AI summaryⓘ
The authors developed a new deep learning model called Dual-Stream Attention (DSA) to better predict flu-like illness rates up to 12 weeks ahead by combining both numbers and news headlines. Their method uses two separate encoders for numbers and text and then connects them with a special attention mechanism that lets each type of data influence the other. Compared to other forecasting models, DSA showed lower prediction errors across multiple tests and datasets. Their tests also revealed that using attention in both directions between text and numbers works better than just one way. Additionally, the model meaningfully uses the news information when making predictions, especially when adjusting the influence of text on numbers.
Influenza-like illness (ILI)Multimodal deep learningTransformerCross-Modal AttentionTime series forecastingNumerical encodingTextual encodingMean Squared Error (MSE)Ablation studyFaithfulness analysis
Abstract
Forecasting long-range influenza-like illness (ILI) matters for public health readiness. Publicly available surveillance datasets typically pair numeric epidemiological signals with textual information that is noisy, loosely structured, only indirectly related to near-term trends, and often lagged relative to the numeric signal. Fusing the two therefore requires careful design. We propose Dual-Stream Attention (DSA), a multimodal deep learning framework that forecasts 12-week-ahead ILI activity from a 36-week multimodal history by letting the numerical and textual streams condition each other. Using the Time-MMD health-domain dataset, DSA separately encodes the two modalities with a Transformer-based numerical encoder and a domain-adapted headline encoder, then couples them through a bidirectional Cross-Modal Attention (CMA) mechanism: the text (news headlines) conditions the interpretation of the numeric signal and vice versa. The CMA output then passes to a causal temporal model for forecasting. Evaluated across ten random seeds, DSA achieves a median test MSE of 0.416, versus 0.668, 0.607, and 0.851 for iTransformer, TaTS, and GPT4MTS, corresponding to mean-error reductions of 54.95%, 37.29%, and 67.23%, with paired Cohen's d of 0.555, 0.337, and 0.345, respectively, and ranks first in 100% of bootstrap draws. It also has substantially lower worst-window error than all baselines. On an external-geography dataset, DSA again ranks first among nine evaluated baselines. Ablations show the advantage does not depend on text-encoder choice or language-model fine-tuning, and that bidirectional attention outperforms either direction alone. Finally, perturbation-based faithfulness analysis shows the learned CMA is functionally informative under targeted masking, with a stronger effect in the text-to-numerical direction.