Joint-Embedding Predictive Architectures
PulseAugur coverage of Joint-Embedding Predictive Architectures — every cluster mentioning Joint-Embedding Predictive Architectures across labs, papers, and developer communities, ranked by signal.
4 day(s) with sentiment data
-
New Med-D-JEPA model advances 3D brain MRI synthesis and classification
Researchers have adapted Joint-Embedding Predictive Architectures (JEPAs), a framework primarily used for self-supervised representation learning, for 3D brain MRI synthesis. The proposed Med-D-JEPA model combines maske…
-
Orthogonal JEPA framework enhances latent world models with factorized states
Researchers have introduced Orthogonal JEPA, a novel latent world-modeling framework designed to improve prediction and reasoning capabilities. This method addresses limitations in standard Joint-Embedding Predictive Ar…
-
New JEPA method uses contrastive inverse dynamics to improve world models
Researchers have developed a new method called Action-Contrastive Masked Transition Modeling (AC-MTM) for Joint-Embedding Predictive Architectures (JEPAs) that addresses the issue of trivial solutions in world models. U…
-
New diagnostic tool assesses AI world models for visual perturbation resilience
Researchers have developed a new diagnostic tool called Action-Conditioned Predictive Consistency (ACPC) to evaluate world models within Joint-embedding predictive architectures (JEPAs). ACPC measures how much a world m…
-
UniJEPA unifies image and video visual world modeling
Researchers have introduced UniJEPA, a novel unified architecture for self-supervised visual world modeling. This new framework integrates both image-level photometric prediction and video-level temporal prediction into…
-
JEPA models face challenges with language's conditional structure
A new paper explores the challenges of applying Joint-Embedding Predictive Architectures (JEPAs) to language processing, contrasting their effectiveness in image and audio domains with their limitations in text. The res…
-
Temporal-Distance JEPA enhances world model predictive control
Researchers have introduced Temporal-Distance JEPA (TD-JEPA), a novel approach to representation learning for latent world model predictive control. This method enhances Joint-Embedding Predictive Architectures (JEPAs) …
-
New regularization method tackles bias in self-supervised learning
Researchers have introduced Unbiased Open World Regularization (UOWReg), a novel framework designed to mitigate biases in self-supervised learning (SSL) and Joint-Embedding Predictive Architectures (JEPAs). Unlike previ…
-
Yann LeCun proposes JEPA to enhance LLMs' physical world understanding
Yann LeCun, in a recent interview, discussed the limitations of current Large Language Models (LLMs) in truly understanding the physical world, contrasting their ability to answer questions with their inability to perfo…
-
MorphologyFM model learns from ECG and pulse oximetry waveforms
Researchers have developed MorphologyFM, a novel foundation model designed to learn representations from electrocardiogram (ECG) and pulse oximetry (SpO2) waveforms. Unlike previous methods that focus on reconstruction …
-
Qantara JEPA model enables multi-paradigm control from single checkpoint
Researchers have introduced Qantara, a novel Joint-Embedding Predictive Architecture (JEPA) that enables a single model checkpoint to support multiple inference paradigms for control from raw pixels. Unlike previous JEP…
-
Fast LeWorldModel accelerates visual planning with parallel prediction
Researchers have developed Fast LeWorldModel (Fast-LeWM), an advancement over existing Joint-Embedding Predictive Architectures (JEPAs) like LeWorldModel (LeWM) for visual planning. Unlike LeWM's computationally intensi…
-
New DCGWM Architecture Prevents Objective Interference Collapse in World Models
Researchers have introduced Dual-Channel Grounded World Modeling (DCGWM), a novel architecture designed to prevent Objective Interference Collapse (OIC) in Joint Embedding Predictive Architectures (JEPAs). OIC occurs wh…
-
New RePAIR architecture learns chess concepts via self-supervised learning
Researchers have developed a new self-supervised learning architecture called RePAIR, which combines elements of MAE, JEPA, and BERT. This architecture is designed to encode sequential data, such as chess positions, int…
-
New Architecture Achieves Near-Infinite Temporal Consistency in World Models
A new research paper introduces the Physics-Grounded Symbolic Architecture (PGSA), which overcomes limitations in current statistical World Models. Unlike existing models that require Gaussian dynamics for linear identi…
-
New AI models advance self-supervised learning for 3D medical imaging
Two new research papers explore advanced self-supervised learning techniques for 3D medical imaging. One paper introduces a framework using Masked Autoencoders (MAE) and Joint Embedding Predictive Architectures (JEPA) t…
-
New JEPA Architectures Achieve Stable End-to-End Training from Pixels
Researchers have developed LeWorldModel (LeWM), a novel Joint Embedding Predictive Architecture (JEPA) that stably trains end-to-end from raw pixels. Unlike previous fragile JEPA methods, LeWM uses only two loss terms a…
-
New JEPA Model Learns Sparse Representations with Rectified Distribution Matching
Researchers have developed Rectified LpJEPA, a novel approach to Joint-Embedding Predictive Architectures (JEPA) that aims to create more efficient and sparse representations. Unlike previous methods that favored dense …
-
Yann LeCun dismisses LLMs as path to AGI, champions JEPA
Yann LeCun argues that current Large Language Models (LLMs) are not on a path to human-level intelligence because they lack the ability to predict consequences or perform search-based reasoning. He advocates for his Joi…