UNIEGO creates a unified egocentric encoder by using proxy models to mediate multi-teacher distillation from ego-exo, multi-modal, and foundation-model sources, with selective distillation and convex initialization, achieving SOTA on action recognition, retrieval, and segmentation.
Theia: Distilling diverse vision foundation models for robot learning
11 Pith papers cite this work. Polarity classification is still indexing.
11
Pith papers citing it
citation-role summary
background 2
citation-polarity summary
roles
background 2polarities
background 2representative citing papers
AEM pretrains compact history representations via masked modeling on interleaved vision-action sequences to boost downstream robot manipulation in simulation and real settings.
PRISM is a two-stage MoE framework that achieves new state-of-the-art results on PASCAL-Context and NYUD-v2 by enabling self-organized expert specialization across diverse vision foundation models.
SceneDiver introduces a coarse-to-fine focus plan generation approach for VLMs that constructs holistic scene graphs then iteratively decomposes tasks, plus a distillation adapter for VLAs, to reduce visual hallucinations in embodied AI benchmarks.
LACE aligns human-robot visual features via semantic distribution matching on corresponding body parts plus Gram loss, yielding 65% better zero-shot policy transfer than baseline DINO.
VECA learns effective visual representations using core-periphery attention where patches interact exclusively via a resolution-invariant set of learned core embeddings, achieving linear O(N) complexity while maintaining competitive performance.
ExpertGen generates high-success expert policies in simulation from imperfect priors by freezing a diffusion behavior model and optimizing its initial noise via RL, then distills them for real-robot deployment.
SigLino distills SigLIP2 and DINOv3 into efficient vision models via asymmetric relation-knowledge distillation, token-balanced batching, and hierarchical data sampling on a new 200M-image corpus, yielding better transfer to grounding VLMs than training from scratch.
RADSeg adapts the RADIO model with targeted enhancements to deliver 6-30% higher mIoU in zero-shot OVSS while using 2.5x fewer parameters and running 3.95x faster than prior large-model combinations.
DreamVLA uses dynamic-region-guided world knowledge prediction, block-wise attention to disentangle information types, and a diffusion transformer for actions, reaching 76.7% success on real robot tasks and 4.44 average length on CALVIN ABC-D.
A label-propagation pipeline combining a segment proposer with Hopfield networks on multi-model embeddings can automatically annotate 60% of household object data for up to 50 classes using only limited initial labels.
citing papers explorer
-
UNIEGO: Proxies as Mediators for Unified Egocentric Video Representation Learning
UNIEGO creates a unified egocentric encoder by using proxy models to mediate multi-teacher distillation from ego-exo, multi-modal, and foundation-model sources, with selective distillation and convex initialization, achieving SOTA on action recognition, retrieval, and segmentation.
-
Action-Effect Memory Pretraining for Robot Manipulation
AEM pretrains compact history representations via masked modeling on interleaved vision-action sequences to boost downstream robot manipulation in simulation and real settings.
-
PRISM: Synergizing Vision Foundation Models via Self-organized Expert Specialization
PRISM is a two-stage MoE framework that achieves new state-of-the-art results on PASCAL-Context and NYUD-v2 by enabling self-organized expert specialization across diverse vision foundation models.
-
Dive into the Scene: Breaking the Perceptual Bottleneck in Vision-Language Decision Making via Focus Plan Generation
SceneDiver introduces a coarse-to-fine focus plan generation approach for VLMs that constructs holistic scene graphs then iteratively decomposes tasks, plus a distillation adapter for VLAs, to reduce visual hallucinations in embodied AI benchmarks.
-
LACE: Latent Visual Representation for Cross-Embodiment Learning
LACE aligns human-robot visual features via semantic distribution matching on corresponding body parts plus Gram loss, yielding 65% better zero-shot policy transfer than baseline DINO.
-
Elastic Attention Cores for Scalable Vision Transformers
VECA learns effective visual representations using core-periphery attention where patches interact exclusively via a resolution-invariant set of learned core embeddings, achieving linear O(N) complexity while maintaining competitive performance.
-
RADSeg: Unleashing Parameter and Compute Efficient Zero-Shot Open-Vocabulary Segmentation Using Agglomerative Models
RADSeg adapts the RADIO model with targeted enhancements to deliver 6-30% higher mIoU in zero-shot OVSS while using 2.5x fewer parameters and running 3.95x faster than prior large-model combinations.
-
DreamVLA: A Vision-Language-Action Model Dreamed with Comprehensive World Knowledge
DreamVLA uses dynamic-region-guided world knowledge prediction, block-wise attention to disentangle information types, and a diffusion transformer for actions, reaching 76.7% success on real robot tasks and 4.44 average length on CALVIN ABC-D.
-
Efficient Image Annotation via Semi-Supervised Object Segmentation with Label Propagation
A label-propagation pipeline combining a segment proposer with Hopfield networks on multi-model embeddings can automatically annotate 60% of household object data for up to 50 classes using only limited initial labels.