Statistical Learning
Random Cuts are Optimal for Explainable k-Medians Konstantin Makarychev Northwestern University Liren Shan
The algorithm's aim is to cluster data sets and present results in a manner easily understood
a8808b75b299d64a23255bc8d30fb786-Paper-Conference.pdf
Can a physicist make only a finite number of errors in the eternal quest to uncover the law of nature? This millennium-old philosophical problem, known as inductive inference, lies at the heart of epistemology.
A new perspective on building efficient and expressive 3D equivariant graph neural networks
Geometric deep learning enables the encoding of physical symmetries in modeling 3D objects.
PAC Learning Linear Thresholds from Label Proportions
The goal is to train a good instance classifier.
Particle-based Variational Inference with Generalized Wasserstein Gradient Flow
We show that GWG exhibits strong convergence guarantees.
Feature Learning for Interpretable, Performant Decision Trees
Points were sampled uniformly in the bands denoted by dashed lines. We posit that these barriers are due, at least in part, to the sensitivity of decision trees to transformations of the input resulting from greedy construction and simple decision rules. Of these, key limitation is the latter; even if we replace greedy construction with a perfect tree learner, simple distributions can nonetheless require an arbitrarily large axis-aligned tree to fit.
Unraveling the Gradient Descent Dynamics of Transformers
By analyzing the loss landscape of a single Transformer layer using Softmax and Gaussian attention kernels, our work provides concrete answers to these questions.