T-SaS: Toward Shift-aware Dynamic Adaptation for Streaming Data

TL;DR

This paper introduces T-SaS, a Bayesian framework for detecting and adapting to abrupt distribution shifts in streaming data, enabling dynamic model selection and improved performance in forecasting and classification tasks.

Contribution

The paper proposes a novel Bayesian approach with discrete variables for shift detection and dynamic network adaptation, addressing the challenge of sudden, unanticipated distribution changes.

Findings

Superior shift boundary detection accuracy

Effective adaptation to downstream tasks

Outperforms existing methods in dynamic environments

Abstract

In many real-world scenarios, distribution shifts exist in the streaming data across time steps. Many complex sequential data can be effectively divided into distinct regimes that exhibit persistent dynamics. Discovering the shifted behaviors and the evolving patterns underlying the streaming data are important to understand the dynamic system. Existing methods typically train one robust model to work for the evolving data of distinct distributions or sequentially adapt the model utilizing explicitly given regime boundaries. However, there are two challenges: (1) shifts in data streams could happen drastically and abruptly without precursors. Boundaries of distribution shifts are usually unavailable, and (2) training a shared model for all domains could fail to capture varying patterns. This paper aims to solve the problem of sequential data modeling in the presence of sudden distribution shifts that occur without any precursors. Specifically, we design a Bayesian framework, dubbed as T-SaS, with a discrete distribution-modeling variable to capture abrupt shifts of data. Then, we design a model that enable adaptation with dynamic network selection conditioned on that discrete variable. The proposed method learns specific model parameters for each distribution by learning which neurons should be activated in the full network. A dynamic masking strategy is adopted here to support inter-distribution transfer through the overlapping of a set of sparse networks. Extensive experiments show that our proposed method is superior in both accurately detecting shift boundaries to get segments of varying distributions and effectively adapting to downstream forecast or classification tasks.