Chen Cui

Amala Sanjay Deshmukh, Kateryna Chumachenko, Tuomas Rintamaki et al. · nvidia

We introduce Nemotron Nano V2 VL, the latest model of the Nemotron vision-language series designed for strong real-world document understanding, long video comprehension, and reasoning tasks. Nemotron Nano V2 VL delivers significant improvements over our previous model, Llama-3.1-Nemotron-Nano-VL-8B, across all vision and text domains through major enhancements in model architecture, datasets, and training recipes. Nemotron Nano V2 VL builds on Nemotron Nano V2, a hybrid Mamba-Transformer LLM, and innovative token reduction techniques to achieve higher inference throughput in long document and video scenarios. We are releasing model checkpoints in BF16, FP8, and FP4 formats and sharing large parts of our datasets, recipes and training code.

Emmy Liu, Chen Cui, Kenneth Zheng et al. · cmu

Figurative and metaphorical language are commonplace in discourse, and figurative expressions play an important role in communication and cognition. However, figurative language has been a relatively under-studied area in NLP, and it remains an open question to what extent modern language models can interpret nonliteral phrases. To address this question, we introduce Fig-QA, a Winograd-style nonliteral language understanding task consisting of correctly interpreting paired figurative phrases with divergent meanings. We evaluate the performance of several state-of-the-art language models on this task, and find that although language models achieve performance significantly over chance, they still fall short of human performance, particularly in zero- or few-shot settings. This suggests that further work is needed to improve the nonliteral reasoning capabilities of language models.

Amala Sanjay Deshmukh, Kateryna Chumachenko, Tuomas Rintamaki et al. · amazon-science, nvidia

We introduce Nemotron 3 Nano Omni, the latest model in the Nemotron multimodal series and the first to natively support audio inputs alongside text, images, and video. Nemotron 3 Nano Omni delivers consistent accuracy improvements over its predecessor, Nemotron Nano V2 VL, across all modalities, enabled by advances in architecture, training data and recipes. In particular, Nemotron 3 delivers leading results in real-world document understanding, long audio-video comprehension, and agentic computer use. Built on the highly efficient Nemotron 3 Nano 30B-A3B backbone, Nemotron 3 Nano Omni further incorporates innovative multimodal token-reduction techniques to deliver substantially lower inference latency and higher throughput than other models of similar size. We are releasing model checkpoints in BF16, FP8, and FP4 formats, along with portions of the training data and codebase to facilitate further research and development.

Zecong Wang, Jiaxi Cheng, Chen Cui et al.

The abundance of information on social media has increased the necessity of accurate real-time rumour detection. Manual techniques of identifying and verifying fake news generated by AI tools are impracticable and time-consuming given the enormous volume of information generated every day. This has sparked an increase in interest in creating automated systems to find fake news on the Internet. The studies in this research demonstrate that the BERT and RobertA models with fine-tuning had the best success in detecting AI generated news. With a score of 98%, tweaked RobertA in particular showed excellent precision. In conclusion, this study has shown that neural networks can be used to identify bogus news AI generation news created by ChatGPT. The RobertA and BERT models' excellent performance indicates that these models can play a critical role in the fight against misinformation.

Jinyin Chen, Guohan Huang, Haibin Zheng et al.

Graph neural network (GNN) has achieved great success on graph representation learning. Challenged by large scale private data collected from user-side, GNN may not be able to reflect the excellent performance, without rich features and complete adjacent relationships. Addressing the problem, vertical federated learning (VFL) is proposed to implement local data protection through training a global model collaboratively. Consequently, for graph-structured data, it is a natural idea to construct a GNN based VFL framework, denoted as GVFL. However, GNN has been proved vulnerable to adversarial attacks. Whether the vulnerability will be brought into the GVFL has not been studied. This is the first study of adversarial attacks on GVFL. A novel adversarial attack method is proposed, named Graph-Fraudster. It generates adversarial perturbations based on the noise-added global node embeddings via the privacy leakage and the gradient of pairwise node. Specifically, first, Graph-Fraudster steals the global node embeddings and sets up a shadow model of the server for the attack generator. Second, noise is added into node embeddings to confuse the shadow model. At last, the gradient of pairwise node is used to generate attacks with the guidance of noise-added node embeddings. Extensive experiments on five benchmark datasets demonstrate that Graph-Fraudster achieves the state-of-the-art attack performance compared with baselines in different GNN based GVFLs. Furthermore, Graph-Fraudster can remain a threat to GVFL even if two possible defense mechanisms are applied. Additionally, some suggestions are put forward for the future work to improve the robustness of GVFL. The code and datasets can be downloaded at https://github.com/hgh0545/Graph-Fraudster.

Jinyin Chen, Dunjie Zhang, Zhaoyan Ming et al.

Graph neural networks (GNNs) have been successfully exploited in graph analysis tasks in many real-world applications. The competition between attack and defense methods also enhances the robustness of GNNs. In this competition, the development of adversarial training methods put forward higher requirement for the diversity of attack examples. By contrast, most attack methods with specific attack strategies are difficult to satisfy such a requirement. To address this problem, we propose GraphAttacker, a novel generic graph attack framework that can flexibly adjust the structures and the attack strategies according to the graph analysis tasks. GraphAttacker generates adversarial examples through alternate training on three key components: the multi-strategy attack generator (MAG), the similarity discriminator (SD), and the attack discriminator (AD), based on the generative adversarial network (GAN). Furthermore, we introduce a novel similarity modification rate SMR to conduct a stealthier attack considering the change of node similarity distribution. Experiments on various benchmark datasets demonstrate that GraphAttacker can achieve state-of-the-art attack performance on graph analysis tasks of node classification, graph classification, and link prediction, no matter the adversarial training is conducted or not. Moreover, we also analyze the unique characteristics of each task and their specific response in the unified attack framework. The project code is available at https://github.com/honoluluuuu/GraphAttacker.

Yun Liu, Chen Cui, Shi Shu et al.

The numerical solution of partial differential equations (PDEs) is fundamental to scientific and engineering computing. In the presence of strong anisotropy, material heterogeneity, and complex geometries, however, classical iterative solvers often suffer from reduced efficiency and require substantial problem-dependent tuning. The Fourier neural solver (FNS) is a learning-based hybrid iterative solver for such problems without extensive manual parameter tuning, but its original design is primarily effective for scalar PDEs on structured meshes and is difficult to extend directly to unstructured meshes or strongly coupled PDE systems. Building on the FNS framework, we introduce block smoothing operators and graph neural networks to construct a solver for unstructured systems, termed the graph Fourier neural solver (G-FNS). We further incorporate a coordinate transformation network to develop the adaptive graph Fourier neural solver (AG-FNS), and then extend this formulation to a frequency-domain multilevel variant, ML-AG-FNS. Rigorous analysis shows that, under suitable mathematical assumptions, the proposed method achieves mesh-independent convergence rate. Error-spectrum visualizations further indicate that AG-FNS can capture complex multiscale error modes. Extensive experiments on two-dimensional anisotropic diffusion and on two- and three-dimensional isotropic/anisotropic linear elasticity problems over unstructured meshes demonstrate strong robustness and efficiency. The proposed framework can be used either as a solver or as a preconditioner for Krylov subspace methods. Overall, it substantially extends the original FNS methodology and broadens the applicability of this class of neural solvers.

Jinyin Chen, Ruoxi Chen, Haibin Zheng et al.

Recent years have witnessed unprecedented success achieved by deep learning models in the field of computer vision. However, their vulnerability towards carefully crafted adversarial examples has also attracted the increasing attention of researchers. Motivated by the observation that adversarial examples are due to the non-robust feature learned from the original dataset by models, we propose the concepts of salient feature(SF) and trivial feature(TF). The former represents the class-related feature, while the latter is usually adopted to mislead the model. We extract these two features with coupled generative adversarial network model and put forward a novel detection and defense method named salient feature extractor (SFE) to defend against adversarial attacks. Concretely, detection is realized by separating and comparing the difference between SF and TF of the input. At the same time, correct labels are obtained by re-identifying SF to reach the purpose of defense. Extensive experiments are carried out on MNIST, CIFAR-10, and ImageNet datasets where SFE shows state-of-the-art results in effectiveness and efficiency compared with baselines. Furthermore, we provide an interpretable understanding of the defense and detection process.

Pingping Lu, Chen Cui, Shaobing Xu et al.

AI-based lane detection algorithms were actively studied over the last few years. Many have demonstrated superior performance compared with traditional feature-based methods. The accuracy, however, is still generally in the low 80% or high 90%, or even lower when challenging images are used. In this paper, we propose a real-time lane detection system, called Scene Understanding Physics-Enhanced Real-time (SUPER) algorithm. The proposed method consists of two main modules: 1) a hierarchical semantic segmentation network as the scene feature extractor and 2) a physics enhanced multi-lane parameter optimization module for lane inference. We train the proposed system using heterogeneous data from Cityscapes, Vistas and Apollo, and evaluate the performance on four completely separate datasets (that were never seen before), including Tusimple, Caltech, URBAN KITTI-ROAD, and X-3000. The proposed approach performs the same or better than lane detection models already trained on the same dataset and performs well even on datasets it was never trained on. Real-world vehicle tests were also conducted. Preliminary test results show promising real-time lane-detection performance compared with the Mobileye.

Zhaohong Deng, Chen Cui, Peng Xu et al.

As the multi-view data grows in the real world, multi-view clus-tering has become a prominent technique in data mining, pattern recognition, and machine learning. How to exploit the relation-ship between different views effectively using the characteristic of multi-view data has become a crucial challenge. Aiming at this, a hidden space sharing multi-view fuzzy clustering (HSS-MVFC) method is proposed in the present study. This method is based on the classical fuzzy c-means clustering model, and obtains associ-ated information between different views by introducing shared hidden space. Especially, the shared hidden space and the fuzzy partition can be learned alternatively and contribute to each other. Meanwhile, the proposed method uses maximum entropy strategy to control the weights of different views while learning the shared hidden space. The experimental result shows that the proposed multi-view clustering method has better performance than many related clustering methods.

Peng Xu, Zhaohong Deng, Chen Cui et al.

The superior interpretability and uncertainty modeling ability of Takagi-Sugeno-Kang fuzzy system (TSK FS) make it possible to describe complex nonlinear systems intuitively and efficiently. However, classical TSK FS usually adopts the whole feature space of the data for model construction, which can result in lengthy rules for high-dimensional data and lead to degeneration in interpretability. Furthermore, for highly nonlinear modeling task, it is usually necessary to use a large number of rules which further weakens the clarity and interpretability of TSK FS. To address these issues, a concise zero-order TSK FS construction method, called ESSC-SL-CTSK-FS, is proposed in this paper by integrating the techniques of enhanced soft subspace clustering (ESSC) and sparse learning (SL). In this method, ESSC is used to generate the antecedents and various sparse subspace for different fuzzy rules, whereas SL is used to optimize the consequent parameters of the fuzzy rules, based on which the number of fuzzy rules can be effectively reduced. Finally, the proposed ESSC-SL-CTSK-FS method is used to construct con-cise zero-order TSK FS that can explain the scenes in high-dimensional data modeling more clearly and easily. Experiments are conducted on various real-world datasets to confirm the advantages.