Toward Time-Continuous Data Inference in Sparse Urban CrowdSensing

TL;DR

This work addresses the limitation of time-discrete assumptions in Sparse Mobile Crowd Sensing by introducing time-continuous data completion. It develops a progression of models—DMF, RNN-DMF, and TIME-DMF—with time gates and a Query-Generate strategy to capture temporal dynamics and infinite future states. The main contributions include reformulating completion on a continuous timeline, enabling fine-grained and time-continuous inference under extreme sparsity, and providing extensive experiments across four real datasets demonstrating improved accuracy, robustness to uneven sampling, and generative capabilities. The results suggest substantial practical impact for real-time sensing applications where data evolves continuously and observation is sparse.

Abstract

Mobile Crowd Sensing (MCS) is a promising paradigm that leverages mobile users and their smart portable devices to perform various real-world tasks. However, due to budget constraints and the inaccessibility of certain areas, Sparse MCS has emerged as a more practical alternative, collecting data from a limited number of target subareas and utilizing inference algorithms to complete the full sensing map. While existing approaches typically assume a time-discrete setting with data remaining constant within each sensing cycle, this simplification can introduce significant errors, especially when dealing with long cycles, as real-world sensing data often changes continuously. In this paper, we go from fine-grained completion, i.e., the subdivision of sensing cycles into minimal time units, towards a more accurate, time-continuous completion. We first introduce Deep Matrix Factorization (DMF) as a neural network-enabled framework and enhance it with a Recurrent Neural Network (RNN-DMF) to capture temporal correlations in these finer time slices. To further deal with the continuous data, we propose TIME-DMF, which captures temporal information across unequal intervals, enabling time-continuous completion. Additionally, we present the Query-Generate (Q-G) strategy within TIME-DMF to model the infinite states of continuous data. Extensive experiments across five types of sensing tasks demonstrate the effectiveness of our models and the advantages of time-continuous completion.

Figures (32)

• Figure 1: Time-continuous and time-discrete sensing data in Sparse MCS.
• Figure 2: Time-discrete and time-continuous formulation.
• Figure 3: Relationship between DMF and TIME-DMF.
• Figure 4: The structure of DMF.
• Figure 5: The inner structure of RNN-DMF.