The success of Transformer-based Language Models (LMs) stems from their attention mechanism. While this mechanism has been extensively studied in explainability research, particularly through the attention values obtained during the forward pass of LMs, the backward pass of attention has been largely overlooked. In this work, we study the mathematics of the backward pass of attention, revealing that it implicitly calculates an attention matrix we refer to as "Reversed Attention". We examine the properties of Reversed Attention and demonstrate its ability to elucidate the models' behavior and edit dynamics. In an experimental setup, we showcase the ability of Reversed Attention to directly alter the forward pass of attention, without modifying the model's weights, using a novel method called "attention patching". In addition to enhancing the comprehension of how LM configure attention layers during backpropagation, Reversed Attention maps contribute to a more interpretable backward pass. Our code will be available at: https://github.

For gradient-based machine learning (ML) methods commonly adopted in practice such as stochastic gradient descent, the de facto differential privacy (DP) technique is perturbing the gradients with random Gaussian noise. Data valuation attributes the ML performance to the training data and is widely used in privacy-aware applications that require enforcing DP such as data pricing, collaborative ML, and federated learning (FL). Can existing data valuation methods still be used when DP is enforced via gradient perturbations? We show that the answer is no with the default approach of injecting i.i.d.~random noise to the gradients because the estimation uncertainty of the data value estimation paradoxically linearly scales with more estimation budget, producing estimates almost like random guesses. To address this issue, we propose to instead inject carefully correlated noise to provably remove the linear scaling of estimation uncertainty w.r.t.~the budget. We also empirically demonstrate that our method gives better data value estimates on various ML tasks and is applicable to use cases including dataset valuation and~FL.

The dramatic increase in consumption of ultra-processed food has been associated with numerous adverse health effects. Given the public health consequences linked to ultra-processed food consumption, it is highly relevant to build computational models to predict the processing of food products. We created a range of machine learning, deep learning, and NLP models to predict the extent of food processing by integrating the FNDDS dataset of food products and their nutrient profiles with their reported NOVA processing level. Starting with the full nutritional panel of 102 features, we further implemented coarse-graining of features to 65 and 13 nutrients by dropping flavonoids and then by considering the 13-nutrient panel of FDA, respectively. LGBM Classifier and Random Forest emerged as the best model for 102 and 65 nutrients, respectively, with an F1-score of 0.9411 and 0.9345 and MCC of 0.8691 and 0.8543. For the 13-nutrient panel, Gradient Boost achieved the best F1-score of 0.9284 and MCC of 0.8425. We also implemented NLP based models, which exhibited state-of-the-art performance.

Test-time adaptation (TTA) aims to fine-tune a trained model online using unlabeled testing data to adapt to new environments or out-of-distribution data, demonstrating broad application potential in real-world scenarios. However, in this optimization process, unsupervised learning objectives like entropy minimization frequently encounter noisy learning signals. These signals produce unreliable gradients, which hinder the model ability to converge to an optimal solution quickly and introduce significant instability into the optimization process. In this paper, we seek to resolve these issues from the perspective of optimizer design. Unlike prior TTA using manually designed optimizers like SGD, we employ a learning-to-optimize approach to automatically learn an optimizer, called Meta Gradient Generator (MGG). Specifically, we aim for MGG to effectively utilize historical gradient information during the online optimization process to optimize the current model. To this end, in MGG, we design a lightweight and efficient sequence modeling layer -- gradient memory layer. It exploits a self-supervised reconstruction loss to compress historical gradient information into network parameters, thereby enabling better memorization ability over a long-term adaptation process. We only need a small number of unlabeled samples to pre-train MGG, and then the trained MGG can be deployed to process unseen samples. Promising results on ImageNet-C, R, Sketch, and A indicate that our method surpasses current state-of-the-art methods with fewer updates, less data, and significantly shorter adaptation iterations. Compared with a previous SOTA method SAR, we achieve 7.4% accuracy improvement and 4.2 times faster adaptation speed on ImageNet-C.

Systems may depend on parameters which one may control, or which serve to optimise the system, or are imposed externally, or they could be uncertain. This last case is taken as the "Leitmotiv" for the following. A reduced order model is produced from the full order model by some kind of projection onto a relatively low-dimensional manifold or subspace. The parameter dependent reduction process produces a function of the parameters into the manifold. One now wants to examine the relation between the full and the reduced state for all possible parameter values of interest. Similarly, in the field of machine learning, also a function of the parameter set into the image space of the machine learning model is learned on a training set of samples, typically minimising the mean-square error. This set may be seen as a sample from some probability distribution, and thus the training is an approximate computation of the expectation, giving an approximation to the conditional expectation, a special case of an Bayesian updating where the Bayesian loss function is the mean-square error. This offers the possibility of having a combined look at these methods, and also introducing more general loss functions.

Machine learning models offer powerful predictive capabilities but often lack transparency. Local Interpretable Model-agnostic Explanations (LIME) addresses this by perturbing features and measuring their impact on a model's output. In text-based tasks, LIME typically removes present words (bits set to 1) to identify high-impact tokens. We propose \textbf{Q-LIME $\pi$} (Quantum LIME $\pi$), a quantum-inspired extension of LIME that encodes a binary feature vector in a quantum state, leveraging superposition and interference to explore local neighborhoods more efficiently. Our method focuses on flipping bits from $1 \rightarrow 0$ to emulate LIME's ``removal'' strategy, and can be extended to $0 \rightarrow 1$ where adding features is relevant. Experiments on subsets of the IMDb dataset demonstrate that Q-LIME $\pi$ often achieves near-identical top-feature rankings compared to classical LIME while exhibiting lower runtime in small- to moderate-dimensional feature spaces. This quantum-classical hybrid approach thus provides a new pathway for interpretable AI, suggesting that, with further improvements in quantum hardware and methods, quantum parallelism may facilitate more efficient local explanations for high-dimensional data.

Artifact detection algorithms are crucial to correcting the output generated by diffusion models. However, because of the variety of artifact forms, existing methods require substantial annotated data for training. This requirement limits their scalability and efficiency, which restricts their wide application. This paper shows that the similarity of denoised images between consecutive time steps during the sampling process is related to the severity of artifacts in images generated by diffusion models. Building on this observation, we introduce the concept of Similarity Trajectory to characterize the sampling process and its correlation with the image artifacts presented. Using an annotated data set of 680 images, which is only 0.1% of the amount of data used in the prior work, we trained a classifier on these trajectories to predict the presence of artifacts in images. By performing 10-fold validation testing on the balanced annotated data set, the classifier can achieve an accuracy of 72.35%, highlighting the connection between the Similarity Trajectory and the occurrence of artifacts. This approach enables differentiation between artifact-exhibiting and natural-looking images using limited training data.

This study introduces the Community Sentiment and Engagement Index (CSEI), developed to capture nuanced public sentiment and engagement variations on social media, particularly in response to major events related to COVID-19. Constructed with diverse sentiment indicators, CSEI integrates features like engagement, daily post count, compound sentiment, fine-grain sentiments (fear, surprise, joy, sadness, anger, disgust, and neutral), readability, offensiveness, and domain diversity. Each component is systematically weighted through a multi-step Principal Component Analysis (PCA)-based framework, prioritizing features according to their variance contributions across temporal sentiment shifts. This approach dynamically adjusts component importance, enabling CSEI to precisely capture high-sensitivity shifts in public sentiment. The development of CSEI showed statistically significant correlations with its constituent features, underscoring internal consistency and sensitivity to specific sentiment dimensions. CSEI's responsiveness was validated using a dataset of 4,510,178 Reddit posts about COVID-19. The analysis focused on 15 major events, including the WHO's declaration of COVID-19 as a pandemic, the first reported cases of COVID-19 across different countries, national lockdowns, vaccine developments, and crucial public health measures. Cumulative changes in CSEI revealed prominent peaks and valleys aligned with these events, indicating significant patterns in public sentiment across different phases of the pandemic. Pearson correlation analysis further confirmed a statistically significant relationship between CSEI daily fluctuations and these events (p = 0.0428), highlighting the capacity of CSEI to infer and interpret shifts in public sentiment and engagement in response to major events related to COVID-19.

Smart grid, through networked smart meters employing the non-intrusive load monitoring (NILM) technique, can considerably discern the usage patterns of residential appliances. However, this technique also incurs privacy leakage. To address this issue, we propose an innovative scheme based on adversarial attack in this paper. The scheme effectively prevents NILM models from violating appliance-level privacy, while also ensuring accurate billing calculation for users. To achieve this objective, we overcome two primary challenges. First, as NILM models fall under the category of time-series regression models, direct application of traditional adversarial attacks designed for classification tasks is not feasible. To tackle this issue, we formulate a novel adversarial attack problem tailored specifically for NILM and providing a theoretical foundation for utilizing the Jacobian of the NILM model to generate imperceptible perturbations. Leveraging the Jacobian, our scheme can produce perturbations, which effectively misleads the signal prediction of NILM models to safeguard users' appliance-level privacy. The second challenge pertains to fundamental utility requirements, where existing adversarial attack schemes struggle to achieve accurate billing calculation for users. To handle this problem, we introduce an additional constraint, mandating that the sum of added perturbations within a billing period must be precisely zero. Experimental validation on real-world power datasets REDD and UK-DALE demonstrates the efficacy of our proposed solutions, which can significantly amplify the discrepancy between the output of the targeted NILM model and the actual power signal of appliances, and enable accurate billing at the same time. Additionally, our solutions exhibit transferability, making the generated perturbation signal from one target model applicable to other diverse NILM models.

The large size of DNNs poses a significant challenge for deployment on devices with limited resources, such as mobile, edge, and IoT platforms. To address this issue, a distributed inference framework can be utilized. In this framework, a small-scale DNN (initial layers) is deployed on mobile devices, a larger version on edge devices, and the full DNN on the cloud. Samples with low complexity (easy) can be processed on mobile, those with moderate complexity (medium) on edge devices, and high complexity (hard) samples on the cloud. Given that the complexity of each sample is unknown in advance, the crucial question in distributed inference is determining the sample complexity for appropriate DNN processing. We introduce a novel method named \our{}, which leverages the Data Cartography approach initially proposed for enhancing DNN generalization. By employing data cartography, we assess sample complexity. \our{} aims to boost accuracy while considering the offloading costs from mobile to edge/cloud. Our experimental results on GLUE datasets, covering a variety of NLP tasks, indicate that our approach significantly lowers inference costs by more than 43\% while maintaining a minimal accuracy drop of less than 0.5\% compared to performing all inferences on the cloud. The source code is available at https://anonymous.4open.science/r/DIMEC-1B04.