In current recommendation systems, temporal data shift poses a significant
challenge. The presence of data shift prevents the system from simply enhancing
the CTR model's adaptability to new data by adding more training data. We
observed that although the correlation between features and labels in
recommendation systems changes over time, if a fixed search space is
established, the relationship between the data and the search space remains
invariant. Therefore, we designed a framework that uses retrieval techniques to
leverage shifting data for training a relevance network. However, due to the
use of BM25 as a retrieval method, this framework is challenging to deploy in
online recommendation systems. We then designed a distillation method using
knowledge distillation to transfer knowledge from the relevance network to a
parameterized module, the search-distill module. We refer to this entire
process as the Retrieval and Distill paradigm (RAD). With the RAD paradigm, we
have an effective method for leveraging shifting data to enhance the
performance of CTR models. In future research directions, we aim to incorporate
a wider variety of data into the CTR model using RAD. On the other hand,
enhancing the performance of the distillation method is also a significant area
of focus.

在当前的推荐系统中，时间数据的转变构成了一个重要挑战。为应对数据转变，我们设计了一种使用检索技术的框架来利用转变的数据来训练相关性网络，并使用知识蒸馏将相关性网络的知识转移到参数化模块中，即搜索蒸馏模块。我们将这整个过程称为检索与蒸馏范式（RAD）。在未来的研究方向中，我们旨在利用 RAD 将更多种类的数据纳入 CTR 模型，并提升蒸馏方法的性能。

在线推荐系统的检索与提取：一种无时序数据偏移的框架

Retrieval and Distill: A Temporal Data Shift Free Framework for Online  Recommendation System

The recently introduced Controlled Text Reduction (CTR) task isolates the
text generation step within typical summarization-style tasks. It does so by
challenging models to generate coherent text conforming to pre-selected content
within the input text ("highlights").
This framing enables increased modularity in summarization-like tasks,
allowing to couple a single CTR model with various content-selection setups and
modules.
However, there are currently no reliable CTR models, while the performance of
the existing baseline for the task is mediocre, falling short of practical
utility.
Here, we address this gap by introducing a high-quality, open-source CTR
model that tackles two prior key limitations: inadequate enforcement of the
content-preservation constraint, and suboptimal silver training data.
Addressing these, we amplify the content-preservation constraint in both
training, via RL, and inference, via a controlled decoding strategy.
Further, we substantially improve the silver training data quality via GPT-4
distillation.
Overall, pairing the distilled dataset with the highlight-adherence
strategies yields marked gains over the current baseline, of up to 30 ROUGE-L
points, providing a reliable CTR model for downstream use.

我们通过增强内容保留限制，改进了已有的基线模型，并利用 GPT-4 精馏方法大幅提高银标注训练数据的质量，从而在满足摘要生成式任务需求的情况下，提供了一个可靠的 CTR 模型。

不添加，不错过：从预选文本片段中有效地生成保留内容

Dont Add, dont Miss: Effective Content Preserving Generation from  Pre-Selected Text Spans

Since clicks usually contain heavy noise, increasing research efforts have
been devoted to modeling implicit negative user behaviors (i.e., non-clicks).
However, they either rely on explicit negative user behaviors (e.g., dislikes)
or simply treat non-clicks as negative feedback, failing to learn negative user
interests comprehensively. In such situations, users may experience fatigue
because of seeing too many similar recommendations. In this paper, we propose
Fatigue-Aware Network (FAN), a novel CTR model that directly perceives user
fatigue from non-clicks. Specifically, we first apply Fourier Transformation to
the time series generated from non-clicks, obtaining its frequency spectrum
which contains comprehensive information about user fatigue. Then the frequency
spectrum is modulated by category information of the target item to model the
bias that both the upper bound of fatigue and users' patience is different for
different categories. Moreover, a gating network is adopted to model the
confidence of user fatigue and an auxiliary task is designed to guide the
learning of user fatigue, so we can obtain a well-learned fatigue
representation and combine it with user interests for the final CTR prediction.
Experimental results on real-world datasets validate the superiority of FAN and
online A/B tests also show FAN outperforms representative CTR models
significantly.

本文提出了一种新的 CTR 模型 —— 疲劳感知网络 (Fatigue-Aware Network，FAN)，可以直接从用户的非点击行为中感知疲劳。通过将离散傅里叶变换应用于非点击生成的时间序列，获取用户疲劳的全面信息，进而结合目标项的类别信息来模拟用户疲劳的上限和耐心不同的偏见，并采用一种门控网络来模拟用户的疲劳信心。通过实验结果表明，该模型在真实数据集上具有优越性，并且在线 A/B 测试也显示 FAN 明显优于 CTR 模型。

FAN: 电子商务推荐中基于疲劳感知的点击率预测网络

FAN: Fatigue-Aware Network for Click-Through Rate Prediction in  E-commerce Recommendation

User behaviors on an e-commerce app not only contain different kinds of
feedback on items but also sometimes imply the cognitive clue of the user's
decision-making. For understanding the psychological procedure behind user
decisions, we present the behavior path and propose to match the user's current
behavior path with historical behavior paths to predict user behaviors on the
app. Further, we design a deep neural network for behavior path matching and
solve three difficulties in modeling behavior paths: sparsity, noise
interference, and accurate matching of behavior paths. In particular, we
leverage contrastive learning to augment user behavior paths, provide behavior
path self-activation to alleviate the effect of noise, and adopt a two-level
matching mechanism to identify the most appropriate candidate. Our model shows
excellent performance on two real-world datasets, outperforming the
state-of-the-art CTR model. Moreover, our model has been deployed on the
Meituan food delivery platform and has accumulated 1.6% improvement in CTR and
1.8% improvement in advertising revenue.

应用深度学习模型和对比学习方法分析用户购买行为路径，能够有效提高电商 app 中 CTR 模型的性能。