FEDIN: Frequency-Enhanced Deep Interest Network for Click-Through Rate Prediction
Zenan Dai, Jinpeng Wang, Junwei Pan, Dapeng Liu, Lei Xiao + 1 more
TLDR
FEDIN uses frequency-domain analysis with target-aware filtering to improve click-through rate prediction by capturing periodic user interests.
Key contributions
- Identifies distinct spectral entropy for user interests vs. noise, conditioned on target items.
- Proposes FEDIN, a novel network with a frequency-domain branch for CTR prediction.
- Introduces target-aware spectrum filtering to isolate periodic user interest signals.
- Achieves state-of-the-art performance and noise robustness on three public datasets.
Why it matters
This paper addresses a key limitation in sequential recommendation by effectively handling noisy time-series data. By leveraging frequency-domain insights and target-aware filtering, FEDIN significantly improves the accuracy and robustness of CTR prediction. This approach offers a new direction for capturing subtle, periodic user behaviors.
Original Abstract
Sequential recommendation models often struggle to capture latent periodic patterns in user interests, primarily due to the noise inherent in time-domain behavioral data. While frequency-domain analysis offers a global perspective to address this, existing approaches typically treat user sequences in isolation, overlooking the crucial context of the target item. In this work, we present a novel empirical observation: user attention scores exhibit distinct spectral entropy distributions when conditioned on positive versus negative target items. Specifically, true user interests manifest as highly concentrated spectral patterns with lower entropy in the frequency domain, whereas irrelevant behaviors appear as high-entropy noise. Leveraging this insight, we propose the Frequency-Enhanced Deep Interest Network (FEDIN). FEDIN introduces a frequency-domain branch that utilizes a target-aware spectrum filtering mechanism to isolate these periodic interest signals. Extensive experiments on three public datasets demonstrate that FEDIN consistently outperforms state-of-the-art sequential recommendation baselines, demonstrating superior robustness against noise. We have released our code at: https://github.com/otokoneko/FEDIN.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.