Deep Neural Networks are vulnerable to data poisoning attacks, which inject a backdoor by poisoning the training data set with a predefined trigger pattern. However, most existing studies design trigger patterns as exogenous features introduced to clean samples (such as a checkerboard patch), whereas the endogenous features inherited from sample origins (such as deep generative models) have not been investigated yet. In this study, we investigate the efficacy of utilizing Generative Adversarial Network fingerprints to design trigger patterns by examining three attack patterns: the all-label attack, label-specific attack, and semantic-specific attack. Specifically, we select training data that satisfies the requirements of various attack patterns, train a GAN model, and employ samples embedded with GAN fingerprints to generate poisoned data sets. Our evaluations on three data sets (CIFAR-10, GTSRB, and LSUN) demonstrate that employing GAN fingerprints as trigger patterns 1) can achieve average attack success rate results of 39.40% in all-label attack, 89.84% in label-specific attack, and 91.06% in semantic-specific attack, 2) is stealthy by reducing the probability of being exposed, and 3) can resist six existing backdoor detection techniques, three backdoor erasing techniques, and two deepfake detection techniques. Furthermore, it exhibits practical applicability in federated learning scenario.

In modern process industries, data-driven models are important tools for real-time monitoring when key performance indicators are difficult to measure directly. While accurate predictions are essential, reliable uncertainty quantification (UQ) is equally critical for safety, reliability, and decision-making, but remains a major challenge in current data-driven approaches. In this work, we introduce a diffusion-based posterior sampling framework that inherently produces well-calibrated predictive uncertainty via faithful posterior sampling, eliminating the need for post hoc calibration. In extensive evaluations on synthetic distributions, the Raman-based phenylacetic acid soft sensor benchmark, and a real ammonia synthesis case study, our method achieves practical improvements over existing UQ techniques in both uncertainty calibration and predictive accuracy. These results highlight diffusion samplers as a principled and scalable paradigm for advancing uncertainty-aware modeling in industrial applications.

Nonlinear probabilistic latent variable models (NPLVMs) are a cornerstone of soft sensor modeling due to their capacity for uncertainty delineation. However, conventional NPLVMs are trained using amortized variational inference, where neural networks parameterize the variational posterior. While facilitating model implementation, this parameterization converts the distributional optimization problem within an infinite-dimensional function space to parameter optimization within a finite-dimensional parameter space, which introduces an approximation error gap, thereby degrading soft sensor modeling accuracy. To alleviate this issue, we introduce KProxNPLVM, a novel NPLVM that pivots to relaxing the objective itself and improving the NPLVM's performance. Specifically, we first prove the approximation error induced by the conventional approach. Based on this, we design the Wasserstein distance as the proximal operator to relax the learning objective, yielding a new variational inference strategy derived from solving this relaxed optimization problem. Based on this foundation, we provide a rigorous derivation of KProxNPLVM's optimization implementation, prove the convergence of our algorithm can finally sidestep the approximation error, and propose the KProxNPLVM by summarizing the abovementioned content. Finally, extensive experiments on synthetic and real-world industrial datasets are conducted to demonstrate the efficacy of the proposed KProxNPLVM.

Step-level reward models (SRMs) can significantly enhance mathematical reasoning performance through process supervision or step-level preference alignment based on reinforcement learning. The performance of SRMs is pivotal, as they serve as critical guidelines, ensuring that each step in the reasoning process is aligned with desired outcomes. Recently, AlphaZero-like methods, where Monte Carlo Tree Search (MCTS) is employed for automatic step-level preference annotation, have proven particularly effective. However, the precise mechanisms behind the success of SRMs remain largely unexplored. To address this gap, this study delves into the counterintuitive aspects of SRMs, particularly focusing on MCTS-based approaches. Our findings reveal that the removal of natural language descriptions of thought processes has minimal impact on the efficacy of SRMs. Furthermore, we demonstrate that SRMs are adept at assessing the complex logical coherence present in mathematical language while having difficulty in natural language. These insights provide a nuanced understanding of the core elements that drive effective step-level reward modeling in mathematical reasoning. By shedding light on these mechanisms, this study offers valuable guidance for developing more efficient and streamlined SRMs, which can be achieved by focusing on the crucial parts of mathematical reasoning.

This paper proposes an AI-assisted speculative narrative design workflow as a critical tool to address and respond to ecological ethics and anthropocentrism in the Anthropocene. Speculating on a post-climate collapse, ocean-dominated future, Project Serum constructs a multi-species narrative through the lens of a deep-sea court trial, where humans, whales, and robots contest ecological justice and species dominance. Combining speculative design methodologies with multispecies worldbuilding, this work builds a narrative blueprint using a structured 5W1H framework and AI-enhanced story design. This blueprint is then remediated into multiple formats, including video, postcards, and comic posters, through the proposed AI-assisted workflows integrating AI image generation and human editorial control. By applying these narrative remediation strategies with AI, this study challenges anthropocentric narrative structures and offers a reusable workflow for multi-modal, multi-species storytelling. It demonstrates the speculative narrative’s potential as a reflective intervention, capable of destabilising normative narrative hierarchies and foregrounding ecological justice with AI-mediated production.

Nonlinear dynamical probabilistic latent variable model (NDPLVM) and its variants, essential in industrial inferential sensors, face challenges in latent space inference and deep learning (DL) backend implementation. The first issue arises from the assumption that covariates directly infer the latent variable, potentially leading to inaccuracies. The second issue involves the discrepancy between the probabilistic distribution function form of NDPLVMs and data sample-based operation of DL backends. Addressing these, this study introduces the optimal control-NDPLVM (OC-NDPLVM), a model designed to enhance performance by analyzing NDPLVMs learning and tackling these issues. For the first problem, NDPLVMs' learning is reinterpreted as an optimization problem, solved by alternating direction method of multipliers, and selecting the inference network's input via studying optimal solution's structure. To address the second issue, OC-NDPLVM adapts mean and covariance equations for compatibility with DL backends. This model's effectiveness is validated through experiments on inferential sensor datasets.

Data-driven soft sensors have been prevalent in industrial key performance indicator prediction. However, rigorously quantifying the uncertainty of predictions has been consistently neglected. This neglect can result in industrial practitioners being ignorant of the reliability of the predictions, potentially leading to dangerous misjudgments. To bridge this gap, this study interprets uncertainty quantification (UQ) from a Bayesian perspective, which conceptualizes model uncertainty as the parameter distribution, thus distinguishing it from data uncertainty. Based on this interpretation, a novel UQ method, heat equation stein variational ensemble (HESVE), is proposed. HESVE introduces an efficient nonparametric approach called Stein variational gradient descent (SVGD) to approximate parameter distributions more precisely. A heat equation-based method is also adopted for adaptive hyperparameter tuning of SVGD. In addition, our method incorporates the variable noise estimator to capture heteroscedastic data noise, which contributes to uncertainty decomposition. Experiments demonstrate the superiority of HESVE in multiple aspects compared to other state-of-the-art methods.

A reasonable configuration of battery energy storage system (BESS) in distribution networks, especially those penetrated by photovoltaic (PV) systems, helps to improve the voltage stability and reduce the configuration and operation cost. In order to solve the sizing and placement problem of BESS in distribution network penetrated by PV systems, this paper proposes an integrated planning strategy for placement and sizing of BESS based on the Amartya Sen index. Inspired by the concept of Amartya Sen poverty index, the Amartya Sen index is adopted to describe the static voltage stability of distribution network in this paper. Considering three aspects of system static voltage stability, network loss, and BESS investment cost, a unified planning model for BESS placement and sizing is constructed with the goal of minimizing the Amartya Sen index and the economic cost. The model is solved based on power flow calculation and multi-objective particle swarm optimization algorithm (MOPSO). The case study conducted in the IEEE 33-node system validates the proposed method, which can effectively improve the static voltage stability of distribution network with reduced investment cost of BESS.