Event time series appear in many software scenarios and are a necessary data type in data analytics systems. Transformers are the preferred type of sequential neural network for advanced analytics on event time series, particularly due to their significant contributions to the recent surge of large language models (LLMs). Event series analytics heavily depends on the quality of input data, which may contain natural measurement errors or adversarial noises. Since the input data deviates from the true state, the opaque nature of neural networks presents a challenge in ensuring the reliability of output, which might be deemed untrustworthy. In this paper, we introduce an innovative formal verification framework for Transformer-based event series systems, leveraging sampling, linear programming, and the extreme value theorem. This framework can support the verification of the dependability of Transformers in managing inputs characterized by unpredictability and uncertainty. To exemplify its utility, we apply our verification approach to verify natural requirements from a real-world event series environments: network traffic classification. It outperforms the current state-of-the-art verifier in terms of effectiveness, providing more stringent verified bounds. Our experimental findings provide valuable benchmarks for guaranteeing reliable deployment of systems in scenarios where the credibility of event data is compromised, and for exposing specific cases in which the expected requirements are not satisfied.

Large language models (LLMs) have demonstrated outstanding performance in various tasks. However, the memory footprint of the key-value (KV) cache generated during model inference poses significant challenges for efficient model deployment. This paper presents a detailed analysis of the KV cache and identifies potential sources of quantization errors. We find that the attention scores exhibit a strong power-law distribution while showing an additional attention to the initial and recent tokens, and that only a small subset of outlier channels in KV cache span a wide numerical range. Based on these findings, we propose QoKV, an efficient post-training quantization (PTQ) method designed for the KV cache, which facilitates accurate low-bit quantization through a novel group evaluation strategy and numerical scaling. Extensive experimental results across various tasks demonstrate that our proposed method outperforms existing methods, achieving near-floating-point performance while saving 73.75% of the cache footprint.

Explainable Fake News Detection (EFND) is a new challenge that aims to verify news authenticity and provide clear explanations for its decisions. Traditional EFND methods often treat the tasks of classification and explanation as separate, ignoring the fact that explanation content can assist in enhancing fake news detection. To overcome this gap, we present a new solution: the End-to-End Explainable Fake News Detection Network ( EExpFND ). Our model includes an evidence-claim variational causal inference component, which not only utilizes explanation content to improve fake news detection but also employs a variational approach to address the distributional bias between the ground truth explanation in the training set and the prediction explanation in the test set. Additionally, we incorporate a masked attention network to detail the nuanced relationships between evidence and claims. Our comprehensive tests across two public datasets show that EExpFND sets a new benchmark in performance. The code is available at https://anonymous.4open.science/r/EExpFND-F5C6 .

Fake news detection has gotten continuous attention during these years with more and more people have been posting and reading news online. To enable fake news detection, existing researchers usually assume labeled posts are provided for two classes (true or false) so that the model can learn a discriminative classifier from the labeled data. However, this supposition may not hold true in reality, as most users may only label a small number of posts in a single category that they are interested in. Furthermore, most existing methods fail to mask the noise or irrelevant context (i.e., regions or words) between different modalities to assist in strengthening the correlations between relevant contexts. To tackle these issues, we present a curriculum-based multi-modal masked transformer network (CMMTN) for positive unlabeled multi-modal fake news detection by jointly modeling the inter-modality and intra-modality relationships of multi-modal information and masking the irrelevant context between modalities. In particular, we adopt BERT and ResNet to obtain better representations for texts and images, separately. Then, the extracted features of images and texts are fed into a multi-modal masked transformer network to fuse the multi-modal content and mask the irrelevant context between modalities by calculating the similarity between inter-modal contexts. Finally, we design a curriculum-based PU learning method to handle the positive and unlabeled data. Massive experiments on three public real datasets prove the effectiveness of the CMMTN.

The rapid proliferation of social media data has led to the widespread dissemination of multi-modal fake news, prompting researchers to develop novel detection methods. Most fake news detection approaches mine the rich context information, including the text and image content of news and associated comments. However, existing methods are often insufficient to filter out irrelevant contexts, such as noisy words, redundant image regions, and spam comments, which may introduce noise into the model. Particularly, these approaches struggle to handle comments, which often contain the most severe noise. In many cases, only a minuscule portion of comments is relevant to the news. To overcome these limitations, our research introduces a novel Comment-Context Dual Collaborative Masked Transformer Network ($C^{2}DCMTN$). To handle the irrelevant contexts, we propose a Multi-modal Masked Transformer Network. This network extends the traditional Transformer with a mask mechanism capable of dynamically obscuring irrelevant multi-modal context information. To effectively deal with comments, which can suffer from more severe noise issues, we have designed a Comment-Context Encoder that focuses solely on the most crucial comments. Comprehensive experiments on two publicly available real-world datasets confirm that$C^{2}DCMTN$outperforms state-of-the-art methods.

In this article, for the nonzero-sum (NZS) differential games problem of uncertain nonlinear systems with state constraints, an adaptive robust stabilization scheme based on the control barrier function (CBF) is presented under the influence of random disturbances and control input matrix uncertainty. To deal with the impact of uncertainty on the system, the nominal system of the original system is adopted and the cost functions associated with each player are appropriately chosen to convert the robust regulation problem of multiplayer differential games into an optimal regulation problem. Furthermore, the purpose of combining the cost function relevant to each player with the CBF is to make the system states evolve in the safe area. Different from the classical actor–critic dual neural network (NN), each player only needs a critic NN to approach the corresponding cost function without the restriction of the initial stabilizing control. Combined with the Lyapunov stability theory, under the combined influence of random disturbances and state constraints, the state and critic NN weights of the closed-loop system are guaranteed to be uniformly ultimately bounded (UUB). Finally, two simulation examples are used to verify the validity of the presented scheme.

Nowadays, detecting fake news on social media platforms has become a top priority since the widespread dissemination of fake news may mislead readers and have negative effects. To date, many algorithms have been proposed to facilitate the detection of fake news from the hand-crafted feature extraction methods to deep learning approaches. However, these methods may suffer from the following limitations: (1) fail to utilize the multi-modal context information and extract high-order complementary information for each news to enhance the detection of fake news; (2) largely ignore the full hierarchical semantics of textual content to assist in learning a better news representation. To overcome these limitations, this paper proposes a novel hierarchical multi-modal contextual attention network (HMCAN) for fake news detection by jointly modeling the multi-modal context information and the hierarchical semantics of text in a unified deep model. Specifically, we employ BERT and ResNet to learn better representations for text and images, respectively. Then, we feed the obtained representations of images and text into a multi-modal contextual attention network to fuse both inter-modality and intra-modality relationships. Finally, we design a hierarchical encoding network to capture the rich hierarchical semantics for fake news detection. Extensive experiments on three public real datasets demonstrate that our proposed HMCAN achieves state-of-the-art performance.