ROMAN: A Multiscale Routing Operator for Convolutional Time Series Models

TL;DR

ROMAN is a deterministic multiscale operator for time series that enhances model interpretability and efficiency by explicitly encoding temporal scale and position into channel structures, improving downstream classifier performance.

Contribution

It introduces ROMAN, a novel multiscale routing operator that constructs explicit channel representations for time series, enabling better control over model inductive biases and computational efficiency.

Findings

ROMAN improves classification accuracy on tasks sensitive to temporal structure.

ROMAN often enhances computational efficiency by shortening the processed time axis.

The effect of ROMAN on accuracy varies depending on the specific task.

Abstract

We introduce ROMAN (ROuting Multiscale representAtioN), a deterministic operator for time series that maps temporal scale and coarse temporal position into an explicit channel structure while reducing sequence length. ROMAN builds an anti-aliased multiscale pyramid, extracts fixed-length windows from each scale, and stacks them as pseudochannels, yielding a compact representation on which standard convolutional classifiers can operate. In this way, ROMAN provides a simple mechanism to control the inductive bias of downstream models: it can reduce temporal invariance, make temporal pooling implicitly coarse-position-aware, and expose multiscale interactions through channel mixing, while often improving computational efficiency by shortening the processed time axis. We formally analyze the ROMAN operator and then evaluate it in two complementary ways by measuring its impact as a preprocessing step for four representative convolutional classifiers: MiniRocket, MultiRocket, a standard CNN-based classifier, and a fully convolutional network (FCN) classifier. First, we design synthetic time series classification tasks that isolate coarse position awareness, long-range correlation, multiscale interaction, and full positional invariance, showing that ROMAN behaves consistently with its intended mechanism and is most useful when class information depends on temporal structure that standard pooled convolution tends to suppress. Second, we benchmark the same models with and without ROMAN on long-sequence subsets of the UCR and UEA archives, showing that ROMAN provides a practically useful alternative representation whose effect on accuracy is task-dependent, but whose effect on efficiency is often favorable. Code is available at https://github.com/gon-uri/ROMAN