Provable quantum speedups for computing persistence in topological data analysis

Researchers have made significant advancements in topological data analysis (TDA) using both quantum algorithms and machine learning approaches.

Topological data analysis aims to extract noise-robust features from a data set by examining the number and persistence of holes in its topology^[1]. A recent study demonstrated that determining whether a given hole persists across different length scales is BQP1-hard, implying an exponential quantum speedup under standard complexity-theoretic assumptions^[1]. This result suggests that a classical solution is extremely unlikely. Meanwhile, another approach combines TDA and machine learning to extract actionable information from high-dimensional time-series data by representing multivariate time-series data as manifolds and using topological descriptors to summarize the structure of such data^[2]. A neural ordinary differential equation is used to learn the dynamic evolution of the topological structure of the system, which is effective at detecting diverse types of events in an industrial process^[2]. This approach is contrasted against reconstruction-based approaches such as principal component analysis and autoencoders, and against a trajectory-based approach that uses Koopman autoencoders^[2].