Overcoming the Modality Gap in Context-Aided Forecasting
This addresses the modality gap in context-aided forecasting for AI systems, but it is incremental as it focuses on dataset quality rather than architectural innovations.
The paper tackled the problem of multimodal models underperforming in context-aided forecasting due to poor context quality in datasets, and introduced a semi-synthetic data augmentation method that generated a 7 million context-augmented time series dataset, showing effective transfer to real-world evaluation and clear context utilization.
Context-aided forecasting (CAF) holds promise for integrating domain knowledge and forward-looking information, enabling AI systems to surpass traditional statistical methods. However, recent empirical studies reveal a puzzling gap: multimodal models often fail to outperform their unimodal counterparts. We hypothesize that this underperformance stems from poor context quality in existing datasets, as verification is challenging. To address these limitations, we introduce a semi-synthetic data augmentation method that generates contexts both descriptive of temporal dynamics and verifiably complementary to numerical histories. This approach enables massive-scale dataset creation, resulting in CAF-7M, a corpus of 7 million context-augmented time series windows, including a rigorously verified test set. We demonstrate that semi-synthetic pre-training transfers effectively to real-world evaluation, and show clear evidence of context utilization. Our results suggest that dataset quality, rather than architectural limitations, has been the primary bottleneck in context-aided forecasting.