Zhouhong Gu, Xingzhou Chen, Xiaoran Shi, Tao Wang, Suhang Zheng, Tianyu Li, Hongwei Feng, Yanghua Xiao. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Explainable fake news detection predicts the authenticity of news items with annotated explanations. Today, Large Language Models (LLMs) are known for their powerful natural language understanding and explanation generation abilities. However, using LLMs for explainable fake news detection remains two main challenges. Firstly, fake news appears reasonable and could easily mislead LLMs, leaving them unable to understand the complex news-faking process. Secondly, utilizing LLMs for this task would generate correct and incorrect explanations, requiring abundant labor in the loop. In this paper, we propose LLM-GAN, a novel framework that utilizes prompting mechanisms to enable an LLM to function as a Generator and a Detector for realistic fake news generation and detection. Extensive experimental results demonstrate LLM-GAN’s effectiveness in both prediction performance and explanation quality.

We introduce *GraphGPT*, a novel self-supervised *generative pre-trained* model for graph learning based on the *Graph Eulerian Transformer* (**GET**). First, we propose **GET**, which combines a standard transformer encoder or decoder architecture with an innovative graph-to-sequence transformation method. This method converts graphs or sampled subgraphs into sequences of tokens representing nodes, edges, and attributes in a reversible manner using Eulerian paths. We pre-train **GET** using either of the two self-supervised tasks: next-token prediction (NTP) and scheduled masked-token prediction (SMTP). The pre-trained model is then fine-tuned for downstream tasks such as graph-, edge-, and node-level prediction. Despite its simplicity, GraphGPT achieves performance comparable to or surpassing state-of-the-art methods on multiple large-scale Open Graph Benchmark (OGB) datasets. It demonstrates exceptional results on the molecular property prediction dataset PCQM4Mv2 and the protein-protein interaction dataset ogbl-ppa. Notably, generative pre-training enables scaling GraphGPT to 2 billion parameters while maintaining performance gains — a breakthrough that overcomes the scalability limitations of traditional Graph Neural Networks (GNNs) and prior graph transformers (GTs). To advance research in graph foundation models and facilitate scientific discovery in chemistry, materials science, and related fields, we have released the source code (https://github.com/alibaba/graph-gpt) and model checkpoints (https://www.modelscope.cn/organization/Alibaba-DT).

When doing private domain marketing with cloud services, the merchants usually have to purchase different machine learning models for the multiple marketing purposes, leading to a very high cost. We present a unified user-item matching framework to simultaneously conduct item recommendation and user targeting with just one model. We empirically demonstrate that the above concurrent modeling is viable via modeling the user-item interaction matrix with the multinomial distribution, and propose a bidirectional bias-corrected NCE loss for the implementation. The proposed loss function guides the model to learn the user-item joint probability p(u,i) instead of the conditional probability p(i|u) or p(u|i) through correcting both the users and items’ biases caused by the in-batch negative sampling. In addition, our framework is model-agnostic enabling a flexible adaptation of different model architectures. Extensive experiments demonstrate that our framework results in significant performance gains in comparison with the state-of-the-art methods, with greatly reduced cost on computing resources and daily maintenance.

Autoencoder-based hybrid recommender systems have become popular recently because of their ability to learn user and item representations by reconstructing various information sources, including users' feedback on items (e.g., ratings) and side information of users and items (e.g., users' occupation and items' title). However, existing systems still use representations learned by matrix factorization (MF) to predict the rating, while using representations learned by neural networks as the regularizer. In this paper, we define the neural representation for prediction (NRP) framework and apply it to the autoencoder-based recommendation systems. We theoretically analyze how our objective function is related to the previous MF and autoencoder-based methods and explain what it means to use neural representations as the regularizer. We also apply the NRP framework to a direct neural network structure which predicts the ratings without reconstructing the user and item information. We conduct extensive experiments which confirm that neural representations are better for prediction than regularization and show that the NRP framework outperforms the state-of-the-art methods in the prediction task, with less training time and memory.

Meta-learning approaches have shown great success in solving challenging knowledge transfer and fast adaptation problems with few samples in vision and language domains. However, few studies discuss the practice of meta-learning for large-scale industrial applications, e.g., representation learning for e-commerce platform users. Although e-commerce companies have spent many efforts on learning accurate and expressive representations to provide a better user experience, we argue that such efforts cannot be stopped at this step. In addition to learning a strong profile of user behaviors, the challenging question about how to effectively transfer the learned representation and quickly adapt the learning process to the subsequent learning tasks or applications is raised simultaneously.

Existing algorithms aiming to learn a binary classifier from positive (P) and unlabeled (U) data generally require estimating the class prior or label noises ahead of building a classification model. However, the estimation and classifier learning are normally conducted in a pipeline instead of being jointly optimized. In this paper, we propose to alternatively train the two steps using reinforcement learning. Our proposal adopts a policy network to adaptively make assumptions on the labels of unlabeled data, while a classifier is built upon the output of the policy network and provides rewards to learn a better strategy. The dynamic and interactive training between the policy maker and the classifier can exploit the unlabeled data in a more effective manner and yield a significant improvement on the classification performance. Furthermore, we present two different approaches to represent the actions sampled from the policy. The first approach considers continuous actions as soft labels, while the other uses discrete actions as hard assignment of labels for unlabeled examples. We validate the effectiveness of the proposed method on two benchmark datasets as well as one e-commerce dataset. The result shows the proposed method is able to consistently outperform state-of-the-art methods in various settings.

User preferences in the form of absolute feedback, s.a., ratings, are widely exploited in Recommender Systems (RSs). Recent research has explored the usage of preferences expressed with pairwise comparisons, which signal relative feedback. It has been shown that pairwise comparisons can be effectively combined with ratings, but, it is important to fine tune the technique that leverages both types of feedback. Previous approaches train a single model by converting ratings into pairwise comparisons, and then use only that type of data. However, we claim that these two types of preferences reveal different information about users interests and should be exploited differently. Hence, in this work, we develop a ranking technique that separately exploits absolute and relative preferences in a hybrid model. In particular, we propose a joint loss function which is computed on both absolute and relative preferences of users. Our proposed ranking model uses pairwise comparisons data to predict the user's preference order between pairs of items and uses ratings to push high rated (relevant) items to the top of the ranking. Experimental results on three different data sets demonstrate that the proposed technique outperforms competitive baseline algorithms on popular ranking-oriented evaluation metrics.