Market-Conditioned JEPA learns asset signals from realized market shocks
On 11 years of minute-level Chinese A-share data, it beat per-asset JEPA on returns, exposure, risk and liquidity.
Stories per day · 30 days
2 RSS
Relatedcomputer-vision2jepa23d-animation1attention1change-point-detection1depth-estimation1dino1encoders1feature-selection1financial-time-series1image-reconstruction1image-segmentation1lejepa1market-factors1masked-autoencoder1monocular-depth1neural-skinning1physics-simulation1
18 stories, newest first · one card per story
Stories tagged #self-supervised-learning
Paper trains Vision Transformers to specialize spatial attention heads
Two annotation-free focusing losses steer multi-head attention toward distinct image regions for denser predictions.
Label-free SSL signals miss known encoder decline
In nine of ten LeJEPA seeds, factor readability dropped while loss and most of eleven signals kept improving.
WildMAE extends masked autoencoders for wildlife camera-trap images
On 4,744 Snapshot Serengeti images at 75% masking, historical PSNR was 27.90 dB for WildMAE versus 24.33 for vanilla MAE.
Paper characterizes JEPA feature selection as predictive closure
Unpredictable features can still be encoded when they help predict other features.
PhysSkin++ learns real-time neural skinning without per-object training
A self-supervised autoencoder builds deformation subspaces from static geometry for meshes, clouds, and Gaussians.
MaSS makes self-supervised monocular depth scale across mixed scenes
A decoupled structure pathway lets mixed-domain training help, with zero-shot tests on eight more datasets.
Less Decoder is More Encoder: Geometric Representation Learning from Novel View Synthesis
SNAP, a self-supervised pose-conditioned transformer for novel view synthesis, learns transferable geometric representations competitive with geometry-supervised methods across five tasks.
On-Board Anomaly Detection for Efficient Marine Environmental Monitoring
Self-supervised encoder plus anomaly detection pipeline compresses hyperspectral satellite imagery onboard to flag oil spills and algal blooms.
Learning Commute-Time-Preserving World Models for Planning
CTWMs learn commute-time-preserving latents for planning and match LeWM with half the parameters.
Image Classifiers are Efficient Self-Supervised Video Representation Learners
VideoMSN turns pretrained image ViTs into efficient video learners, cutting pretraining by up to 160x.
A Width-Matched Comparison of Hybrid Quantum-Classical Self-Supervised Learning for Fingerprint Recognition
Hybrid quantum SSL beats matched classical models on fingerprints only for contrastive objectives, not reliably for BYOL.
I Have a Stream: Making Self-Supervised Learning Work on Continuous Video
StreamMAE adapts masked autoencoders so pretraining works on continuous video streams without global shuffling.
What masking geometry works best for EEG foundation models?
An EEG masking ablation matches REVE-level decoding performance at a fraction of the pre-training compute.
TT-VidT: Decoupling the Temporal Axis for Efficient Motion-Centric Video Pretraining
TT-VidT pairs a DINOv3 spatial ViT with a temporal layer, leading motion benchmarks at lower FLOPs.
BeatGraph: Self-Supervised Heartbeat Graphs for Infant ECG Representations from the Home Environment
BeatGraph learns infant ECG from heartbeat graphs and releases 3,408 hours of home recordings.
Tactile-JEPA: Topology-Aware Self-Supervised Representation Learning for Distributed Tactile Sensors
Tactile-JEPA pretrains topology-aware encoders for sparse robot tactile skins using graph-guided masking.
Do speech foundation models really learn words?
Researchers show via residualization that later layers of HuBERT and wav2vec 2.0 encode word identity and semantics independently of phonetic content.