Cover Image

Is the reconstruction loss culprit? An attempt to outperform JEPA

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Potapov, Alexey, Shcherbakov, Oleg, Kravchenko, Ivan
Format:	Preprint
Published:	2026
Subjects:	Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

On (Normalised) Discounted Cumulative Gain as an Off-Policy Evaluation Metric for Top-$n$ Recommendation
by: Jeunen, Olivier, et al.
Published: (2023)

SPHERE-JEPA: Spherical Prediction with Homogeneous Embeddings
by: Nicollier, Léo, et al.
Published: (2026)

Semantic Tube Prediction: Beating LLM Data Efficiency with JEPA
by: Huang, Hai, et al.
Published: (2026)

T-JEPA: Augmentation-Free Self-Supervised Learning for Tabular Data
by: Thimonier, Hugo, et al.
Published: (2024)

Fine-tuning on simulated data outperforms prompting for agent tone of voice
by: Marquardt, Ingo, et al.
Published: (2025)

Knowledge, Rules and Their Embeddings: Two Paths towards Neuro-Symbolic JEPA
by: Huang, Yongchao, et al.
Published: (2026)

TD-JEPA: Latent-predictive Representations for Zero-Shot Reinforcement Learning
by: Bagatella, Marco, et al.
Published: (2025)

Error whitening: Why Gauss-Newton outperforms Newton
by: McKay, Maricela Best, et al.
Published: (2026)

Why and How Auxiliary Tasks Improve JEPA Representations
by: Yu, Jiacan, et al.
Published: (2025)

Sanity Checks for Sparse Autoencoders: Do SAEs Beat Random Baselines?
by: Korznikov, Anton, et al.
Published: (2026)

Crys-JEPA: Accelerating Crystal Discovery via Embedding Screening and Generative Refinement
by: Liu, Nian, et al.
Published: (2026)

OrtSAE: Orthogonal Sparse Autoencoders Uncover Atomic Features
by: Korznikov, Anton, et al.
Published: (2025)

Label-Efficient Grasp Joint Prediction with Point-JEPA
by: Guzelkabaagac, Jed, et al.
Published: (2025)

Value-guided action planning with JEPA world models
by: Destrade, Matthieu, et al.
Published: (2025)

TREB: a BERT attempt for imputing tabular data imputation
by: Wang, Shuyue, et al.
Published: (2024)

BiJEPA: Bi-directional Joint Embedding Predictive Architecture for Symmetric Representation Learning
by: Huang, Yongchao
Published: (2026)

Representation Without Reward: A JEPA Audit for LLM Fine-Tuning
by: Sengupta, Biswa
Published: (2026)

Mini-JEPA Foundation Model Fleet Enables Agentic Hydrologic Intelligence
by: Rahman, Mashrekur
Published: (2026)

T-JEPA: A Joint-Embedding Predictive Architecture for Trajectory Similarity Computation
by: Li, Lihuan, et al.
Published: (2024)

SparseJEPA: Sparse Representation Learning of Joint Embedding Predictive Architectures
by: Hartman, Max, et al.
Published: (2025)

MTS-JEPA: Multi-Resolution Joint-Embedding Predictive Architecture for Time-Series Anomaly Prediction
by: He, Yanan, et al.
Published: (2026)

AeroJEPA: Learning Semantic Latent Representations for Scalable 3D Aerodynamic Field Modeling
by: Giral, Francisco, et al.
Published: (2026)

How JEPA Avoids Noisy Features: The Implicit Bias of Deep Linear Self Distillation Networks
by: Littwin, Etai, et al.
Published: (2024)

Pretrained deep models outperform GBDTs in Learning-To-Rank under label scarcity
by: Hou, Charlie, et al.
Published: (2023)

Cropping outperforms dropout as an augmentation strategy for self-supervised training of text embeddings
by: González-Márquez, Rita, et al.
Published: (2025)

Sub-JEPA: Subspace Gaussian Regularization for Stable End-to-End World Models
by: Zhao, Kai, et al.
Published: (2026)

GLiRA: Black-Box Membership Inference Attack via Knowledge Distillation
by: Galichin, Andrey V., et al.
Published: (2024)

S-JEPA: towards seamless cross-dataset transfer through dynamic spatial attention
by: Guetschel, Pierre, et al.
Published: (2024)

CSI-JEPA: Towards Foundation Representations for Ubiquitous Sensing with Minimal Supervision
by: Luo, Xuanhao, et al.
Published: (2026)

BERT-JEPA: Reorganizing CLS Embeddings for Language-Invariant Semantics
by: Gillin, Taj, et al.
Published: (2026)

Text-Conditional JEPA for Learning Semantically Rich Visual Representations
by: Huang, Chen, et al.
Published: (2026)

Rethinking JEPA: Compute-Efficient Video SSL with Frozen Teachers
by: Li, Xianhang, et al.
Published: (2025)

KerJEPA: Kernel Discrepancies for Euclidean Self-Supervised Learning
by: Zimmermann, Eric, et al.
Published: (2025)

Deep learning outperforms traditional machine learning methods in predicting childhood malnutrition: evidence from survey data
by: Bastola, Deepak, et al.
Published: (2026)

Laya: A LeJEPA Approach to EEG via Latent Prediction over Reconstruction
by: Panchavati, Saarang, et al.
Published: (2026)

Parameter-free representations outperform single-cell foundation models on downstream benchmarks
by: Souza, Huan, et al.
Published: (2026)

ACT-JEPA: Novel Joint-Embedding Predictive Architecture for Efficient Policy Representation Learning
by: Vujinovic, Aleksandar, et al.
Published: (2025)

BRo-JEPA: Learning Modular Arithmetic in Latent Space
by: Jha, Divyansh, et al.
Published: (2026)

Steering LLMs? Actually, Sparse Autoencoders can outperform simple baselines
by: Jørgensen, Mikkel Godsk, et al.
Published: (2026)

The Rogue Scalpel: Activation Steering Compromises LLM Safety
by: Korznikov, Anton, et al.
Published: (2025)