The Label Horizon Paradox: Rethinking Supervision Targets in Financial Forecasting
This work addresses a fundamental but overlooked issue in financial forecasting for practitioners and researchers, though it appears incremental as it builds on existing deep learning approaches.
The paper tackles the problem of suboptimal supervision signals in financial forecasting by identifying the Label Horizon Paradox, where optimal training labels differ from prediction targets due to market dynamics. The proposed bi-level optimization framework achieves consistent improvements over conventional baselines on large-scale financial datasets.
While deep learning has revolutionized financial forecasting through sophisticated architectures, the design of the supervision signal itself is rarely scrutinized. We challenge the canonical assumption that training labels must strictly mirror inference targets, uncovering the Label Horizon Paradox: the optimal supervision signal often deviates from the prediction goal, shifting across intermediate horizons governed by market dynamics. We theoretically ground this phenomenon in a dynamic signal-noise trade-off, demonstrating that generalization hinges on the competition between marginal signal realization and noise accumulation. To operationalize this insight, we propose a bi-level optimization framework that autonomously identifies the optimal proxy label within a single training run. Extensive experiments on large-scale financial datasets demonstrate consistent improvements over conventional baselines, thereby opening new avenues for label-centric research in financial forecasting.