Abstract:Many high-dimensional practical data sets have hierarchical structures induced by graphs or time series. Such data sets are hard to process in Euclidean spaces and one often seeks low-dimensional embeddings in other space forms to perform the required learning tasks. For hierarchical data, the space of choice is a hyperbolic space because it guarantees low-distortion embeddings for tree-like structures. The geometry of hyperbolic spaces has properties not encountered in Euclidean spaces that pose challenges when trying to rigorously analyze algorithmic solutions. We propose a unified framework for learning scalable and simple hyperbolic linear classifiers with provable performance guarantees. The gist of our approach is to focus on Poincaré ball models and formulate the classification problems using tangent space formalisms. Our results include a new hyperbolic perceptron algorithm as well as an efficient and highly accurate convex optimization setup for hyperbolic support vector machine classifiers. Furthermore, we adapt our approach to accommodate second-order perceptrons, where data is preprocessed based on second-order information (correlation) to accelerate convergence, and strategic perceptrons, where potentially manipulated data arrives in an online manner and decisions are made sequentially. The excellent performance of the Poincaré second-order and strategic perceptrons shows that the proposed framework can be extended to general machine learning problems in hyperbolic spaces. Our experimental results, pertaining to synthetic, single-cell RNA-seq expression measurements, CIFAR10, Fashion-MNIST and mini-ImageNet, establish that all algorithms provably converge and have complexity comparable to those of their Euclidean counterparts. Accompanying codes can be found at: <a class="link-external link-https" href="https://github.com/thupchnsky/PoincareLinearClassification" rel="external noopener nofollow">this https URL</a>.

Embeddings and near-neighbor searching with constant additive error for hyperbolic spaces

A Quadtree, a Steiner Spanner, and Approximate Nearest Neighbours in Hyperbolic Space

Representation Tradeoffs for Hyperbolic Embeddings

Efficient Approximate Algorithms for the Closest Pair Problem in High Dimensional Spaces.

Diversity Embeddings and the Hypergraph Sparsest Cut

On the metric spaces of lattices and periodic point sets

Embedding Heterogeneous Information Network in Hyperbolic Spaces

Provably Accurate and Scalable Linear Classifiers in Hyperbolic Spaces

Hyperbolic Distance Matrices

No-dimensional Tverberg Partitions Revisited

Data Structures for Approximate Discrete Fréchet Distance

Optimal Private Halfspace Counting via Discrepancy

Central limit theorems for the nearest neighbour embracing graph in Euclidean and hyperbolic space

Additive approximation algorithm for geodesic centers in $δ$-hyperbolic graphs

Random Laplacian Features for Learning with Hyperbolic Space

Dynamic Metric Embedding into $\ell_p$ Space

Physical Schemata Underlying Biological Pattern Formation - Examples, Issues and Strategies

A discretized point-hyperplane incidence bound in $\mathbb{R}^d$

Hyperbolic Metric Spaces and Stochastic Embeddings

Some geometric applications of the discrete heat flow