with an unbounded interaction horizon via causal pretraining, a distilled real-time runtime that drives 720p@60fps, a wider action/event repertoire, and a pilot–director agent split for behavior planning and environment synthesis.
AI Video Papers2026
Recovers and predicts RGB video from sparse event-camera streams by fine-tuning pre-trained video diffusion priors; jointly addresses reconstruction, long-horizon prediction, and bidirectional frame interpolation with mechanisms to reduce temporal drift and enforce interpolation consistency.
Provides IdeaGene-Bench, a dataset and evaluation suite for scientific-lineage reasoning and lineage-grounded idea generation, representing papers as minimal, typed Idea Genome objects and GenomeDiffs that record inheritance, mutation, loss, import and novel insertion. Includes 1,961 lineage traces, IG-Exam (42 task types) and IG-Arena with a Population-Evolution Score for generation.