Hedging Is Not All You Need: A Simple Baseline for Online Learning Under Haphazard Inputs

TL;DR

This paper introduces HapNet, a simple, scalable baseline for online learning with haphazard streaming data, outperforming complex hedging-based methods and addressing variable window scaling challenges.

Contribution

HapNet simplifies online learning under inconsistent data streams by using plain self-attention, eliminating the need for complex architectures and specialized components.

Findings

HapNet achieves competitive performance on five benchmarks.

It effectively handles variable window scaling with uncorrelated data.

The approach is scalable and does not require online backpropagation.

Abstract

Handling haphazard streaming data, such as data from edge devices, presents a challenging problem. Over time, the incoming data becomes inconsistent, with missing, faulty, or new inputs reappearing. Therefore, it requires models that are reliable. Recent methods to solve this problem depend on a hedging-based solution and require specialized elements like auxiliary dropouts, forked architectures, and intricate network design. We observed that hedging can be reduced to a special case of weighted residual connection; this motivated us to approximate it with plain self-attention. In this work, we propose HapNet, a simple baseline that is scalable, does not require online backpropagation, and is adaptable to varying input types. All present methods are restricted to scaling with a fixed window; however, we introduce a more complex problem of scaling with a variable window where the data becomes positionally uncorrelated, and cannot be addressed by present methods. We demonstrate that a variant of the proposed approach can work even for this complex scenario. We extensively evaluated the proposed approach on five benchmarks and found competitive performance.