Topic: machine learning

Latest on “machine learning”

The AI news that matters for people who make things — what happened, and how to put it to work in Grafy.

TabPFN Predicts Energetic Material Properties Without Tuning

A new study finds the TabPFN foundation model outperforms tuned XGBoost and other baselines at predicting energetic-material properties, with no tuning.

Try in Readertabpfnenergetic materialsmachine learningtabular foundation model

Grafy Research Labyesterday4 min read

Outliers Reshape Hard-Margin Linear SVM in High Dimensions

Grafy study found outliers can shift hard-margin linear SVM behavior in high dimensions and low samples, with thresholds for margin and support loss. Read more

Try in Readeroutlierssupport vector machine

Grafy Research Labyesterday4 min read

VC Dimension Found in Geometric AI

A new paper calculates VC-dimension for metric-ball classifiers in geometric machine learning, with bounds across Riemannian and normed spaces. Why it matters.

Try in Readervc-dimensionmachine learning

Reddit ML3 days ago4 min read

8B Model Outperforms LoRA in Continual Learning Experiments

A new experiment shows an 8B model retaining 90.7% accuracy on news data, significantly outperforming LoRA methods which suffer from catastrophic forgetting.

Try in Readercontinual learningmachine learning

Reddit ML3 days ago4 min read

New GAP architecture merges attention and MLP into one block

A researcher has introduced Gated Attention Projection (GAP), a Transformer alternative that reduces parameter counts by nearly 20% through architectural.

Try in Readertransformer architecturemachine learning